# Jetson Orin

Jetson Orin is an NVIDIA AI computer and module series for energy-efficient autonomous machines, supporting concurrent AI inference pipelines and sensor interfaces.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Achieving Near-Zero-Overhead Multi-Model Hierarchical Classification in Real-Time Detection Pipelines

DevFeed: [Achieving Near-Zero-Overhead Multi-Model Hierarchical Classification in Real-Time Detection Pipelines](<https://devfeed.tech/articles/achieving-near-zero-overhead-multi-model-hierarchical-classification-in-real-time-detection-pipelines-144919.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2608.11770>)

Author: Vaishnav Raju

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [quantization](<https://devfeed.tech/topics/quantization.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [real-time](<https://devfeed.tech/topics/real-time.md>)

Tags: [accelerators](<https://devfeed.tech/tags/accelerators.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-vehicles](<https://devfeed.tech/tags/autonomous-vehicles.md>), [bottleneck](<https://devfeed.tech/tags/bottleneck.md>), [classification](<https://devfeed.tech/tags/classification.md>), [classifier](<https://devfeed.tech/tags/classifier.md>), [classifiers](<https://devfeed.tech/tags/classifiers.md>), [compilation](<https://devfeed.tech/tags/compilation.md>), [concurrent](<https://devfeed.tech/tags/concurrent.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [custom](<https://devfeed.tech/tags/custom.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [detection](<https://devfeed.tech/tags/detection.md>), [drone](<https://devfeed.tech/tags/drone.md>), [edge](<https://devfeed.tech/tags/edge.md>), [overhead](<https://devfeed.tech/tags/overhead.md>), [pipelines](<https://devfeed.tech/tags/pipelines.md>), [real-time](<https://devfeed.tech/tags/real-time.md>)

### AI overview

The paper reports that a GPU detector paired with DLA classification on a Jetson Orin NX ran at 12.5 FPS at 1080p, compared with 13.3 FPS for the detector alone. Its five-step deployment method covers adapting classification architectures, repairing TensorRT implicit quantization ranges for pipeline validation (raising reported accuracy from 75% to 94%), quantization-aware training, ONNX graph surgery, and concurrent GPU/DLA inference. The authors also describe nine engineering constraints and report dual-DLA scaling without additional cost. Evidence here is limited to the abstract.

### Source excerpt

arXiv:2608.11770v2 Announce Type: replace-cross Abstract: Edge-deployed vision systems in target recognition, surveillance, autonomous vehicles, and drone domains require hierarchical inference pipelines where a detection model identifies objects of interest and downstream classifiers provide fine-grained attribute analysis. Running all models on the GPU creates a serial bottleneck that limits real-time throughput as pipeline stages grow. Modern edge SoCs pair GPUs with dedicated neural accelerators (NPUs, DLAs) capable of concurrent execution, yet deploying custom models on these accelerators remains impractical due to strict operator constraints, quantization incompatibilities, and an undocumented end-to-end pipeline. We target NVIDIA Jetson DLA cores as the representative platform. We present a five-step methodology for zero GPU fallback DLA INT8 deployment of classification backbones, comprising architecture adaptation, manual dynamic range workaround to rescue TensorRT's implicit quantization (recovering 94.0% accuracy from implicit quantization's 75%) for rapid pipeline validation before explicit quantization, quantization-aware training, ONNX graph surgery for DLA compilation, and a concurrent GPU-detection/DLA-classification inference pipeline. We document nine engineering constraints with root-cause analysis and generalizable solutions. Validation on a dual-head person attribute classifier running on DLA alongside a GPU object detector on a Jetson Orin NX demonstrates near-zero pipeline overhead (12.5 vs. 13.3~FPS detector-only at 1080p), with dual-DLA scaling at no additional cost. The methodology is backbone-agnostic and generalizes to any detection-classification edge pipeline.

## NVIDIA and Microsoft Expand Accelerated Computing From Azure to Windows PCs

DevFeed: [NVIDIA and Microsoft Expand Accelerated Computing From Azure to Windows PCs](<https://devfeed.tech/articles/nvidia-and-microsoft-expand-accelerated-computing-from-azure-to-windows-pcs-144456.md>)

Original publisher: [Read original article](<https://blogs.nvidia.com/blog/microsoft-ignite-accelerated-computing-azure-windows-pcs/>)

Author: Dave Salvator

Published: 2026-11-19T20:00:15Z

Content type: release

Language: en

Sources: [NVIDIA](<https://devfeed.tech/sources/nvidia-blog.md>)

Topics: [Nvidia](<https://devfeed.tech/topics/nvidia.md>), [Microsoft](<https://devfeed.tech/topics/microsoft.md>), [Azure](<https://devfeed.tech/topics/azure.md>), [Vera Rubin NVL72](<https://devfeed.tech/topics/vera-rubin-nvl72.md>), [NVIDIA Vera Rubin](<https://devfeed.tech/topics/nvidia-vera-rubin.md>), [RTX PRO](<https://devfeed.tech/topics/rtx-pro.md>), [Jetson](<https://devfeed.tech/topics/jetson.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Thor](<https://devfeed.tech/topics/thor.md>), [Windows](<https://devfeed.tech/topics/windows.md>), [virtual machines](<https://devfeed.tech/topics/virtual-machines.md>)

Tags: [accelerated-computing](<https://devfeed.tech/tags/accelerated-computing.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [applications](<https://devfeed.tech/tags/applications.md>), [availability](<https://devfeed.tech/tags/availability.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-virtual-machines](<https://devfeed.tech/tags/azure-virtual-machines.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [collaboration](<https://devfeed.tech/tags/collaboration.md>), [cosmos](<https://devfeed.tech/tags/cosmos.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [full-stack](<https://devfeed.tech/tags/full-stack.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [ignite](<https://devfeed.tech/tags/ignite.md>), [industrial](<https://devfeed.tech/tags/industrial.md>), [inference](<https://devfeed.tech/tags/inference.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [microsoft-ignite](<https://devfeed.tech/tags/microsoft-ignite.md>), [nemotron](<https://devfeed.tech/tags/nemotron.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-blackwell](<https://devfeed.tech/tags/nvidia-blackwell.md>), [nvidia-vera](<https://devfeed.tech/tags/nvidia-vera.md>), [nvidia-vera-rubin](<https://devfeed.tech/tags/nvidia-vera-rubin.md>), [nvl72](<https://devfeed.tech/tags/nvl72.md>), [organizations](<https://devfeed.tech/tags/organizations.md>), [software](<https://devfeed.tech/tags/software.md>), [windows](<https://devfeed.tech/tags/windows.md>), [windows-pcs](<https://devfeed.tech/tags/windows-pcs.md>)

### AI overview

NVIDIA and Microsoft are extending AI deployment from Azure data centers to enterprise systems and Windows PCs. Azure is running Vera Rubin NVL72 systems, with work on confidential computing protections and power efficiency; reported MLPerf results describe Azure scaling to 10,240 GB200 GPUs for training and leading cloud submissions on a DeepSeek-R1 inference measure. Microsoft Foundry adds managed post-training and deployment options for customized open models. EY reports that adapting Nemotron for purchasing-record classification improved evaluation accuracy by 55% and reduced cost 15x versus proprietary frontier models. Azure Local is adding Nemotron capabilities and support for RTX PRO 6000 Blackwell GPUs, Jetson edge platforms, and physical AI models. The companies are also bringing scientific AI tools into Microsoft Discovery and using NVIDIA Dynamo for Microsoft AI post-training inference. Windows 365 GPU Cloud PCs are slated for a limited preview, with public preview planned for early 2027.

### Source excerpt

NVIDIA and Microsoft are extending their full-stack collaboration to help organizations build, customize and deploy intelligent applications -- from next-generation infrastructure on Azure to industrial systems and Windows PCs. At Microsoft Ignite, the companies are expanding access to Azure virtual machines powered by NVIDIA Vera Rubin NVL72 systems, as well as adding customer availability to [...]

## Jetson Orin Nano vs Orin NX: How to Choose for Industrial Edge AI

DevFeed: [Jetson Orin Nano vs Orin NX: How to Choose for Industrial Edge AI](<https://devfeed.tech/articles/jetson-orin-nano-vs-orin-nx-how-to-choose-for-industrial-edge-ai-142317.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/11/19/jetson-orin-nano-vs-orin-nx-how-to-choose-for-industrial-edge-ai/>)

Author: Sunny Wang1

Published: 2026-11-19T07:40:33Z

Content type: comparison

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Jetson](<https://devfeed.tech/topics/jetson.md>), [seeed studio](<https://devfeed.tech/topics/seeed-studio.md>)

Tags: [4k](<https://devfeed.tech/tags/4k.md>), [ai](<https://devfeed.tech/tags/ai.md>), [camera](<https://devfeed.tech/tags/camera.md>), [edge](<https://devfeed.tech/tags/edge.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [feature](<https://devfeed.tech/tags/feature.md>), [industrial](<https://devfeed.tech/tags/industrial.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [know-how](<https://devfeed.tech/tags/know-how.md>), [news](<https://devfeed.tech/tags/news.md>), [nvdia](<https://devfeed.tech/tags/nvdia.md>), [nx](<https://devfeed.tech/tags/nx.md>), [selection-guide](<https://devfeed.tech/tags/selection-guide.md>), [tech](<https://devfeed.tech/tags/tech.md>), [vs](<https://devfeed.tech/tags/vs.md>)

### AI overview

For multi-stream industrial vision, the Orin NX 16GB can offer meaningful pipeline headroom over the Orin Nano 8GB, especially when recording is required: in a ten-minute four-stream 4K H.264 benchmark, it delivered 1.33x inference throughput, 1.51x decode throughput, 63.3% lower normalized CPU load, and 69.7% lower end-to-end latency. The benchmark used TensorRT FP16 YOLOv8n, 5 FPS recording per stream, and excluded 30 seconds of warm-up; the Nano used CPU-based x264 while the NX used NVENC, so the results reflect the full configured pipeline, not isolated encoder or inference performance. The NX also doubles memory and raises bandwidth, but its higher-power configuration reached a 97.38°C peak under the stated passive-cooling setup, making thermal validation important. The Nano may suit bounded workloads that meet throughput and memory needs without substantial recording. Check actual camera codecs and rates, measure the complete workload and memory use, and validate sustained operation at intended power and cooling conditions before choosing.

### Source excerpt

A proof of concept passes on an Orin Nano: one camera, one model, 30 FPS, The post Jetson Orin Nano vs Orin NX: How to Choose for Industrial Edge AI appeared first on Latest News from Seeed Studio.

## HELENA for 5G NR LEO NTN Channel Estimation: A Comparative Evaluation

DevFeed: [HELENA for 5G NR LEO NTN Channel Estimation: A Comparative Evaluation](<https://devfeed.tech/articles/helena-for-5g-nr-leo-ntn-channel-estimation-a-comparative-evaluation-137016.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.14735>)

Author: Miguel Camelo Botero, Nina Slamnik-Krije\\v{s}torac, Johann Marquez-Barja

Published: 2026-11-17T05:00:00Z

Content type: comparison

Language: en

Sources: [arXiv: Networking](<https://devfeed.tech/sources/arxiv-networking.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Deep learning](<https://devfeed.tech/topics/deep-learning.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [TensorRT](<https://devfeed.tech/topics/tensorrt.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [RTX PRO](<https://devfeed.tech/topics/rtx-pro.md>)

Tags: [5g](<https://devfeed.tech/tags/5g.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [attention](<https://devfeed.tech/tags/attention.md>), [channel](<https://devfeed.tech/tags/channel.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ni](<https://devfeed.tech/tags/cs-ni.md>), [cs-pf](<https://devfeed.tech/tags/cs-pf.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [doppler](<https://devfeed.tech/tags/doppler.md>), [eess-sp](<https://devfeed.tech/tags/eess-sp.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [energy](<https://devfeed.tech/tags/energy.md>), [estimator](<https://devfeed.tech/tags/estimator.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fp16](<https://devfeed.tech/tags/fp16.md>), [high-performance](<https://devfeed.tech/tags/high-performance.md>), [inference](<https://devfeed.tech/tags/inference.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [latency](<https://devfeed.tech/tags/latency.md>), [leo](<https://devfeed.tech/tags/leo.md>), [low-earth-orbit](<https://devfeed.tech/tags/low-earth-orbit.md>), [non-terrestrial-networks](<https://devfeed.tech/tags/non-terrestrial-networks.md>), [ntn](<https://devfeed.tech/tags/ntn.md>)

### AI overview

The study finds HELENA's unchanged dual-attention architecture remains competitive for the evaluated LEO NTN channel-estimation task after retraining, without an NTN-specific redesign. Its SNR-averaged NMSE is within 0.29 dB of AttRNet with receiver compensation and 0.72 dB with residual impairments. HELENA beats global LMMSE in the compensated condition and all evaluated LMMSE variants under residual impairments, though profile-aware filtering is more accurate with compensation. FP16 TensorRT inference on an RTX PRO 4500 reaches 0.0573 ms P99, under the 0.5 ms budget. On a 10 W Jetson Orin NX, HELENA uses 66.0% less energy per sample than AttRNet, but its 0.9620 ms P99 misses that budget; only the less accurate LSiDNN-48 meets it. Results suggest embedded latency remains a challenge for accurate estimation.

### Source excerpt

arXiv:2609.14735v2 Announce Type: replace-cross Abstract: Deep Learning (DL)-based channel estimation has shown high accuracy and low latency in terrestrial 5G NR, but Low Earth Orbit (LEO) Non-Terrestrial Networks (NTNs) introduce Doppler and synchronization impairments that may require NTN-specific architectures. We investigate whether HELENA, a dual-attention neural channel estimator originally designed for terrestrial channels, remains effective after NTN retraining and suitable across high-performance and power-constrained inference platforms. Its unchanged architecture is trained on paired receiver-compensated and residual-impaired datasets and evaluated alongside terrestrial-origin, NTN-specific, and classical estimators. HELENA's SNR-averaged Normalized Mean Squared Error (NMSE), computed by arithmetic averaging in linear scale, remains within 0.29 and 0.72 dB of the most accurate learned estimator, AttRNet, in the two conditions, respectively. It outperforms global Linear Minimum Mean Squared Error (LMMSE) under compensation and all evaluated LMMSE variants under residual impairments, although profile-aware filtering achieves lower error under compensation. On an RTX PRO 4500, FP16 TensorRT inference reaches 0.0573 ms 99th-Percentile (P99) latency, below the adopted 0.5 ms estimator budget. On a 10 W Jetson Orin NX, HELENA consumes 66.0% less energy per sample than AttRNet, but its 0.9620 ms P99 exceeds the budget; only the less accurate LSiDNN-48 meets it. Thus, HELENA remains effective without NTN-specific redesign for the evaluated task, while accurate embedded inference remains a latency challenge.

## Run Native DLA Inference in Python on NVIDIA Jetson AGX Orin Using cuDLA

DevFeed: [Run Native DLA Inference in Python on NVIDIA Jetson AGX Orin Using cuDLA](<https://devfeed.tech/articles/run-native-dla-inference-in-python-on-nvidia-jetson-agx-orin-using-cudla-133567.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/run-native-dla-inference-in-python-on-nvidia-jetson-agx-orin-using-cudla/>)

Author: Nikshay Shrivastava

Published: 2026-11-13T18:57:10Z

Content type: tutorial

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Jetson](<https://devfeed.tech/topics/jetson.md>), [Deep learning](<https://devfeed.tech/topics/deep-learning.md>), [On-device AI](<https://devfeed.tech/topics/on-device-ai.md>), [TensorRT](<https://devfeed.tech/topics/tensorrt.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-applications](<https://devfeed.tech/tags/ai-applications.md>), [ai-inference](<https://devfeed.tech/tags/ai-inference.md>), [applications](<https://devfeed.tech/tags/applications.md>), [compute](<https://devfeed.tech/tags/compute.md>), [computer-vision-video-analytics](<https://devfeed.tech/tags/computer-vision-video-analytics.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [edge](<https://devfeed.tech/tags/edge.md>), [inference](<https://devfeed.tech/tags/inference.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [networks](<https://devfeed.tech/tags/networks.md>), [neural-networks](<https://devfeed.tech/tags/neural-networks.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [nvidia-jetson-agx-orin](<https://devfeed.tech/tags/nvidia-jetson-agx-orin.md>), [python](<https://devfeed.tech/tags/python.md>), [robotics](<https://devfeed.tech/tags/robotics.md>)

### AI overview

The excerpt describes using `cuda.bindings.cudla` 13.3.0+ to run a TensorRT-compiled DLA loadable from Python on Jetson Orin. JetPack provides `libcudla.so`; inference needs that runtime, while TensorRT is needed only to compile the model. Python 3.10 or later is required. The demonstrated workflow uses hybrid `CUDA_DLA` mode and one CUDA stream to coordinate GPU work with DLA inference: load the model bytes, query tensor descriptors, allocate and register GPU buffers, enqueue the task, synchronize, and copy results back for Python postprocessing. Descriptor fields provide sizes and layouts, but preprocessing and output interpretation remain model-specific. Not every layer is DLA-compatible, so the ONNX graph may need trimming. Standalone mode is unsupported in this release. The excerpt says the cuDLA API is identical on RTX Spark, though its detailed setup is centered on Jetson.

### Source excerpt

AI applications at the edge often need to run multiple neural networks within limited compute and power budgets. NVIDIA Deep Learning Accelerator (DLA)...

## How to Set Up a Basler Camera on NVIDIA Jetson with Roboflow

DevFeed: [How to Set Up a Basler Camera on NVIDIA Jetson with Roboflow](<https://devfeed.tech/articles/how-to-set-up-a-basler-camera-on-nvidia-jetson-with-roboflow-132522.md>)

Original publisher: [Read original article](<https://roboflow.com/blog/basler-jetson-deployment>)

Author: Shantanu Bala

Published: 2026-11-13T11:36:00Z

Content type: tutorial

Language: en

Sources: [Roboflow](<https://devfeed.tech/sources/roboflow-blog.md>)

Topics: [Jetson](<https://devfeed.tech/topics/jetson.md>), [webcam](<https://devfeed.tech/topics/webcam.md>), [Documentation](<https://devfeed.tech/topics/documentation.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>)

Tags: [automate](<https://devfeed.tech/tags/automate.md>), [browser](<https://devfeed.tech/tags/browser.md>), [camera](<https://devfeed.tech/tags/camera.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [deployment-manager-edge-fleet](<https://devfeed.tech/tags/deployment-manager-edge-fleet.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [live](<https://devfeed.tech/tags/live.md>), [manager](<https://devfeed.tech/tags/manager.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [predictions](<https://devfeed.tech/tags/predictions.md>), [view](<https://devfeed.tech/tags/view.md>)

### AI overview

Roboflow Deployment Manager simplifies connecting Basler GigE or USB3 cameras to an NVIDIA Jetson: register a Debian device, connect the camera, then create a Stream by entering its serial number and choosing a Workflow. The device discovers the camera and, for GigE, assigns a reachable address; automatic network configuration and camera recovery are enabled by default, with optional network overrides. No local monitor or pylon installation is required. Streams may take several minutes to provision, and an unsupported Workflow block requires updating the device's Inference version. The browser view provides live predictions, frame-rate and sharpness metrics, per-block performance, and camera/network diagnostics. Deployment Manager is built primarily for Orin-based Jetsons; shared cameras across Streams require Multiprocessing.

### Source excerpt

Set up a Basler camera on NVIDIA Jetson with Roboflow Deployment Manager. Automate camera configuration and view live predictions in your browser.

## LazyWeights: Decode-on-Demand Edge LLMs

DevFeed: [LazyWeights: Decode-on-Demand Edge LLMs](<https://devfeed.tech/articles/lazyweights-decode-on-demand-edge-llms-133173.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.17108>)

Author: Arnab Sanyal, Gourav Datta, Sandeep P. Chinchali

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Jetson](<https://devfeed.tech/topics/jetson.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [CPU Cache](<https://devfeed.tech/topics/cpu-cache.md>), [communications](<https://devfeed.tech/topics/communications.md>), [Tensor Cores](<https://devfeed.tech/topics/tensor-cores.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [amd](<https://devfeed.tech/tags/amd.md>), [area](<https://devfeed.tech/tags/area.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [asic](<https://devfeed.tech/tags/asic.md>), [awq](<https://devfeed.tech/tags/awq.md>), [bandwidth](<https://devfeed.tech/tags/bandwidth.md>), [beats](<https://devfeed.tech/tags/beats.md>), [best](<https://devfeed.tech/tags/best.md>), [bits](<https://devfeed.tech/tags/bits.md>), [but](<https://devfeed.tech/tags/but.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [code](<https://devfeed.tech/tags/code.md>), [compress](<https://devfeed.tech/tags/compress.md>), [cs-ar](<https://devfeed.tech/tags/cs-ar.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-it](<https://devfeed.tech/tags/cs-it.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [decode](<https://devfeed.tech/tags/decode.md>), [demand](<https://devfeed.tech/tags/demand.md>), [edge](<https://devfeed.tech/tags/edge.md>), [llms](<https://devfeed.tech/tags/llms.md>), [math-it](<https://devfeed.tech/tags/math-it.md>)

### AI overview

LazyWeights reduces edge LLM memory traffic by tuning GPTQ codes for compression, Huffman-coding them losslessly in interleaved lanes, and decoding on demand between DRAM and L2 while leaving GPU kernels unchanged. At stated 1% and 3.5% perplexity budgets, three tested models were 5.7-31.4% smaller than the best compared llama.cpp formats. The abstract reports up to 36% lower latency and 38% lower energy per token on Jetson Orin Nano Super, with larger reductions using prompt-lookup speculation; on Jetson Nano it reports 58-64% less time and 53-57% less energy than CPU llama.cpp. A streamed Mistral-7B ran 6.6x faster at 3.89 bits per weight. The decoder was evaluated as a small ASIC unit and shown bit-exact on a Zynq UltraScale+ FPGA. These are abstract-level reported results; experimental details are not included.

### Source excerpt

arXiv:2611.17108v1 Announce Type: new Abstract: Memory decides which LLMs an edge device runs, and how fast: capacity on a $4$ GB Jetson Nano, bandwidth on an $8$ GB Jetson Orin Nano Super. Deployed formats (GPTQ, AWQ, $\texttt{llama.cpp}$'s K- and I-quants) spend equal bits on every integer code of a matrix, although the codes are far from uniform. $\texttt{LazyWeights}$ keeps these formats but tunes GPTQ so the codes compress, Huffman-codes them losslessly in interleaved lanes, and decodes them on demand in a small unit between DRAM and L2, so GPU kernels run unchanged and DRAM moves only coded bytes. At $1\%$ and $3.5\%$ perplexity budgets, coded Phi-3-mini, Mistral-7B and SmolLM-1.7B are $5.7-31.4\%$ smaller than the best $\texttt{llama.cpp}$ format. A bit-exact $0.21$ mm$^2$ ASIC decoder unit ($1$ GHz, $7$ nm) cuts Orin latency and energy per token by up to $36\%$ and $38\%$ ($54\%$ and $44\%$ with prompt-lookup speculation). On the Nano, $\texttt{LazyWeights}$ needs $58-64\%$ less time and $53-57\%$ less energy per token than CPU $\texttt{llama.cpp}$, and runs a streamed Mistral-7B 6.6$\times$ faster at $3.89$ bits per weight. A Huffman decoder beats an rANS decoder in rate and area at equal throughput; the ASIC decoder unit runs bit-exact on a Zynq UltraScale+ FPGA (AMD ZCU102).

## Joint Quantization and Dynamic Partitioning for Context-Aware Edge Vision Transformers

DevFeed: [Joint Quantization and Dynamic Partitioning for Context-Aware Edge Vision Transformers](<https://devfeed.tech/articles/joint-quantization-and-dynamic-partitioning-for-context-aware-edge-vision-transformers-128406.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.14196>)

Author: Zhu Jiaqi, Jain Prakriti, Katare Dewant, Kumar Saurabh

Published: 2026-11-11T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [quantization](<https://devfeed.tech/topics/quantization.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [autonomous vehicles](<https://devfeed.tech/topics/autonomous-vehicles.md>), [object-detection](<https://devfeed.tech/topics/object-detection.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [bandwidth](<https://devfeed.tech/tags/bandwidth.md>), [communication](<https://devfeed.tech/tags/communication.md>), [compute](<https://devfeed.tech/tags/compute.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [edge](<https://devfeed.tech/tags/edge.md>), [int8](<https://devfeed.tech/tags/int8.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [jetson-orin-nx](<https://devfeed.tech/tags/jetson-orin-nx.md>), [latency](<https://devfeed.tech/tags/latency.md>), [link](<https://devfeed.tech/tags/link.md>), [model](<https://devfeed.tech/tags/model.md>), [orchestration](<https://devfeed.tech/tags/orchestration.md>), [partition](<https://devfeed.tech/tags/partition.md>), [partitioning](<https://devfeed.tech/tags/partitioning.md>), [quantization](<https://devfeed.tech/tags/quantization.md>), [transformers](<https://devfeed.tech/tags/transformers.md>), [vision](<https://devfeed.tech/tags/vision.md>), [vision-transformers](<https://devfeed.tech/tags/vision-transformers.md>)

### AI overview

The framework jointly chooses model split points, per-layer quantization bitwidths, and execution placement across heterogeneous vehicle compute nodes. It formulates these choices as a MILP minimizing compute and communication energy under latency and task-quality constraints, addressing changing V2V bandwidth, processing load, and concurrent streams. In OPV2V-driven workloads across Jetson Orin Nano, Orin NX, and AGX Orin nodes with 2-8 streams, it reduces V2V communication energy by 33.3-36.0% versus a Uniform INT8 split. At eight streams, latency falls by 34.0% on Nano and 33.1% on NX versus that baseline, while 3D detection accuracy is 3.1-4.1 mAP50 points below FP16. These are reported evaluation results; the excerpt does not establish performance beyond the tested configurations.

### Source excerpt

arXiv:2611.14196v1 Announce Type: new Abstract: Deploying vision transformer models across connected and automated vehicles (CAVs) requires joint management of computation and Vehicle-to-Vehicle (V2V) communication. Static split-inference methods use fixed partition points and cannot adapt to time-varying link bandwidth, processing load, and concurrent multi-stream traffic. This paper proposes an edge orchestration framework that jointly selects model partition points, per-layer quantization bitwidths, and execution placement across heterogeneous vehicular compute nodes. The joint partitioning and mixed-precision allocation problem is formulated as a Mixed-Integer Linear Program (MILP) that minimizes compute and communication energy subject to end-to-end latency and task-quality constraints. The framework is evaluated on Jetson Orin Nano (8~GB), Jetson Orin NX (16~GB), and Jetson AGX Orin (32~GB) nodes using OPV2V-driven multi-stream workloads with $K \in \{2,4,6,8\}$. Across the evaluated configurations, the proposed formulation reduces V2V communication energy by $33.3$--$36.0\%$ relative to a Uniform INT8 split. At $K=8$, end-to-end task completion latency is reduced by $34.0\%$ on the Orin Nano and $33.1\%$ on the Orin NX relative to the same baseline, while limiting the reduction in 3D detection accuracy to $3.1$--$4.1$ percentage points in $\mathrm{mAP}_{50}$ relative to FP16 execution.

## A Unified Framework for License Plate Identification and Character-Recognition

DevFeed: [A Unified Framework for License Plate Identification and Character-Recognition](<https://devfeed.tech/articles/a-unified-framework-for-license-plate-identification-and-character-recognition-125354.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.11289>)

Author: Saurabh Kumar, Dewant Katare, Prakriti Jain

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [image recognition](<https://devfeed.tech/topics/image-recognition.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [TensorRT](<https://devfeed.tech/topics/tensorrt.md>)

Tags: [ai-assisted](<https://devfeed.tech/tags/ai-assisted.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [detection](<https://devfeed.tech/tags/detection.md>), [edge](<https://devfeed.tech/tags/edge.md>), [framework](<https://devfeed.tech/tags/framework.md>), [int8](<https://devfeed.tech/tags/int8.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [latency](<https://devfeed.tech/tags/latency.md>), [license-plate](<https://devfeed.tech/tags/license-plate.md>), [memory](<https://devfeed.tech/tags/memory.md>), [power-consumption](<https://devfeed.tech/tags/power-consumption.md>), [recognition](<https://devfeed.tech/tags/recognition.md>), [tensorrt](<https://devfeed.tech/tags/tensorrt.md>), [throughput](<https://devfeed.tech/tags/throughput.md>)

### AI overview

The framework combines license plate detection and character recognition in one network for edge hardware. A shared CSP-Darknet/PANet encoder feeds an anchor-free localization head and a RoIAlign sequence-recognition branch that decodes characters from aligned feature maps, avoiding image-space cropping and a second recognition backbone. Trained on about 37,000 manually verified images covering all 50 U.S. states and varied capture conditions, its INT8 TensorRT deployment on an 8 GB NVIDIA Jetson Orin Nano reports 97.8% mAP50 and 90.7% full-sequence accuracy. It reaches 147 FPS at 6.8 ms end-to-end latency, using 315 MB inference memory and 9.4 W average power. Compared with the YOLOv8s+CRNN baseline, it reports 7.1x throughput and 77.8% less allocated inference memory.

### Source excerpt

arXiv:2611.11289v1 Announce Type: new Abstract: This paper presents a unified framework for license plate detection and character recognition on low-power edge hardware. The proposed network uses a shared CSP-Darknet and PANet feature encoder with an anchor-free localization head and a RoIAlign-based sequence-recognition branch. Character decoding is performed directly on aligned feature maps, avoiding image-space plate cropping and a second recognition backbone. The model is trained on an AI-assisted dataset of approximately 37,000 manually verified images covering license plate layouts from all 50 United States. The dataset includes variations in illumination, viewpoint, weather, blur, occlusion, and plate design. The framework is converted to an INT8 TensorRT engine and evaluated on an NVIDIA Jetson Orin Nano with 8 GB memory. The INT8 model shows an mAP50 of 97.8% and a full-sequence accuracy of 90.7%. It records 147 FPS at 6.8 ms end-to-end latency, with 315 MB inference memory and 9.4 W average power consumption. Relative to the YOLOv8s+CRNN baseline, the proposed deployment provides a 7.1x throughput increase and a 77.8% reduction in allocated inference memory

## Multi-Node Transformers for Object Detection and Captioning Across Distributed Edge Networks

DevFeed: [Multi-Node Transformers for Object Detection and Captioning Across Distributed Edge Networks](<https://devfeed.tech/articles/multi-node-transformers-for-object-detection-and-captioning-across-distributed-edge-networks-125355.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.11291>)

Author: Saurabh Kumar, Dewant Katare, Prakriti Jain

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Transformers](<https://devfeed.tech/topics/transformers.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [distributed algorithm](<https://devfeed.tech/topics/distributed-algorithm.md>), [Jetson](<https://devfeed.tech/topics/jetson.md>)

Tags: [autoregressive](<https://devfeed.tech/tags/autoregressive.md>), [concurrent](<https://devfeed.tech/tags/concurrent.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [detection](<https://devfeed.tech/tags/detection.md>), [device](<https://devfeed.tech/tags/device.md>), [distributed](<https://devfeed.tech/tags/distributed.md>), [edge](<https://devfeed.tech/tags/edge.md>), [execution](<https://devfeed.tech/tags/execution.md>), [int8](<https://devfeed.tech/tags/int8.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [latency](<https://devfeed.tech/tags/latency.md>), [memory](<https://devfeed.tech/tags/memory.md>), [networks](<https://devfeed.tech/tags/networks.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [object-detection](<https://devfeed.tech/tags/object-detection.md>), [pi-5](<https://devfeed.tech/tags/pi-5.md>), [raspberry-pi-5](<https://devfeed.tech/tags/raspberry-pi-5.md>), [transformers](<https://devfeed.tech/tags/transformers.md>)

### AI overview

The framework shares one visual encoder across object detection and dense region captioning, reusing image features for localization and description instead of encoding the image a second time. It allocates model stages and input streams across edge nodes based on execution time, memory, processor use, power limits, and tensor-transfer cost. On a six-device testbed, the unified INT8 model reached 38.1 FPS on a Jetson Orin Nano, using 3.5 GB peak memory and 11.8 W. With four concurrent streams, the proposed allocation achieved 35.2 FPS and 28.4 ms mean end-to-end latency, with 6.8 MB/s internode transfer.

### Source excerpt

arXiv:2611.11291v1 Announce Type: cross Abstract: This paper presents a unified transformer-based framework for object detection and dense region captioning across distributed edge networks. The architecture uses a shared visual encoder followed by detection, region extraction, and autoregressive language-decoding stages, allowing localization and region description to reuse the same image features. This design removes the second image-encoding path required by decoupled detection and captioning pipelines. To support execution on resource-limited devices, a workload-allocation formulation assigns model stages and input streams across connected edge nodes while accounting for execution time, memory capacity, processor utilization, power limits, and intermediate tensor transfer cost. The framework is evaluated on a six-device edge testbed comprising four NVIDIA Jetson Orin Nano and Orin NX inference nodes and two Raspberry Pi 5 auxiliary CPU nodes, with TensorRT FP16 and INT8 execution on the NVIDIA devices. On a Jetson Orin Nano, the unified INT8 model achieves 38.1 FPS with 3.5 GB peak memory use and 11.8 W measured power. Under four concurrent streams, the proposed allocation achieves 35.2 FPS and 28.4 ms mean end-to-end latency with an internode transfer rate of 6.8 MB/s.

## HMMF-Net: A Lightweight Hierarchical Multi-Frame Multimodal Fusion Network for Accurate Fish Feeding Intensity Assessment

DevFeed: [HMMF-Net: A Lightweight Hierarchical Multi-Frame Multimodal Fusion Network for Accurate Fish Feeding Intensity Assessment](<https://devfeed.tech/articles/hmmf-net-a-lightweight-hierarchical-multi-frame-multimodal-fusion-network-for-accurate-fish-feeding-intensity-assessment-124638.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.09224>)

Author: Phi-Hung Hoang, Van-Manh Tran, Nam-Thuan Trinh, Thi-Thu-Hong Phan

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [TensorRT](<https://devfeed.tech/topics/tensorrt.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3d](<https://devfeed.tech/tags/3d.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [assessment](<https://devfeed.tech/tags/assessment.md>), [attention](<https://devfeed.tech/tags/attention.md>), [automated](<https://devfeed.tech/tags/automated.md>), [av](<https://devfeed.tech/tags/av.md>), [axis](<https://devfeed.tech/tags/axis.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [best](<https://devfeed.tech/tags/best.md>), [branch](<https://devfeed.tech/tags/branch.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [fish](<https://devfeed.tech/tags/fish.md>), [frame](<https://devfeed.tech/tags/frame.md>), [fusion](<https://devfeed.tech/tags/fusion.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [network](<https://devfeed.tech/tags/network.md>)

### AI overview

HMMF-Net combines temporally separated video frames and log-mel audio features, then progressively exchanges information between them through compact messenger tokens. Its visual branch adapts 2D EfficientNetB0; its acoustic branch uses adaptive multi-axis residual units to model temporal, frequency, and local time-frequency patterns. On AV-FFIA five-fold cross-validation, it reports 96.77% accuracy and 96.88% F1 with 3.987M parameters and 1.002 GFLOPs. Against the best visual configuration's 95.49% accuracy, the authors report a 28.38% reduction in classification error. On Jetson Orin Nano with TensorRT FP32, inference took 11.54 ms per 2-second sample and the model was 18.81 MB; the abstract says accuracy remained comparable.

### Source excerpt

arXiv:2611.09224v1 Announce Type: cross Abstract: Automated fish feeding intensity assessment (FFIA) is important for precision aquaculture and feed cost optimization. However, unimodal methods are vulnerable to acoustic noise and visual disturbances, while existing multimodal frameworks often rely on computationally intensive 3D convolutions or Transformers, limiting their use on edge devices. To address these limitations, we propose HMMF-Net, a lightweight multimodal architecture that combines early multi-frame visual encoding and multi-axis acoustic modeling through a hierarchical bidirectional messenger cross-modal fusion (BMCF) mechanism. The visual branch captures temporal variations in feeding behavior by stacking temporally distant frames as a multi-channel input to an adapted 2D EfficientNetB0 backbone. The acoustic branch employs adaptive multi-axis residual units (AMARU) to extract temporal, frequency, and local time-frequency representations from log-mel spectrograms. BMCF progressively exchanges information between the two modalities at intermediate representation levels through compact messenger tokens, reducing the computational overhead of dense cross-modal attention. Evaluated on the AV-FFIA dataset using five-fold cross-validation, HMMF-Net achieves 96.77% accuracy and 96.88% F1-score on the held-out test folds, with 3.987M parameters and 1.002 GFLOPs. Compared with the best-performing visual configuration (95.49% accuracy), HMMF-Net reduces the classification error by 28.38%. Deployment on an NVIDIA Jetson Orin Nano using TensorRT with FP32 precision maintains comparable accuracy, with an inference time of 11.54 ms per 2-second sample and a model size of 18.81 MB. These results demonstrate that HMMF-Net combines effective multimodal representation learning with computational efficiency, enabling accurate and practical FFIA on resource-constrained edge platforms.

## Energy-Aware Open-Vocabulary Detection Scheduling for Low-Power Robotic Autonomy

DevFeed: [Energy-Aware Open-Vocabulary Detection Scheduling for Low-Power Robotic Autonomy](<https://devfeed.tech/articles/energy-aware-open-vocabulary-detection-scheduling-for-low-power-robotic-autonomy-120229.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.07663>)

Author: Abdul Basit, Muhammad Shafique

Published: 2026-11-06T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [scheduling](<https://devfeed.tech/topics/scheduling.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Sustainable Computing](<https://devfeed.tech/topics/sustainable-computing.md>)

Tags: [agilex](<https://devfeed.tech/tags/agilex.md>), [autonomy](<https://devfeed.tech/tags/autonomy.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [detection](<https://devfeed.tech/tags/detection.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [energy](<https://devfeed.tech/tags/energy.md>), [inference](<https://devfeed.tech/tags/inference.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [low-power](<https://devfeed.tech/tags/low-power.md>), [mission](<https://devfeed.tech/tags/mission.md>), [mobile](<https://devfeed.tech/tags/mobile.md>), [navigation](<https://devfeed.tech/tags/navigation.md>), [object-detection](<https://devfeed.tech/tags/object-detection.md>), [policies](<https://devfeed.tech/tags/policies.md>), [power](<https://devfeed.tech/tags/power.md>), [power-management](<https://devfeed.tech/tags/power-management.md>), [reliability](<https://devfeed.tech/tags/reliability.md>), [robot](<https://devfeed.tech/tags/robot.md>), [robotic](<https://devfeed.tech/tags/robotic.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [scheduling](<https://devfeed.tech/tags/scheduling.md>)

### AI overview

A robot-state-aware scheduler reduces open-vocabulary detection workload by switching among low-, mid-, and high-power YOLO-World-S modes using odometry, map changes, and detector confidence. On an AgileX LIMO across 40 navigation trials, it achieved 8/10 mission successes versus 7/10 for each fixed-rate baseline, with 5.2% lower mean Jetson input power and a 34% lower detection call rate than the 8 Hz baseline. It did not minimize Jetson input energy over full missions, since longer runtimes can offset power savings. In a separate 60-second stationary test, it reduced mean power from 8.045 W to 7.301 W and energy from 484.7 J to 438.1 J. The results suggest competitive task reliability alongside workload power savings, while mission energy should be considered with task progress and runtime.

### Source excerpt

arXiv:2611.07663v1 Announce Type: cross Abstract: Open-vocabulary object detection (OVD) lets robots search for run-time object queries, but fixed-rate inference spends embedded compute on low-value frames during idle motion, map revisits, or stable viewpoints. This paper presents an energy-aware OVD scheduler for low-power mobile robots. The scheduler uses odometry, map change, and detector confidence to switch among low-, mid-, and high-power YOLO-World-S modes on a Jetson Orin Nano. We evaluate the energy-quality trade-off on an AgileX LIMO using 40 real navigation trials: two target objects, five placements per target-policy pair, and four perception policies. Across bottle and cup trials, the scheduler records 8/10 mission success compared with 7/10 for each fixed-rate baseline, while reducing mean Jetson input power (a Jetson-side power-draw metric) by 5.2% relative to an 8 Hz high-rate baseline and lowering achieved OVD call rate by 34.0%. It does not minimize mission-window Jetson input energy because longer missions can erase instantaneous power savings. A separate 60 s stationary microbenchmark isolates the perception workload: the scheduler lowers mean power from 8.045 W to 7.301 W and energy from 484.7 J to 438.1 J relative to the 8 Hz baseline. These results support robot-state-aware OVD scheduling as a promising power-management mechanism with competitive task reliability, while showing that mission-window Jetson input energy must be read together with task progress and runtime.

## Communication-Efficient Model-Distributed Inference for CBRS Radar Detection at the Edge

DevFeed: [Communication-Efficient Model-Distributed Inference for CBRS Radar Detection at the Edge](<https://devfeed.tech/articles/communication-efficient-model-distributed-inference-for-cbrs-radar-detection-at-the-edge-113779.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.00625>)

Author: Teng Li, Hulya Seferoglu

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>)

Topics: [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [Data transmission](<https://devfeed.tech/topics/data-transmission.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [communications](<https://devfeed.tech/topics/communications.md>)

Tags: [5g](<https://devfeed.tech/tags/5g.md>), [cbrs](<https://devfeed.tech/tags/cbrs.md>), [communication](<https://devfeed.tech/tags/communication.md>), [compression](<https://devfeed.tech/tags/compression.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [data](<https://devfeed.tech/tags/data.md>), [detection](<https://devfeed.tech/tags/detection.md>), [device](<https://devfeed.tech/tags/device.md>), [distributed](<https://devfeed.tech/tags/distributed.md>), [distributed-inference](<https://devfeed.tech/tags/distributed-inference.md>), [edge](<https://devfeed.tech/tags/edge.md>), [eess-sp](<https://devfeed.tech/tags/eess-sp.md>), [i-q](<https://devfeed.tech/tags/i-q.md>), [inference](<https://devfeed.tech/tags/inference.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [quality](<https://devfeed.tech/tags/quality.md>), [radar](<https://devfeed.tech/tags/radar.md>), [radio](<https://devfeed.tech/tags/radio.md>), [raw](<https://devfeed.tech/tags/raw.md>)

### AI overview

CEDAR reduces communication costs in edge-distributed CBRS radar detection by redesigning both the detector and model partitioning. CEDAR-YOLO removes YOLO's multi-scale neck and two detection scales, using a single-scale path with independently configured frequency and time resolutions. CEDAR-MDI adds lightweight activation compressors and chooses each compressor's partition to minimize the pipeline bottleneck. On an NVIDIA Jetson Orin Nano with real-world over-the-air data, CEDAR-YOLO reduced parameters by 35-57% and inference time by 16-33%; CEDAR-MDI achieved up to 3.27x the throughput of single-device inference. The abstract reports preserved detection quality for both, but provides no further evaluation details.

### Source excerpt

arXiv:2611.00625v1 Announce Type: new Abstract: The Citizens Broadband Radio Service (CBRS) shares the 3550-3700 MHz band between incumbent radar systems and commercial LTE/5G users, so radar activity must be detected accurately and promptly. Spectrogram-based YOLO detectors can localize coexisting radar and cellular signals, but running them on resource-constrained edge devices is challenging. Model-distributed inference (MDI) can increase throughput by partitioning a model across edge devices. Using an end-to-end pipeline from raw I/Q samples to on-device detection, we show that partitioning recent YOLO generations incurs high activation-transfer costs due to their multi-scale structures. To reduce these costs, we propose CEDAR, a communication-efficient, edge-distributed architecture for radar detection. Its RF-aware detector, CEDAR-YOLO, removes YOLO's multi-scale neck and two of its three detection scales, leaving a single-scale path with independently configured frequency and time resolutions, which lowers the communication cost of partitioning. Its compression-aware MDI framework, CEDAR-MDI, integrates lightweight activation compressors and selects, for each compressor, the partition that minimizes the pipeline bottleneck. On an NVIDIA Jetson Orin Nano testbed with real-world over-the-air data, CEDAR-YOLO reduces the parameter count by 35-57% and model inference time by 16-33%, and CEDAR-MDI achieves up to $3.27\times$ the throughput of single-device inference, with detection quality preserved in both cases.

## MERIS: Adaptive Real-Time Scheduling of Refined Inference DAGs in Dynamic Robotic Environments

DevFeed: [MERIS: Adaptive Real-Time Scheduling of Refined Inference DAGs in Dynamic Robotic Environments](<https://devfeed.tech/articles/meris-adaptive-real-time-scheduling-of-refined-inference-dags-in-dynamic-robotic-environments-113468.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.24044>)

Author: Zexin Li, Tao Ren, Johnathan Liu, Xiaoxi He, Cong Liu

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [scheduling](<https://devfeed.tech/topics/scheduling.md>), [real-time](<https://devfeed.tech/topics/real-time.md>), [autonomous vehicles](<https://devfeed.tech/topics/autonomous-vehicles.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [admission](<https://devfeed.tech/tags/admission.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-driving](<https://devfeed.tech/tags/autonomous-driving.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [completion](<https://devfeed.tech/tags/completion.md>), [compute](<https://devfeed.tech/tags/compute.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [cs-sy](<https://devfeed.tech/tags/cs-sy.md>), [dag](<https://devfeed.tech/tags/dag.md>), [decoder](<https://devfeed.tech/tags/decoder.md>), [dnn](<https://devfeed.tech/tags/dnn.md>), [eess-sy](<https://devfeed.tech/tags/eess-sy.md>), [effectiveness](<https://devfeed.tech/tags/effectiveness.md>), [event-driven](<https://devfeed.tech/tags/event-driven.md>), [execution](<https://devfeed.tech/tags/execution.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [framework](<https://devfeed.tech/tags/framework.md>), [graph](<https://devfeed.tech/tags/graph.md>), [inference](<https://devfeed.tech/tags/inference.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [robotic](<https://devfeed.tech/tags/robotic.md>), [scheduling](<https://devfeed.tech/tags/scheduling.md>)

### AI overview

MERIS schedules changing inference workloads on resource-constrained robots by refining task graphs, assigning proportional intermediate deadlines, and synchronizing events. Its end-to-end deadline argument depends on profiled execution bounds, release assumptions, precedence barriers, and per-level capacity premises. For non-partitionable tasks, it uses a profile-bounded admission rule, with direct per-task completion checks needed to establish empirical timing evidence. Evaluation reports run-level QoE separately from the rate of offered nodes missing their deadlines, and describes a minibenchmark for DNN-driven autonomous driving schedulers. Experiments are reported on Nano Orin, AGX Orin, and M3 Max; the supplied excerpt gives no quantitative results.

### Source excerpt

arXiv:2605.24044v2 Announce Type: replace-cross Abstract: Robots deployed in dynamic environments execute changing directed-acyclic-graph (DAG) workloads under limited compute and timing budgets. This article presents MERIS, a framework for model scheduling under real-time constraints that combines proportional intermediate deadlines, MIMONet-aware graph refinement, and event-driven synchronization for multi-task inference on resource-constrained robotic platforms. We distinguish three representations: the application DAG, each MIMONet's internal encoder/decoder graph, and the refined scheduling DAG, and make the scope of the timing argument explicit. If the profiled execution bounds, release assumptions, precedence barriers, and per-level capacity premises hold, satisfaction of the assigned sub-deadlines composes to the end-to-end deadline. For non-partitionable tasks, we define a profile-bounded admission rule; before interpreting that rule as empirical timing evidence, we require direct per-task completion checks. We report run-level QoE alongside the offered-node deadline missing rate, making explicit that run completion and successful service of every node are different outcomes. We propose a comprehensive minibenchmark for evaluating the performance of schedulers under various DNN-driven autonomous driving scenarios. Extensive experiments on Nano Orin, AGX Orin, and M3 Max demonstrate the effectiveness of our proposed method.

## From Single-Shot QA to Continuous Perception: Revisiting the Energy Composition and Optimization Boundary of Edge VLMs

DevFeed: [From Single-Shot QA to Continuous Perception: Revisiting the Energy Composition and Optimization Boundary of Edge VLMs](<https://devfeed.tech/articles/from-single-shot-qa-to-continuous-perception-revisiting-the-energy-composition-and-optimization-boundary-of-edge-vlms-114968.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.00036>)

Author: Yiting Duan, Miaoxin Wang, Aoqing Guo, Bo Peng

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [vlm](<https://devfeed.tech/topics/vlm.md>), [Sustainable Computing](<https://devfeed.tech/topics/sustainable-computing.md>), [LLM optimization](<https://devfeed.tech/topics/llm-optimization.md>), [Jetson AGX Thor](<https://devfeed.tech/topics/jetson-agx-thor.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Image](<https://devfeed.tech/topics/image.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [composition](<https://devfeed.tech/tags/composition.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [decode](<https://devfeed.tech/tags/decode.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [edge](<https://devfeed.tech/tags/edge.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [encoding](<https://devfeed.tech/tags/encoding.md>), [energy](<https://devfeed.tech/tags/energy.md>), [image](<https://devfeed.tech/tags/image.md>), [independent](<https://devfeed.tech/tags/independent.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [qa](<https://devfeed.tech/tags/qa.md>), [vlms](<https://devfeed.tech/tags/vlms.md>)

### AI overview

Profiling nine edge VLMs across three architecture families on Jetson AGX Thor and Orin, the study finds that energy costs follow two distinct patterns: decode energy per output token rises with model scale, while visual-prefill energy depends mainly on visual-encoder efficiency. In continuous perception with N=10 and sparse output, visual prefill accounts for 92-97% of total energy, shifting the optimization focus toward visual frame reuse, adaptive output truncation, task-aware routing and encoder throughput. The findings are based on 4,508 profiling records calibrated with a pure-text task.

### Source excerpt

arXiv:2611.00036v1 Announce Type: cross Abstract: Visual-prefill energy conclusions for edge vision-language models (VLMs) are commonly drawn under single-shot workloads; we revisit them under continuous perception. Profiling nine VLMs across three architecture families on Jetson AGX Thor and Orin (4,508 records, calibrated with a pure-text task), we find two mutually independent energy laws: the per-output-token decode cost grows monotonically with model scale, whereas the visual-prefill (image-encoding) cost is dominated by visual-encoder architectural efficiency, not scale. Under continuous perception ($N{=}10$) with sparse output, visual-prefill energy reaches 92--97\% of total. We derive optimization strategies---visual frame reuse, adaptive output truncation, task-aware routing, and encoder throughput optimization---for edge VLM deployment.

## ORBITAL: Optimization of Robust Brain--Computer Inference Through Architecture-Aware Low-Precision Compression

DevFeed: [ORBITAL: Optimization of Robust Brain--Computer Inference Through Architecture-Aware Low-Precision Compression](<https://devfeed.tech/articles/orbital-optimization-of-robust-brain-computer-inference-through-architecture-aware-low-precision-compression-103352.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.36411>)

Author: Saim Rehman, Muhammad Shafique

Published: 2026-10-28T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [bci](<https://devfeed.tech/topics/bci.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [quantization](<https://devfeed.tech/topics/quantization.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [tinyml](<https://devfeed.tech/topics/tinyml.md>), [On-device AI](<https://devfeed.tech/topics/on-device-ai.md>)

Tags: [adversarial](<https://devfeed.tech/tags/adversarial.md>), [bci](<https://devfeed.tech/tags/bci.md>), [brain](<https://devfeed.tech/tags/brain.md>), [compression](<https://devfeed.tech/tags/compression.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [eeg](<https://devfeed.tech/tags/eeg.md>), [eess-sp](<https://devfeed.tech/tags/eess-sp.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [energy](<https://devfeed.tech/tags/energy.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [inference](<https://devfeed.tech/tags/inference.md>), [int8](<https://devfeed.tech/tags/int8.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [latency](<https://devfeed.tech/tags/latency.md>), [models](<https://devfeed.tech/tags/models.md>), [native](<https://devfeed.tech/tags/native.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [robust](<https://devfeed.tech/tags/robust.md>), [trades](<https://devfeed.tech/tags/trades.md>), [union](<https://devfeed.tech/tags/union.md>), [validation](<https://devfeed.tech/tags/validation.md>)

### AI overview

ORBITAL combines structured compression, recovery-aware operating-point selection, multi-attack testing, and edge-device validation to balance EEG decoder robustness with hardware costs. Across nine subjects and three seeds on BCI Competition IV-2a, S25-TRADES had the highest observed mean true-union robustness among compressed EEGNet models (23.06%, versus 21.27% for standard Full) while using 26.27% fewer parameters and 25.10% fewer MACs. Matched controls attribute much of the robustness recovery to TRADES: S25 union accuracy improved 6.74 percentage points over clean recovery (p=0.0039); per-example analysis suggests boundary redistribution rather than uniform margin expansion. On Jetson Orin Nano, FP16 compression cut board energy by up to 16.20% and latency by 8.00%. Native INT8 Q/DQ had 99.3% prediction agreement but about twice FP16's latency and energy, showing that prediction agreement alone did not ensure hardware efficiency.

### Source excerpt

arXiv:2610.36411v1 Announce Type: cross Abstract: Compact electroencephalography (EEG) decoders enable embedded brain--computer interfaces (BCIs), but compression can alter adversarial robustness without proportional hardware savings. We present ORBITAL, a robustness-constrained framework combining hardware-realizable structured compression, recovery-aware operating-point selection, multi-attack evaluation, and native edge validation. Across nine subjects and three seeds on BCI Competition IV-2a, S25-TRADES achieves the highest observed mean true-union robustness among compressed EEGNet models (23.06% vs 21.27% for standard Full), with 26.27% fewer parameters and 25.10% fewer MACs. Matched controls identify TRADES as the main source of robustness recovery, improving S25 union accuracy by 6.74 percentage points over clean recovery (p=0.0039), while per-example analysis reveals boundary redistribution rather than uniform margin expansion. On Jetson Orin Nano, FP16 compression reduces board energy by up to 16.20% and latency by 8.00%. Despite 99.3% prediction agreement, native INT8 Q/DQ incurs about twice the FP16 latency and energy, demonstrating the need for joint robustness and hardware-aware optimization.

## CrowdCC: Jointly Benchmarking Corruption Robustness and Deployment Quantization in Crowd Counting

DevFeed: [CrowdCC: Jointly Benchmarking Corruption Robustness and Deployment Quantization in Crowd Counting](<https://devfeed.tech/articles/crowdcc-jointly-benchmarking-corruption-robustness-and-deployment-quantization-in-crowd-counting-103119.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.36340>)

Author: S Mahmudul Hasan, Lu Peng

Published: 2026-10-28T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [quantization](<https://devfeed.tech/topics/quantization.md>), [Post-training optimization](<https://devfeed.tech/topics/post-training-optimization.md>), [performance-engineering](<https://devfeed.tech/topics/performance-engineering.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [TensorRT](<https://devfeed.tech/topics/tensorrt.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>)

Tags: [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [inference-efficiency](<https://devfeed.tech/tags/inference-efficiency.md>), [measurements](<https://devfeed.tech/tags/measurements.md>), [precision](<https://devfeed.tech/tags/precision.md>), [quantization](<https://devfeed.tech/tags/quantization.md>), [robustness](<https://devfeed.tech/tags/robustness.md>)

### AI overview

CrowdCC jointly benchmarks corruption robustness and deployment quantization for crowd counting. It evaluates 26 models across four datasets and 23 corruptions at three severity levels, using TensorRT engines at FP32, FP16, and post-training INT8, with throughput and energy measurements on Jetson Orin NX configurations. Capture-chain corruptions caused the largest degradation, and clean accuracy did not fully predict corrupted performance. INT8 changed robustness more than FP16, without a consistent direction, and did not reliably improve latency or energy; reformat overhead and limited INT8 coverage were associated with inefficiency or slowdowns. The results show that precision can change error rankings and robustness-cost tradeoffs, so model selection should account for both robustness and measured deployment cost.

### Source excerpt

arXiv:2610.36340v1 Announce Type: cross Abstract: Crowd-counting models deployed on edge devices face two shifts: inputs are degraded along the capture and transmission pipeline, while models are quantized to satisfy resource constraints. Yet corruption robustness and deployment quantization are typically evaluated separately. We introduce CrowdCC, a controlled corruption benchmark that jointly evaluates these effects using 23 deployment-motivated corruptions at three severity levels, four crowd-counting datasets, and 26 models using A100 TensorRT engines at FP32, FP16, and post-training INT8 precision, paired with throughput and energy measurements from corresponding Jetson Orin NX configurations. The corruptions span six families, with head annotations adjusted when geometry or visibility changes. Our evaluation shows that capture-chain corruptions consistently produce the largest degradation across the evaluated models, while crowd density affects absolute and relative error differently. Clean accuracy does not fully predict corrupted performance. INT8 alters corruption robustness substantially more than FP16, with no systematic direction, and does not consistently improve latency or energy. INT8 inference efficiency is strongly associated with reformat overhead in the compiled engine, while slowdowns are also observed under limited INT8 coverage. Precision changes corrupted-error rankings and robustness-cost frontiers, showing why model selection must consider corruption robustness and measured deployment cost together. The code is available at https://github.com/Tulane-Computer-Architecture-Systems/CrowdCC-Release.

## Autonomous excavator startup Bedrock Robotics picks up $600M in funding

DevFeed: [Autonomous excavator startup Bedrock Robotics picks up $600M in funding](<https://devfeed.tech/articles/autonomous-excavator-startup-bedrock-robotics-picks-up-600m-in-funding-100936.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/10/26/autonomous-excavator-startup-bedrock-robotics-picks-up-600m-in-funding/>)

Author: Maria Deutscher

Published: 2026-10-26T23:48:33Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [autonomous vehicles](<https://devfeed.tech/topics/autonomous-vehicles.md>), [Robotics](<https://devfeed.tech/topics/robotics.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Physical AI](<https://devfeed.tech/topics/physical-ai.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-vehicle](<https://devfeed.tech/tags/autonomous-vehicle.md>), [bedrock](<https://devfeed.tech/tags/bedrock.md>), [bedrock-robotics](<https://devfeed.tech/tags/bedrock-robotics.md>), [emerging-tech](<https://devfeed.tech/tags/emerging-tech.md>), [excavator](<https://devfeed.tech/tags/excavator.md>), [funding](<https://devfeed.tech/tags/funding.md>), [funding-round](<https://devfeed.tech/tags/funding-round.md>), [investment](<https://devfeed.tech/tags/investment.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [startup](<https://devfeed.tech/tags/startup.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>)

### AI overview

Bedrock Robotics raised $600 million in a Series C led by CapitalG and Emergence Capital, with Nvidia, Bessemer and other investors participating; the company says it is valued at nearly $1 billion. Its Bedrock Operator combines an in-cab computing appliance with a roof-mounted array of eight cameras and lidar to help excavators navigate and perform earthmoving with minimal human input. An inertial measurement unit helps correct GPS errors that can arise from obstructions or uneven terrain. Bedrock says it has begun paid work on commercial construction sites and will use the funding to expand adoption and field service support, with plans to eventually cover other construction vehicles.

### Source excerpt

Bedrock Robotics Inc. today announced that it has closed a $600 million investment led by Alphabet Inc.'s CapitalG fund and Emergence Capital. The Series C round also drew contributions from Nvidia Corp., Bessemer and over a half-dozen others. Bedrock says that it's now valued at nearly $1 billion. The cash infusion comes a few weeks [...] The post Autonomous excavator startup Bedrock Robotics picks up $600M in funding appeared first on SiliconANGLE.

## Yocto on Jetson: Build and Flash on Seeed Carrier Boards

DevFeed: [Yocto on Jetson: Build and Flash on Seeed Carrier Boards](<https://devfeed.tech/articles/yocto-on-jetson-build-and-flash-on-seeed-carrier-boards-91862.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/10/21/yocto-on-jetson-build-and-flash-on-seeed-carrier-boards/>)

Author: Frank

Published: 2026-10-21T07:17:26Z

Content type: tutorial

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [seeed studio](<https://devfeed.tech/topics/seeed-studio.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Build tool](<https://devfeed.tech/topics/build-tool.md>), [Test automation](<https://devfeed.tech/topics/test-automation.md>), [software bill of materials](<https://devfeed.tech/topics/software-bill-of-materials.md>)

Tags: [acceptance](<https://devfeed.tech/tags/acceptance.md>), [application](<https://devfeed.tech/tags/application.md>), [archive](<https://devfeed.tech/tags/archive.md>), [board](<https://devfeed.tech/tags/board.md>), [boot](<https://devfeed.tech/tags/boot.md>), [bsp](<https://devfeed.tech/tags/bsp.md>), [build](<https://devfeed.tech/tags/build.md>), [build-tools](<https://devfeed.tech/tags/build-tools.md>), [building](<https://devfeed.tech/tags/building.md>), [but](<https://devfeed.tech/tags/but.md>), [camera](<https://devfeed.tech/tags/camera.md>), [carrier](<https://devfeed.tech/tags/carrier.md>), [carrier-board](<https://devfeed.tech/tags/carrier-board.md>), [clone](<https://devfeed.tech/tags/clone.md>), [command](<https://devfeed.tech/tags/command.md>), [embedded-linux](<https://devfeed.tech/tags/embedded-linux.md>), [flash](<https://devfeed.tech/tags/flash.md>), [image](<https://devfeed.tech/tags/image.md>), [jetpack](<https://devfeed.tech/tags/jetpack.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [package](<https://devfeed.tech/tags/package.md>), [recomputer](<https://devfeed.tech/tags/recomputer.md>), [tech](<https://devfeed.tech/tags/tech.md>), [yocto](<https://devfeed.tech/tags/yocto.md>)

### AI overview

For reproducible Yocto builds on Jetson, keep the carrier, module SKU, source revision, image recipe, flash artifact, and target results tied together. The example uses Seeed's reComputer Super J401 with OE4T wrynose and meta-tegra for Jetson Linux R39.2.0 / JetPack 7.2; verify the current Wiki for other releases. The development image includes target-side nvcc and build tools, while the x86_64 cross-development SDK is a separate output. The workflow records inputs, builds and checksums a tegraflash archive, prepares flashing with the same recipe, then tests the interfaces the application uses. The support matrix's physical validation is limited, so treat untested interfaces as project work. The article reports no build, flash, or hardware test results.

### Source excerpt

A completed image build leaves several questions open: does the flash package match the installed The post Yocto on Jetson: Build and Flash on Seeed Carrier Boards appeared first on Latest News from Seeed Studio.

## Mobile Robot Full-stack Development In Practice Ⅰ: Build the Robot Brain First -- Jetson, ROS 2, and the Development Foundation

DevFeed: [Mobile Robot Full-stack Development In Practice Ⅰ: Build the Robot Brain First -- Jetson, ROS 2, and the Development Foundation](<https://devfeed.tech/articles/mobile-robot-full-stack-development-in-practice-i-build-the-robot-brain-first-jetson-ros-2-and-the-development-foundation-62481.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/09/30/jetson-powered-mobile-robot-full-stack-development-in-practice-module-1-jetson-ros2/>)

Author: Frank

Published: 2026-09-30T12:59:53Z

Content type: tutorial

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [Isaac](<https://devfeed.tech/topics/isaac.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Learning](<https://devfeed.tech/topics/learning.md>), [remote-development](<https://devfeed.tech/topics/remote-development.md>), [container native](<https://devfeed.tech/topics/container-native.md>), [performance-engineering](<https://devfeed.tech/topics/performance-engineering.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-development](<https://devfeed.tech/tags/ai-development.md>), [ai-robotics](<https://devfeed.tech/tags/ai-robotics.md>), [algorithms](<https://devfeed.tech/tags/algorithms.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [feature](<https://devfeed.tech/tags/feature.md>), [full-stack](<https://devfeed.tech/tags/full-stack.md>), [full-stack-development](<https://devfeed.tech/tags/full-stack-development.md>), [isaac-ros](<https://devfeed.tech/tags/isaac-ros.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [mobile-robot](<https://devfeed.tech/tags/mobile-robot.md>), [mobile-robotics](<https://devfeed.tech/tags/mobile-robotics.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia-jetson-agx-orin](<https://devfeed.tech/tags/nvidia-jetson-agx-orin.md>), [open-source-robotics](<https://devfeed.tech/tags/open-source-robotics.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [recomputer-robotics-j501](<https://devfeed.tech/tags/recomputer-robotics-j501.md>), [robot](<https://devfeed.tech/tags/robot.md>), [robot-development](<https://devfeed.tech/tags/robot-development.md>), [ros-2](<https://devfeed.tech/tags/ros-2.md>)

### AI overview

The article's practical takeaway is to build and verify the robot's system foundation in layers before debugging applications: confirm hardware and Linux device visibility, validate the JetPack acceleration stack, establish a reproducible development workflow, then test ROS 2 communication and GPU access in Docker. Using a Jetson AGX Orin J5011 with an Ubuntu host as its reference setup, it shows remote development, isolated Python environments, monitoring, ROS 2 nodes for robot status and commands across two machines, and an Isaac ROS container check. It highlights platform-specific limits: CUDA installed does not ensure OpenCV uses CUDA, Jetson PyTorch builds must match JetPack/CUDA, and power, heat, and sustained performance matter alongside peak throughput. For failures, it recommends checking hardware and drivers before the development environment, middleware, and application.

### Source excerpt

If this is your first time coming across the series, Mobile Robot Full-stack Development In The post Mobile Robot Full-stack Development In Practice Ⅰ: Build the Robot Brain First -- Jetson, ROS 2, and the Development Foundation appeared first on Latest News from Seeed Studio.

## NOCL Making Progress As Mesa Gallium/OpenCL Atop NVIDIA CUDA

DevFeed: [NOCL Making Progress As Mesa Gallium/OpenCL Atop NVIDIA CUDA](<https://devfeed.tech/articles/nocl-making-progress-as-mesa-gallium-opencl-atop-nvidia-cuda-61946.md>)

Original publisher: [Read original article](<https://www.phoronix.com/news/NOCL-OpenCL-CUDA>)

Author: Michael Larabel

Published: 2026-09-29T13:53:05Z

Content type: news

Language: en

Sources: [Phoronix](<https://devfeed.tech/sources/phoronix.md>)

Topics: [OpenCL](<https://devfeed.tech/topics/opencl.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>), [NVLink](<https://devfeed.tech/topics/nvlink.md>)

Tags: [compute](<https://devfeed.tech/tags/compute.md>), [desktop-linux](<https://devfeed.tech/tags/desktop-linux.md>), [gallium](<https://devfeed.tech/tags/gallium.md>), [linux](<https://devfeed.tech/tags/linux.md>), [linux-benchmarking](<https://devfeed.tech/tags/linux-benchmarking.md>), [linux-hardware-benchmarks](<https://devfeed.tech/tags/linux-hardware-benchmarks.md>), [linux-hardware-reviews](<https://devfeed.tech/tags/linux-hardware-reviews.md>), [linux-how-to](<https://devfeed.tech/tags/linux-how-to.md>), [linux-performance](<https://devfeed.tech/tags/linux-performance.md>), [linux-server-benchmarks](<https://devfeed.tech/tags/linux-server-benchmarks.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-source-graphics](<https://devfeed.tech/tags/open-source-graphics.md>), [opencl](<https://devfeed.tech/tags/opencl.md>), [performance](<https://devfeed.tech/tags/performance.md>), [phoronix](<https://devfeed.tech/tags/phoronix.md>), [phoronix-test-suite](<https://devfeed.tech/tags/phoronix-test-suite.md>), [ptx](<https://devfeed.tech/tags/ptx.md>), [ubuntu-benchmarks](<https://devfeed.tech/tags/ubuntu-benchmarks.md>), [ubuntu-hardware](<https://devfeed.tech/tags/ubuntu-hardware.md>)

### AI overview

NOCL is a developing Mesa Gallium driver that uses NVIDIA's proprietary CUDA driver and PTX representation to provide an open-source OpenCL option on Linux. Its merge request is pending; it supports more OpenCL extensions than NVIDIA's proprietary OpenCL driver and may aid performance comparisons and PTX development. PTX-to-NIR support could also matter for DLSS with the NVK Mesa driver.

### Source excerpt

In development the past year has been the Mesa Gallium3D API implemented atop NVIDIA's CUDA driver API. Initially named "CLUDA" and now going by "NOCL", this Mesa code paired with Rusticl is working well for OpenCL atop the NVIDIA proprietary driver as a new compute option...

## A comparative study of YOLOv8 drone detection on Raspberry Pi 5 and Jetson Orin

DevFeed: [A comparative study of YOLOv8 drone detection on Raspberry Pi 5 and Jetson Orin](<https://devfeed.tech/articles/can-real-time-drone-detection-run-on-the-edge-60580.md>)

Original publisher: [Read original article](<https://hackernoon.com/can-real-time-drone-detection-run-on-the-edge?source=rss>)

Author: vishwa vimukthi gammuduwaththage

Published: 2026-09-27T14:22:25Z

Content type: article

Language: en

Sources: [HackerNoon](<https://devfeed.tech/sources/hackernoon.md>)

Topics: [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Holoscan](<https://devfeed.tech/topics/holoscan.md>)

Tags: [computer-vision](<https://devfeed.tech/tags/computer-vision.md>), [drone-technology](<https://devfeed.tech/tags/drone-technology.md>), [edge](<https://devfeed.tech/tags/edge.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [inference](<https://devfeed.tech/tags/inference.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nano](<https://devfeed.tech/tags/jetson-orin-nano.md>), [jetson-orin-nx](<https://devfeed.tech/tags/jetson-orin-nx.md>), [model-quantization](<https://devfeed.tech/tags/model-quantization.md>), [yolo-object-detection](<https://devfeed.tech/tags/yolo-object-detection.md>), [yolov8](<https://devfeed.tech/tags/yolov8.md>)

### AI overview

The article discusses a study comparing a real-time UAV image-processing pipeline across edge devices, including Raspberry Pi 5 and Jetson Orin. It highlights the trade-offs among inference speed, energy use, and real-time processing for drone detection.

### Source excerpt

A 2025 study compares YOLOv8 across Raspberry Pi 5 and Jetson Orin hardware, revealing the trade-offs behind real-time drone vision.

## US Developers: You Can Now Shop Seeed Studio Products at Micro Center

DevFeed: [US Developers: You Can Now Shop Seeed Studio Products at Micro Center](<https://devfeed.tech/articles/us-developers-you-can-now-shop-seeed-studio-products-at-micro-center-59235.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/09/24/seeed-studio-at-micro-center/>)

Author: Kezang Loday

Published: 2026-09-24T09:02:22Z

Content type: release

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [Jetson](<https://devfeed.tech/topics/jetson.md>), [Internet of things](<https://devfeed.tech/topics/iot.md>), [ESP32](<https://devfeed.tech/topics/esp32.md>), [lora](<https://devfeed.tech/topics/lora.md>), [meshtastic](<https://devfeed.tech/topics/meshtastic.md>), [Robotics](<https://devfeed.tech/topics/robotics.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [voice ai](<https://devfeed.tech/topics/voice-ai.md>)

Tags: [ai-robotics](<https://devfeed.tech/tags/ai-robotics.md>), [availability](<https://devfeed.tech/tags/availability.md>), [community](<https://devfeed.tech/tags/community.md>), [developers](<https://devfeed.tech/tags/developers.md>), [esp32](<https://devfeed.tech/tags/esp32.md>), [feature](<https://devfeed.tech/tags/feature.md>), [iot](<https://devfeed.tech/tags/iot.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [let](<https://devfeed.tech/tags/let.md>), [lora](<https://devfeed.tech/tags/lora.md>), [maker](<https://devfeed.tech/tags/maker.md>), [mesh-networking](<https://devfeed.tech/tags/mesh-networking.md>), [meshtastic](<https://devfeed.tech/tags/meshtastic.md>), [micro-center](<https://devfeed.tech/tags/micro-center.md>), [news](<https://devfeed.tech/tags/news.md>), [products](<https://devfeed.tech/tags/products.md>), [rebot](<https://devfeed.tech/tags/rebot.md>), [recamera](<https://devfeed.tech/tags/recamera.md>), [recomputer](<https://devfeed.tech/tags/recomputer.md>), [respeaker](<https://devfeed.tech/tags/respeaker.md>), [retail](<https://devfeed.tech/tags/retail.md>), [reterminal-epaper](<https://devfeed.tech/tags/reterminal-epaper.md>), [seeed-studio-xiao](<https://devfeed.tech/tags/seeed-studio-xiao.md>), [sensecap](<https://devfeed.tech/tags/sensecap.md>), [us](<https://devfeed.tech/tags/us.md>), [xiao](<https://devfeed.tech/tags/xiao.md>)

### AI overview

Seeed Studio announces that its development boards, ePaper displays, voice-AI microphones, vision-AI cameras, Meshtastic devices, robotics hardware, and Jetson-based edge-AI computers are now available at Micro Center stores and online in the United States.

### Source excerpt

Good news for makers and developers in the US -- Seeed Studio products are now The post US Developers: You Can Now Shop Seeed Studio Products at Micro Center appeared first on Latest News from Seeed Studio.

## Edge Computing, The Rise of Decentralized Models, and More

DevFeed: [Edge Computing, The Rise of Decentralized Models, and More](<https://devfeed.tech/articles/edge-computing-the-rise-of-decentralized-models-and-more-58362.md>)

Original publisher: [Read original article](<https://hackernoon.com/edge-computing-the-rise-of-decentralized-models-and-more?source=rss>)

Author: ZyVOP

Published: 2026-09-22T20:30:23Z

Content type: article

Language: en

Sources: [HackerNoon](<https://devfeed.tech/sources/hackernoon.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Computing](<https://devfeed.tech/topics/computing.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [On-device AI](<https://devfeed.tech/topics/on-device-ai.md>), [model-deployment](<https://devfeed.tech/topics/model-deployment.md>), [Latency](<https://devfeed.tech/topics/latency.md>), [Linux](<https://devfeed.tech/topics/linux.md>), [Jetson Orin](<https://devfeed.tech/topics/jetson-orin.md>), [Cloud](<https://devfeed.tech/topics/cloud.md>), [Chip design](<https://devfeed.tech/topics/chip-design.md>), [Robotics](<https://devfeed.tech/topics/robotics.md>), [Internet of things](<https://devfeed.tech/topics/iot.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-fractures](<https://devfeed.tech/tags/ai-fractures.md>), [algorithms](<https://devfeed.tech/tags/algorithms.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [automation](<https://devfeed.tech/tags/automation.md>), [cloud-computing](<https://devfeed.tech/tags/cloud-computing.md>), [decentralized-models](<https://devfeed.tech/tags/decentralized-models.md>), [developers](<https://devfeed.tech/tags/developers.md>), [edge-computing](<https://devfeed.tech/tags/edge-computing.md>), [export-controls](<https://devfeed.tech/tags/export-controls.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [inference](<https://devfeed.tech/tags/inference.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [latency](<https://devfeed.tech/tags/latency.md>), [linux](<https://devfeed.tech/tags/linux.md>), [reliability](<https://devfeed.tech/tags/reliability.md>)

### AI overview

Regulatory restrictions, export controls, and reliability concerns are pushing AI inference from centralized clouds toward locally controlled edge, on-premises, and regional systems. The article highlights speculative decoding, deterministic routing, Linux-based KV-cache trimming, and AI-designed chips as technologies supporting this shift.

### Source excerpt

Regulatory bans, export controls, and reliability failures are forcing AI developers to move inference to the edge.

[Next page](<https://devfeed.tech/topics/jetson-orin.md?cursor=WyIyMDI2LTA5LTIyVDIwOjMwOjIzKzAwOjAwIiwgIjI1YzY4YTFhLWM1NjItNDZhZC05NDUxLWM5NDA2ZWI3MjBhYiJd>)