# Explore Where It Matters: Targeted Exploration for Efficient Rover Adaptation

DevFeed: [Explore Where It Matters: Targeted Exploration for Efficient Rover Adaptation](<https://devfeed.tech/articles/explore-where-it-matters-targeted-exploration-for-efficient-rover-adaptation-108106.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.38657>)

Author: Jacob Levy, Arthur Bouton, Tristan D. Hassler, Hari Nayar, David Fridovich-Keil

Published: 2026-10-30T04:00:00Z

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: [World models](<https://devfeed.tech/topics/world-models.md>)

Tags: [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [errors](<https://devfeed.tech/tags/errors.md>), [exploration](<https://devfeed.tech/tags/exploration.md>), [mobility](<https://devfeed.tech/tags/mobility.md>), [performance](<https://devfeed.tech/tags/performance.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [reward](<https://devfeed.tech/tags/reward.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [simulation](<https://devfeed.tech/tags/simulation.md>)

## AI overview

TERRA adapts a simulation-trained rover world model by weighting prediction errors according to task reward, while a learned loss predictor identifies where exploration may reveal future performance-relevant errors. It uses that signal to guide exploration during model predictive control, aiming to reduce deployment-time trial and error. In simulation under severe out-of-distribution pretraining, it achieved about 90% slope-climbing success within five trials and outperformed exploration baselines on reward, completion time, and energy to first success. On the ERNEST rover, it recovered coordinated active-suspension locomotion on a deformable Mars-simulant slope in three to five real-world trials.

## Source excerpt

arXiv:2610.38657v1 Announce Type: cross Abstract: Future planetary rovers will traverse steep, deformable terrain where slip, sinkage, and terrain deformation can threaten mobility. Although simulation can capture some wheel-soil effects, unmodeled deformable-terrain phenomena create sim-to-real gaps requiring deployment-time adaptation. Yet, rover operations leave little room for extensive trial-and-error. We introduce TERRA (Targeted Exploration for Reward-Relevant Adaptation), an online adaptation framework that efficiently seeks deployment data exposing task-relevant model errors and corrects them. TERRA adapts a simulation-trained world model with a reward-weighted prediction loss and trains a loss predictor to estimate where future performance-relevant errors are likely to occur, using this signal to guide exploration during MPC. In simulation, TERRA reaches approximately 90% slope-climbing success within five trials under severe out-of-distribution pretraining conditions, outperforming exploration baselines in reward, completion time, and energy to first success. On the ERNEST rover, TERRA recovers coordinated active-suspension locomotion on a deformable Mars-simulant slope within three to five real-world trials. Project website: https://terra-explore.github.io