Occluding the Solution Space: Planner-Agnostic Adversarial Attacks on Tolerance-Aware Manipulation
For robotic manipulation systems, this work provides a more practical and efficient method to evaluate robustness against adversarial obstacles without requiring access to the planner.
The paper proposes a planner-agnostic adversarial attack framework for tolerance-aware manipulation that uses a kinematic occupancy heatmap to characterize workspace capabilities and formulates the attack as a budgeted maximum-coverage optimization. Experiments show it outperforms planner-in-the-loop baselines in efficiency and efficacy.
Adversarial attacks on motion planning are crucial for evaluating and quantifying the intrinsic robustness of robotic manipulation. However, existing approaches are typically limited by restrictive exact-pose objectives and their reliance on planner-in-the-loop queries. To address these limitations, we propose a planner-agnostic attack framework for tolerance-aware manipulation. Our approach shifts the evaluation paradigm to task-level feasibility over goal regions, efficiently inserting adversarial obstacles without requiring oracle access to the victim system. Offline, we characterize the robot's intrinsic workspace capabilities via a kinematic occupancy heatmap, which encodes the density of feasible trajectories and robustness priors without invoking a specific planner. Online, we formulate the attack as a budgeted maximum-coverage optimization, strategically deploying obstacles subject to explicit geometric constraints to occlude the solution space. Extensive experiments across simulation and real-world scenarios demonstrate that our method reliably induces planning failures, significantly outperforming planner-in-the-loop baselines in both computational efficiency and attack efficacy.