← Back to Benchmarks
simmediumquadrupedmetric · varies

Safety-Aware Imitation Learning via MPC-Guided Disturbance Injection

Description

Imitation Learning has provided a promising approach to learning complex robot behaviors from expert demonstrations. However, learned policies can make errors that lead to safety violations, which limits their deployment in safety-critical applications. We propose MPC-SafeGIL, a design-time approach that enhances the safety of imitation learning by injecting adversarial disturbances during expert demonstrations. This exposes the expert to a broader range of safety-critical scenarios and allows t

Source

http://arxiv.org/abs/2508.03129v1