Testing Rare Downstream Safety Violations via Upstream Adaptive Sampling of Perception Error Models
This work addresses safety testing for autonomous systems, offering a method to efficiently simulate rare failures, though it is incremental as it builds on existing sampling techniques.
The paper tackles the problem of efficiently testing rare safety violations in black-box perceptual-control systems by combining perception error models with adaptive importance sampling, achieving accurate failure probability estimates with a reduced number of simulations in an autonomous braking system.
Testing black-box perceptual-control systems in simulation faces two difficulties. Firstly, perceptual inputs in simulation lack the fidelity of real-world sensor inputs. Secondly, for a reasonably accurate perception system, encountering a rare failure trajectory may require running infeasibly many simulations. This paper combines perception error models -- surrogates for a sensor-based detection system -- with state-dependent adaptive importance sampling. This allows us to efficiently assess the rare failure probabilities for real-world perceptual control systems within simulation. Our experiments with an autonomous braking system equipped with an RGB obstacle-detector show that our method can calculate accurate failure probabilities with an inexpensive number of simulations. Further, we show how choice of safety metric can influence the process of learning proposal distributions capable of reliably sampling high-probability failures.