Modeling Preemptive Behaviors for Uncommon Hazardous Situations From Demonstrations

TL;DR

This paper introduces a learning from demonstration method to program safe autonomous driving behaviors in uncommon hazardous scenarios, using simulation and a novel linear combination approach to generalize behaviors across variations.

Contribution

It presents a model-agnostic linear combination technique for extending demonstrated behaviors to new, complex hazardous situations in autonomous driving.

Findings

Behavior generalizes well to multiple hazard variations

Decision-making balances road rules and immediate rewards

Simulation-based demonstration effectively captures rare scenarios

Abstract

This paper presents a learning from demonstration approach to programming safe, autonomous behaviors for uncommon driving scenarios. Simulation is used to re-create a targeted driving situation, one containing a road-side hazard creating a significant occlusion in an urban neighborhood, and collect optimal driving behaviors from 24 users. Paper employs a key-frame based approach combined with an algorithm to linearly combine models in order to extend the behavior to novel variations of the target situation. This approach is theoretically agnostic to the kind of LfD framework used for modeling data and our results suggest it generalizes well to variations containing an additional number of hazards occurring in sequence. The linear combination algorithm is informed by analysis of driving data, which also suggests that decision-making algorithms need to consider a trade-off between road-rules and immediate rewards to tackle some complex cases.