RAP: Risk-Aware Prediction for Robust Planning.

Consider an autonomous vehicle planning to drive ahead along the arrow. In this example, the biker has 95% chance to turn right and 5% chance to cross the road. Should 95% of the samples from a trajectory forecasting model point towards turning right? How to estimate the risk for the autonomous vehicle? How to capture the dangerous events?


Robust planning in interactive scenarios requires predicting the uncertain future to make risk-aware decisions. Unfortunately, due to long-tail safety-critical events, the risk is often under-estimated by finite-sampling approximations of probabilistic motion forecasts. This can lead to overconfident and unsafe robot behavior, even with robust planners. Instead of assuming full prediction coverage that robust planners require, we propose to make prediction itself risk-aware. We introduce a new prediction objective to learn a risk-biased distribution over trajectories, so that risk evaluation simplifies to an expected cost estimation under this biased distribution. This reduces the sample complexity of the risk estimation during online planning, which is needed for safe real-time performance. Evaluation results in a didactic simulation environment and on a real-world dataset demonstrate the effectiveness of our approach. The code and a demo are available.

In Conference on Robot Learning.