Multi-Agent Reachability Calibration with Conformal Prediction
Authors:
- 13 December 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 6596-6603
Abstract
We investigate methods to provide safety assurances for autonomous agents that incorporate learning-based predictions of other, uncontrolled agents' behavior into their own trajectory planning. Given a learning-based forecasting model that predicts agents' trajectories, we introduce a method for providing probabilistic assurances on the model's prediction error with calibrated confidence intervals. Through quantile regression, conformal prediction, and reachability analysis, our method generates probabilistically safe and dynamically feasible prediction sets. We showcase their utility in certifying the safety of planning algorithms, both in simulations using actual autonomous driving data and in an experiment with Boeing vehicles.Keywords
All Related Versions
This publication has 19 references indexed in Scilit:
- Sample-Efficient Safety Assurances Using Conformal Prediction2022
- Data-Driven Reachability Analysis with Christoffel Functions2021
- Large Scale Interactive Motion Forecasting for Autonomous Driving : The Waymo Open Motion Dataset2021
- Assurance monitoring of learning-enabled cyber-physical systems using inductive conformal prediction based on distance learningArtificial Intelligence for Engineering Design, Analysis and Manufacturing, 2021
- Trajectron++: Dynamically-Feasible Trajectory Forecasting with Heterogeneous Data2020
- nuScenes: A Multimodal Dataset for Autonomous Driving2020
- Neural Network-Based Uncertainty Quantification: A Survey of Methodologies and ApplicationsIEEE Access, 2018
- Hamilton-Jacobi reachability: A brief overview and recent advances2017
- Quantile Regression2005
- Algorithmic Learning in a Random World2005