Fast and Accurate Collision Probability Estimation for Autonomous Vehicles using Adaptive Sigma-Point Sampling

TL;DR

This paper introduces an adaptive sigma-point sampling algorithm for fast, accurate collision probability estimation in autonomous vehicles, explicitly considering temporal dependence to improve overestimations.

Contribution

The paper presents a novel adaptive sigma-point sampling method that efficiently estimates collision probabilities with high accuracy and accounts for temporal dependence, outperforming prior approaches.

Findings

Median error of 3.5% in collision probability estimates

Median runtime of 0.21ms per estimate

Validated on 400 real-world autonomous vehicle scenarios

Abstract

A novel algorithm is presented for the estimation of collision probabilities between dynamic objects with uncertain trajectories, where the trajectories are given as a sequence of poses with Gaussian distributions. We propose an adaptive sigma-point sampling scheme, which ultimately produces a fast, simple algorithm capable of estimating the collision probability with a median error of 3.5%, and a median runtime of 0.21ms, when measured on an Intel Xeon Gold 6226R Processor. Importantly, the algorithm explicitly accounts for the collision probability's temporal dependence, which is often neglected in prior work and otherwise leads to an overestimation of the collision probability. Finally, the method is tested on a diverse set of relevant real-world scenarios, consisting of 400 6-second snippets of autonomous vehicle logs, where the accuracy and latency is rigorously evaluated.