On-Demand Trajectory Predictions for Interaction Aware Highway Driving

TL;DR

This paper introduces a probabilistic trajectory prediction model for highway driving that performs well with very limited observation data, enhancing safety and decision-making in dense traffic scenarios.

Contribution

It extends a deterministic car-following model into a probabilistic framework with real-time inference, enabling accurate predictions from as little as 400ms of data.

Findings

Achieves state-of-the-art accuracy in dense traffic prediction

Performs well with short observation windows (400ms)

Demonstrates effectiveness on NGSIM dataset

Abstract

Highway driving places significant demands on human drivers and autonomous vehicles (AVs) alike due to high speeds and the complex interactions in dense traffic. Merging onto the highway poses additional challenges by limiting the amount of time available for decision-making. Predicting others' trajectories accurately and quickly is crucial to safely executing maneuvers. Many existing prediction methods based on neural networks have focused on modeling interactions to achieve better accuracy while assuming the existence of observation windows over 3s long. This paper proposes a novel probabilistic model for trajectory prediction that performs competitively with as little as 400ms of observations. The proposed model extends a deterministic car-following model to the probabilistic setting by treating model parameters as unknown random variables and introducing regularization terms. A realtime inference procedure is derived to estimate the parameters from observations in this new model. Experiments on dense traffic in the NGSIM dataset demonstrate that the proposed method achieves state-of-the-art performance with both highly constrained and more traditional observation windows.