Publication

Behavior-aware online prediction of obstacle occupancy using zonotopes

Conference Article

Conference

IEEE Conference on Decision and Control (CDC)

Edition

2025

Pages

7337-7342

Doc link

http://dx.doi.org/DOI: 10.1109/CDC57313.2025.11312375

File

Download the digital copy of the doc pdf document

Authors

Abstract

Predicting the motion of surrounding vehicles is key to achieving safe autonomous driving, especially in scenarios where road geometry is unknown or no prior information about the surrounding obstacles is available. This paper proposes a novel method to accurately and efficiently predict the occupancy sets of surrounding vehicles based on online observations of their behaviors. The approach is divided in two stages: First, a zonotopic set is computed to represent the estimated behavior of the vehicle. This is achieved using optimal observers and solving a linear programming (LP) problem. Then, a reachability analysis is performed to predict the corresponding zonotopic occupancy sets over a fixed prediction horizon. The effectiveness of the method has been validated through simulations in an urban environment, showing accurate and compact predictions without relying on assumptions or prior training data.

Categories

control system analysis, mobile robots, optimal control, optimisation, predictive control.

Author keywords

Set theory, autonomous driving, reachability analysis, formal methods, zonotopes

Scientific reference

A. Carrizosa, J. Zhou, E. Frisk, V. Puig and F. Nejjari. Behavior-aware online prediction of obstacle occupancy using zonotopes, 2025 IEEE Conference on Decision and Control, 2025, Rio de Janeiro, Brazil, pp. 7337-7342, .