Multi-modal system improves object tracking for autonomous racing cars
A Multi-Modal Perception Pipeline for Object Detection and Tracking in Autonomous Racing
RoboticsArtificial Intelligence
Summary
Detecting and tracking other vehicles is tricky for self-driving race cars because they move very fast and face tough conditions like bumps and sensor noise. The authors designed a system that combines data from different sensors like cameras, LiDAR, and RADAR to get better and faster information about nearby vehicles. Their method also uses what is known about car behavior and the race track to predict movements more accurately. Tests on real racing data showed the system works well, even in tricky situations, and can help cars make safer driving decisions.
What this means in practice
- •For autonomous vehicle developers: Use combined camera, LiDAR, and RADAR data with tracking that accounts for delays and vehicle behavior to enhance perception in high-speed racing environments.
- •For urban autonomous vehicle teams: Improve vehicle detection and tracking systems under challenging conditions encountered in city driving by adapting techniques tested in racing scenarios.
Authors
Davide Malvezzi, Michele Pestarino, Vittoria Cavicchioli, Valentina La Gamba, Silvia Severi, Fabio Bagni, Luca Bartoli, Massimiliano Bosi, Francesco Gatti, Micaela Verucchi, Ayoub Raji, Marko Bertogna
Abstract
Object detection and tracking are fundamental components of perception systems for autonomous driving. Achieving robust performance under adverse conditions such as limited visibility, sensor noise, and failures remains an open challenge, particularly in autonomous racing, where vehicles operate at very high speeds, experience strong vibrations, and interact under small safety margins. This paper presents a multi-modal late-fusion perception pipeline for object detection and tracking in the autonomous racing domain. The proposed system extends previous work by exploiting all onboard sensors through a late-fusion approach and a dedicated multi-object tracking framework. Independent detections from cameras, LiDARs, and RADARs are combined to provide timely and robust state estimates of surrounding vehicles. The tracking method explicitly compensates for detection delays and embeds in its model prior knowledge of vehicle dynamics and track layout. Experimental evaluation on real-world data across diverse critical scenarios, representative of challenging edge cases also in urban driving, confirms the effectiveness of the proposed pipeline and its suitability to support safe and adaptive planning decisions.