A Perception-Based Architecture for Autonomous Convoying in GNSS-Denied Areas
Alexander Bienemann, Lukas Beer, Andreas Reich, Thomas Steinecker, Bianca Forkel, Anton Backhaus, Philipp Berthold, Juan David González González, Peter Mortimer, Thorsten Luettel, Maehlisch Mirko
AI summary
Problem
Existing autonomous convoying systems rely heavily on GNSS, prerecorded maps, or line-of-sight tracking, causing failures in GPS-denied areas, unstructured terrain, or during occlusions. They also frequently lack real-world validation under adverse weather and dynamic obstacle conditions.
Approach
The authors designed a full-stack architecture that fuses perception-based vehicle tracking with simultaneous localization and mapping (SLAM), managed by a waypoint generator and controlled via model predictive control, allowing convoying without external positioning or maps.
Key results
- High path-following accuracy across 5–100 m distances at speeds up to 20 m/s
- Successful real-world validation across varying light, weather, and off-road conditions
- Robust performance demonstrated during the 12th European Land Robot Trial (ELROB) 2024
- Redundant SLAM-tracking fusion maintains convoying during line-of-sight loss or GNSS jamming
Why it matters
Enables reliable autonomous logistics and military convoy operations in GPS-denied, unstructured environments where current systems fail.
Abstract
In this article, we present a perception-based full-stack system for autonomous vehicle fol- lowing that does not rely on accurate global localization or map data. Our architecture consists of modules for vehicle communication, localization, object tracking, waypoint management, static environment modeling, trajectory planning, and control, which all are covered in the article. To test our system, we conducted several practical experiments in various scenarios on our two autonomous vehicles. Those experiments include the handling of static and dynamic obstacles, driving on- and off-road under different light and weather conditions with distances between the vehicles ranging from 5 to 100 m and with speeds of up to 20 m/s. Furthermore, we showcased our system’s performance during the 12th European Land Robot Trial (ELROB) 2024, where our institute participated in the convoying scenario. The tests from the trial and our own experiments showed satisfactory results. Our system archives a high path-following accuracy and is able to cope with various challenging scenarios.