This article discusses the development of HOUND, a cost-effective platform for off-road vehicle research. Off-road vehicles often face the risk of rollovers on challenging terrains. While ruggedization through industrial-grade parts can protect against rollovers, it is often too expensive for research labs. The authors propose a software-based rollover-prevention system (RPS) that allows the use of cheaper commercial-off-the-shelf hardware parts. They integrated this RPS into HOUND, a small-scale off-road autonomy platform. The platform can handle challenging terrains at high speeds and has been tested in real-world conditions. HOUND also integrates with BeamNG, a driving simulator, demonstrating a significant reduction in rollovers.
Publication date: 21 Nov 2023
Project Page: https://sites.google.com/view/prl-hound
Paper: https://arxiv.org/pdf/2311.11199
