YOPO-Rally: A Sim-to-Real Single-Stage Planner for Off-Road Terrain
🤖 AI Summary
Autonomous navigation for robots in unstructured off-road terrains—such as forests—faces challenges including dense obstacles, complex traversability assessment, and unreliable localization. Method: We propose an end-to-end, single-stage navigation planner featuring a novel unified neural architecture that jointly models terrain traversability estimation and path search; employs non-uniform cubic Hermite splines to represent multimodal expert trajectories; and accepts raw depth images, ego-velocity, and goal vectors to directly output smooth, robust, multi-candidate trajectories. Contribution/Results: Our key contribution is zero-shot sim-to-real transfer—enabling direct deployment in real forests without fine-tuning. Trained via behavioral cloning on the YOPO-Sim simulator, our method achieves significant improvements in both success rate (+23.6%) and real-time performance (inference <30 ms) in both simulation and physical experiments, demonstrating robustness across soft, sloped, and high-obstacle-density terrains.
Application Category
📝 Abstract
Off-road navigation remains challenging for autonomous robots due to the harsh terrain and clustered obstacles. In this letter, we extend the YOPO (You Only Plan Once) end-to-end navigation framework to off-road environments, explicitly focusing on forest terrains, consisting of a high-performance, multi-sensor supported off-road simulator YOPO-Sim, a zero-shot transfer sim-to-real planner YOPO-Rally, and an MPC controller. Built on the Unity engine, the simulator can generate randomized forest environments and export depth images and point cloud maps for expert demonstrations, providing competitive performance with mainstream simulators. Terrain Traversability Analysis (TTA) processes cost maps, generating expert trajectories represented as non-uniform cubic Hermite curves. The planner integrates TTA and the pathfinding into a single neural network that inputs the depth image, current velocity, and the goal vector, and outputs multiple trajectory candidates with costs. The planner is trained by behavior cloning in the simulator and deployed directly into the real-world without fine-tuning. Finally, a series of simulated and real-world experiments is conducted to validate the performance of the proposed framework.
Problem
Research questions and friction points this paper is trying to address.
Develops sim-to-real planner for off-road robot navigation
Integrates terrain analysis and pathfinding into neural network
Enables zero-shot transfer from simulation to real-world
Innovation
Methods, ideas, or system contributions that make the work stand out.
Unity-based simulator for randomized forest environments
Single neural network integrating TTA and pathfinding
Zero-shot transfer from simulation to real-world
🔎 Similar Papers
No similar papers found.
