Problem
Research questions and friction points this paper is trying to address.
Quadruped Robot
Terrain Adaptation
Stable Locomotion
Quadrupedal Footstep Planning Using Learned Motion Models of a Black-Box Controller
🤖 AI Summary
To address the limitations of commercial quadrupedal robots—namely, their non-programmable black-box controllers that accept only high-level velocity commands—this paper proposes an end-to-end autonomous gait planning framework. First, a deep neural network is employed to learn the kinematic and dynamic behavior of the black-box controller, effectively transforming it into a differentiable, plannable surrogate model. Second, monocular vision-based terrain estimation is tightly integrated with an improved A* graph search algorithm to jointly optimize, in real time, the center-of-mass trajectory, footstep sequence, and velocity commands. The framework achieves 10 Hz autonomous navigation over unstructured terrains—including gravel, stairs, and slopes—on physical hardware. Compared to conventional approaches, the system demonstrates over threefold improvement in robustness and cross-terrain generalization. This work marks the first demonstration of learning-based motion modeling for black-box platforms and establishes a fully closed-loop perception–planning–control autonomy pipeline.
📝 Abstract
Legged robots are increasingly entering new domains and applications, including search and rescue, inspection, and logistics. However, for such a systems to be valuable in real-world scenarios, they must be able to autonomously and robustly navigate irregular terrains. In many cases, robots that are sold on the market do not provide such abilities, being able to perform only blind locomotion. Furthermore, their controller cannot be easily modified by the end-user, requiring a new and time-consuming control synthesis. In this work, we present a fast local motion planning pipeline that extends the capabilities of a black-box walking controller that is only able to track high-level reference velocities. More precisely, we learn a set of motion models for such a controller that maps high-level velocity commands to Center of Mass (CoM) and footstep motions. We then integrate these models with a variant of the $A$ * algorithm to plan the CoM trajectory, footstep sequences, and corresponding high-level velocity commands based on visual information, allowing the quadruped to safely traverse irregular terrains at demand.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Dynamic Body Adjustment
Real-time Path Planning
Visual Information Integration
🔎 Similar Papers
No similar papers found.
