Alapfogalmak
Efficient real-time motion control system using Control Barrier Functions for the six-wheeled-telescopic-legged robot Tachyon 3.
Kivonat
The content discusses a lightweight real-time perceptive motion control system for the newly developed six-wheeled-telescopic-legged robot, Tachyon 3. It introduces analytically smoothed constraints integrating Smooth Separating Axis Theorem (SSAT) with Control Barrier Functions (CBF) to achieve online motion generation while ensuring joint limitations, environmental collision avoidance, and safe convex foothold constraints. The proposed method is validated through stair-climbing motions in both simulation and real machine experiments. The paper outlines hardware structure, recognition process, reference motion generation, CBF QP formulation, and analytical smoothed CBF constraints for Tachyon 3.
I. INTRODUCTION
- Legged robots face challenges with fall risks due to high center of gravity.
- Newly developed Tachyon 3 improves stability and energy efficiency.
- Motion generation under constraints is crucial for safe locomotion.
II. SIX-WHEELED-TELESCOPIC-LEGGED ROBOT TACHYON 3
- Hardware structure includes telescopic legs with driving wheels.
- Unique joint configuration enhances stability and energy efficiency.
III. REAL-TIME PERCEPTIVE MOTION CONTROL FOR TACHYON 3
- Overview of the proposed system involving recognition, reference motion generation, and safety filter using ECBF.
IV. ANALYTICAL SMOOTH CBF CONSTRAINTS FOR TACHYON 3
- Detailed description of continuously differentiable constraints using SSAT.
V. EXPERIMENTS
- Benchmarking computational efficiency of SSAT against other methods.
- Simulation experiments validate safety performance of proposed CBF.
VI. CONCLUSIONS
- Proposed analytically smooth CBF formulations enhance real-time control for Tachyon 3.
Statisztikák
"The computational time of random cuboid’s collision detection."
"The computational time for calculating collision detection value."
Idézetek
"The proposed method integrating the CBF achieves online motion generation in a short control cycle of 1 ms."
"Furthermore, compared with the optimization-based method, the proposed method is more than 10 times faster."