Robots gain faster motion planning with new trajectory optimization method
LEAP: Gradient Flow Augmented Lagrangian Trajectory Optimization for Contact Constrained Robots
Robotics
Summary
Planning movements for legged and manipulating robots is difficult because their motions must follow complex rules, especially where their feet or hands touch the ground. The authors developed a new way called LEAP that solves these planning problems more reliably and quickly than older methods. LEAP uses a special mathematical approach to ensure the robot’s contact forces behave properly throughout the motion. Tests on a variety of robots showed LEAP can find workable motions consistently and faster, and these motions worked well on real robots and simulators.
What this means in practice
- •For robotics software engineers: Create faster and more reliable motion planners for robots with many joints and complex contact interactions using LEAP’s optimization technique.
- •For robotics hardware integrators: Use LEAP-generated trajectories to improve the execution of dynamic motions on legged and manipulating robots in real-world scenarios.
Authors
César E. Ramos Chuquiure, Ryan Tsai, Yutong Zhou, Zihang Huang, Vansh Thakur, Akshunn Trivedi, Zachary Lin, Pearl Lin, Ruikai Yang, Challen Enninful Adu, Ram Vasudevan
Abstract
Generating dynamically feasible motions for legged and manipulating robots requires solving a large, nonlinear, non-convex optimal control problem (OCP) in which contact wrenches must satisfy friction-cone constraints throughout the motion, not only at isolated instants. Existing solvers are sensitive to the initial guess and scale poorly with the number of joints. This paper introduces LEAP, which transcribes the OCP with pseudospectral collocation and, instead of a Newton or SQP solve, integrates a primal--dual gradient flow of a Powell--Hestenes--Rockafellar augmented Lagrangian to compute a solution. Because the flow is stiff, this paper designs a linearly implicit integrator with a Gauss--Newton metric, momentum, and a merit-based step controller that keeps the augmented Lagrangian monotone. Across locomotion benchmarks on five robot platforms, from planar hoppers to 32-joint humanoids, LEAP is the only method to solve every tested problem, achieves substantial speedups over Crocoddyl and Aligator on several benchmarks, and reaches trackable references at smaller compute budgets. The resulting trajectories are validated by tracking them in a high-fidelity simulator and on hardware. We will make the code public upon final acceptance of this paper.