Search papers, labs, and topics across Lattice.
This paper introduces a three-stage whole-body planning framework for humanoids navigating confined spaces, addressing the limitations of traditional trajectory optimizers that struggle with self-collision and dense obstacles. By integrating differentiable collision avoidance into a reachability-constrained formulation, the authors synthesize volume-informed guides that enhance the performance of a full-order trajectory optimizer. The proposed method demonstrates superior performance on the Unitree G1 humanoid, achieving high-quality trajectories in complex tasks that meet NIST emergency response standards, where conventional methods fail.
Humanoids can now navigate confined spaces with unprecedented efficiency, generating feasible trajectories in complex environments that traditional methods can't handle.
Humanoid locomotion in highly confined environments requires navigating dense environmental obstacles and complex self-collision bounds while maintaining multi-contact dynamic feasibility. Traditional trajectory optimizers frequently struggle in these restricted spaces, as navigating the large collision space with splines on particle abstractions is insufficient and leads to poor local minima. To address this, we propose a three-stage whole-body planning framework that formulates kinematic path planning directly over kinematically reachable rigid-body volumes. By integrating differentiable collision avoidance into a reachability-constrained formulation, our framework synthesizes volume-informed guides that reliably guide a full-order trajectory optimizer over long horizons. We show that these optimized plans serve as high-quality references to train a residual reinforcement learning policy for robust online execution. We validate our approach on the Unitree G1 humanoid across three benchmark testbeds exceeding NIST emergency response standards, achieving restricted confinement ratios ($C_r<1.5$). Our framework generates feasible trajectories across 12-to-18-second tasks with complex foot and hand contacts where standard baselines fail, while the learned policy successfully tracks these plans under extensive domain randomization in physics simulation.
