Abstract:This study proposes a hierarchically integrated framework for safe task and motion planning (TAMP) of bipedal locomotion in a partially observable environment with dynamic obstacles and uneven terrain. The high-level task planner employs linear temporal logic for a reactive game synthesis between the robot and its environment and provides a formal guarantee on navigation safety and task completion. To address environmental partial observability, a belief abstraction model is designed by partitioning the environment into multiple belief regions and employed at the high-level navigation planner to estimate the dynamic obstacles' location. This additional location information of dynamic obstacles offered by belief abstraction enables less conservative long-horizon navigation actions beyond guaranteeing immediate collision avoidance. Accordingly, a synthesized action planner sends a set of locomotion actions to the middle-level motion planner while incorporating safe locomotion specifications extracted from safety theorems based on a reduced-order model (ROM) of the locomotion process. The motion planner employs the ROM to design safety criteria and a sampling algorithm to generate nonperiodic motion plans that accurately track high-level actions. At the low level, a foot placement controller based on an angular-momentum linear inverted pendulum model is implemented and integrated with an ankle-actuated passivity-based controller for full-body trajectory tracking. To address external perturbations, this study also investigates the safe sequential composition of the keyframe locomotion state and achieves robust transitions against external perturbations through reachability analysis. The overall TAMP framework is validated with extensive simulations and hardware experiments on bipedal walking robots Cassie and Digit designed by Agility Robotics.

Footstep Planning with Encoded Linear Temporal Logic Specifications

Safe Planning through Incremental Decomposition of Signal Temporal Logic Specifications

Solving Footstep Planning as a Feasibility Problem using L1-norm Minimization (Extended Version)

SPIN-Based Linear Temporal Logic Path Planning for Ground Vehicle Missions with Motion Constraints on Digital Elevation Models

Footstep Planning for Autonomous Walking Over Rough Terrain

A Generalized Mixed-Integer Convex Program for Multilegged Footstep Planning on Uneven Terrain

LSTP: Long Short-Term Motion Planning for Legged and Legged-Wheeled Systems

Time-Robust Path Planning with Piece-Wise Linear Trajectory for Signal Temporal Logic Specifications

Physics-based Motion Planning with Temporal Logic Specifications

End-to-End Path Planning Under Linear Temporal Logic Specifications

Synthesis of output feedback control for motion planning based on LTL specifications

Vision-Based Reactive Temporal Logic Motion Planning for Quadruped Robots in Unstructured Dynamic Environments

Integrated Task and Motion Planning for Safe Legged Navigation in Partially Observable Environments

LTL-D*: Incrementally Optimal Replanning for Feasible and Infeasible Tasks in Linear Temporal Logic Specifications

Extended LTLvis Motion Planning interface (Extended Technical Report)

Deep Learning-Based Path Planning Under Co-Safe Temporal Logic Specifications

DT*: Temporal Logic Path Planning in a Dynamic Environment

Optimization-based Task and Motion Planning under Signal Temporal Logic Specifications using Logic Network Flow

SafeSteps: Learning Safer Footstep Planning Policies for Legged Robots via Model-Based Priors

Multi-Agent Motion Planning From Signal Temporal Logic Specifications

Perception-Based Temporal Logic Planning in Uncertain Semantic Maps