Abstract:Humanoids' abilities to navigate stairs and uneven terrain make them well-suited for disaster response efforts. However, humanoid navigation in such environments is currently limited by the capabilities of navigation planners. Such planners typically consider only footstep locations, but planning with palm contacts may be necessary to cross a gap, avoid an obstacle, or maintain balance. However, considering palm contacts greatly increases the branching factor of the search, leading to impractical planning times for large environments. Planning a contact transition sequence in a large environment is important because it verifies that the robot will be able to reach a given goal. In previous work we explored using library-based methods to address difficult navigation planning problems requiring palm contacts, but such methods are not efficient when navigating an easy-to-traverse part of the environment. To maximize planning efficiency, we would like to use discrete planners when an area is easy to traverse and switch to the library-based method only when traversal becomes difficult. Thus, in this paper we present a method that (1) Plans a torso guiding path which accounts for the difficulty of traversing the environment as predicted by learned regressors; and (2) Decomposes the guiding path into a set of segments, each of which is assigned a motion mode (i.e. a set of feet and hands to use) and a planning method. Easily-traversable segments are assigned a discrete-search planner, while other segments are assigned a library-based method that fits existing motion plans to the environment near the given segment. Our results suggest that the proposed approach greatly outperforms standard discrete planning in success rate and planning time. We also show an application of the method to a real robot in a mock disaster scenario.

Global Navigation for Humanoid Robots Using Sampling-Based Footstep Planners

A Random Sampling-Based Approach to Goal-Directed Footstep Planning for Humanoid Robots

A deterministic sampling-based approach to global footstep planning for humanoid robots

A Real-Time Randomized Navigation Method for Biped Robot

Long-horizon humanoid navigation planning using traversability estimates and previous experience

Footstep Planning For Rapid Path Following In Humanoid Robots

Modeling and Algorithm Realization of Footstep Planning for Humanoid Robots

Global footstep planning with greedy and heuristic optimization guided by velocity for biped robot

Scaling Sampling-based Motion Planning to Humanoid Robots

Omni-directional Bipedal Walking Path Planning

Advances in research on motion planning for humanoid robots

Fast Kinodynamic Bipedal Locomotion Planning with Moving Obstacles

Effective Footstep Planning Using Homotopy-Class Guidance

Motion planning for humanoid robots in complex environments based on stance sequence and ZMP reference

Combined Sampling and Optimization Based Planning for Legged-Wheeled Robots

Footstep Planning for Autonomous Walking Over Rough Terrain

Parameter Optimization In Non-Uniform Randomized Footstep Planning For Biped Navigation

Footstep and Motion Planning in Semi-unstructured Environments Using Randomized Possibility Graphs

Traversability-Aware Legged Navigation by Learning from Real-World Visual Data

Dynamic Humanoid Locomotion over Uneven Terrain With Streamlined Perception-Control Pipeline

Robust Humanoid Contact Planning with Learned Zero- and One-Step Capturability Prediction