Real time A* Adaptive Action Set Footstep Planning with Human Locomotion Energy Approximations Considering Angle Difference for Heuristic Function

TL;DR

This paper introduces a real-time adaptive footstep planning algorithm for bipedal robots that uses energy-based heuristics, angle difference considerations, and dynamic action sets to improve navigation efficiency and collision avoidance.

Contribution

It proposes a novel heuristic function considering angle difference, integrates energy approximations of human walking, and adapts the action set dynamically for efficient real-time navigation.

Findings

Effective collision avoidance demonstrated in simulations.

Reduced local minima issues in footstep planning.

Successful real robot navigation validation.

Abstract

The problem of navigating a bipedal robot to a desired destination in various environments is very important. However, it is very difficult to solve the navigation problem in real time because the computation time is very long due to the nature of the biped robot having a high degree of freedom. In order to overcome this, many scientists suggested navigation through the footstep planning. Usually footstep planning use the shortest distance or angles as the objective function based on the A * algorithm. Recently, the energy required for human walking, which is widely used in human dynamics, approximated by a polynomial function is proposed as a better cost function that explains the movement of the bipedal robot. In addition, for the real time navigation, using the action set of the A * algorithm not fixed, but the number changing according to the situation, so that the computation time does not increase much and the methods of considering the collision with the external environment are suggested as a practical method. In this thesis, polynomial function approximating the energy required for human walking is adopted as a cost function, and heuristic function considering the angular difference between the robot and the destination which is not shown in the previous studies is newly proposed and proved. In addition, a new method to integrate the adaptive behavior set and energy related to human walking is proposed. Furthermore, efficient collision avoidance method and a method to reduce the local minimum problem is proposed in this framework. Finally, footstep planning algorithm with all of these features into the mapping algorithm and the walking algorithm to solve the navigation problem is validated with simulation and real robot.