Learning to Generate Cost-to-Go Functions for Efficient Motion Planning

TL;DR

This paper introduces c2g-HOF, a neural network architecture that directly generates cost-to-go functions for motion planning, significantly reducing computation time and enabling trajectory generation from workspace images.

Contribution

The paper proposes a novel neural network architecture that directly outputs cost-to-go functions, streamlining motion planning without extensive collision checking or iterative propagation.

Findings

c2g-HOF is significantly faster than traditional methods.

It can generate trajectories directly from workspace images.

The approach works in 2D and 3D environments.

Abstract

Traditional motion planning is computationally burdensome for practical robots, involving extensive collision checking and considerable iterative propagation of cost values. We present a novel neural network architecture which can directly generate the cost-to-go (c2g) function for a given configuration space and a goal configuration. The output of the network is a continuous function whose gradient in configuration space can be directly used to generate trajectories in motion planning without the need for protracted iterations or extensive collision checking. This higher order function (i.e. a function generating another function) representation lies at the core of our motion planning architecture, c2g-HOF, which can take a workspace as input, and generate the cost-to-go function over the configuration space map (C-map). Simulation results for 2D and 3D environments show that c2g-HOF can be orders of magnitude faster at execution time than methods which explore the configuration space during execution. We also present an implementation of c2g-HOF which generates trajectories for robot manipulators directly from an overhead image of the workspace.