Task planning method and device based on end point backtracking under signal sequential logic constraint

Abstract

The invention discloses a task planning method and device based on end point backtracking under signal sequential logic constraint, and the method comprises the steps of firstly building a search tree offline, namely adding the time dimension information in a two-dimensional position working space of an intelligent agent, and building a three-dimensional space; under the signal sequential logic constraint, in the three-dimensional space, taking an end point estimated according to a given task of the intelligent agent as a root node, and growing a fast random search tree towards an initial state plane until each possible position point in the initial state plane is covered by a leaf node; during the online task planning, giving an initial state of the intelligent agent, searching all leaf nodes in a reachable range of the intelligent agent in the fast random search tree, and selecting the leaf node with the minimum cost value as a planning node, wherein the path of the intelligent agent from the initial state to the planning node and then to the root node is a path planning result. According to the method, the reachable state space can be traversed, the convergence success rate is improved, and the online planning complexity is reduced.

Description

technical field [0001] The invention relates to the technical field of intelligent agent task planning, in particular to a method and device for task planning based on end-point backtracking under signal timing logic constraints. Background technique [0002] Signal Temporal Logic is a technical field with research prospects in the field of robot planning. Signal timing logic can use a programming language to model a series of complex tasks with timing relationships and strict time constraints, so as to add high-level timing constraints to the control and planning of the robot, and provide an interface for the underlying controller. In the field of robot planning, sometimes robot tasks are complex and have the characteristics of coupling in timing relationships. Some traditional methods are difficult to model, but signal timing logic can well model complex tasks of multi-agent systems. [0003] For signal timing logic planning tasks, the existing solutions are as follows: ...

AI Technical Summary

Problems solved by technology

[0007] For the third scheme, it is hoped to find the path with the largest reward, but because the state space to be explored is often too large, and because the signal timing logic comes with strict time constraints, many exploration paths violate the time constraints before completing the task. , so it is difficult to converge, and the optimal path cannot be found, resulting in a decrease in the success rate of path planning

Images