Plugin Framework-Based Neuro-Symbolic Grounded Task Planning for Multi-Agent System.

Jiyoun Moon

doi:10.3390/s21237896

Abstract

As the roles of robots continue to expand in general, there is an increasing demand for research on automated task planning for a multi-agent system that can independently execute tasks in a wide and dynamic environment. This study introduces a plugin framework in which multiple robots can be involved in task planning in a broad range of areas by combining symbolic and connectionist approaches. The symbolic approach for understanding and learning human knowledge is useful for task planning in a wide and static environment. The network-based connectionist approach has the advantage of being able to respond to an ever-changing dynamic environment. A planning domain definition language-based planning algorithm, which is a symbolic approach, and the cooperative–competitive reinforcement learning algorithm, which is a connectionist approach, were utilized in this study. The proposed architecture is verified through a simulation. It is also verified through an experiment using 10 unmanned surface vehicles that the given tasks were successfully executed in a wide and dynamic environment.

Highlights

The demands for robots are consistently increasing in various domains including ground, air, and the ocean as the robots are proven to be efficient based on the convergence of new technologies such as artificial intelligence and big data [1]
The environment information is updated in real-time using the knowledge base, and it is verified whether the task is executed even when an unexpected situation occurs
The proposed plugin-based neuro-symbolic framework is verified in a wide and dynamic environment using a multi-agent system consisting of four teams

Summary

Introduction

The demands for robots are consistently increasing in various domains including ground, air, and the ocean as the robots are proven to be efficient based on the convergence of new technologies such as artificial intelligence and big data [1]. Some of the fields that utilize multi-agent systems include surveillance [4,5,6], data collection [7,8], and inspection [9,10,11]. Wang et al [12] proposed a multi-agent system based on generalized covariance intersection for multi-view surveillance in centralized and decentralized situations. Laport et al [13] proposed a multiagent architecture for collecting massive data using mobile sensing devices. Jing et al [14] introduced a coverage path planning framework for the large and complex structure inspection of multiple unmanned aerial vehicles

Objectives

Methods

Results

Conclusion