Online Connectivity-aware Dynamic Distribution for Heterogeneous Multi-Robot

Abstract

In many multi-robot applications, the robot team often needs to execute multiple tasks simultaneously with diverse capabilities and task-prescribed controllers. To ensure effective coordination and collaboration, robots from different tasks not only need to stay collision-free but also connected within each task subgroups as well as across different subgroups. In this thesis, we consider the dynamic task allocation and control problem, where a heterogeneous group of networked robots must be assigned and moved to multiple dynamic task places (not known beforehand) to maximize the overall task performance over time. As each task requires various capabilities from the assigned robots and has different weights of importance, the objective is to find the optimal combination of robots assigned to each task and the controllers to drive the robots towards the task sites, while enforcing safety and connectivity (global and subgroup) guarantee. In particular, we propose a unified optimization-based framework for the robots to compute the real-time task assignments with bounded optimality and the controllers that ensure inter-robot collision avoidance and connectivity maintenance. Such a framework allows for autonomous task re-assignment and redistribution of robots to improve task fulfillments during execution in the presence of dynamic task requirements. Simulation and numerical results are provided to demonstrate the effectiveness of the proposed approaches.