Multi-Robot Enhanced MANET Routing with Situation Aware Online Reinforcement Learning

Abstract

This paper studies the problem of real-time routing in a multi-autonomous robot enhanced network under uncertain and complex environment. Emerging network protocols, such as opportunistic mobile network routing protocols, engaged social network in communication network that can increase the interoperability by using social mobility and opportunistic carry and forward routing algorithms. However, the uncertainty of social mobility and complexity of harsh environment would seriously affect the effectiveness and practicality of those emerging network protocols. This paper presents a SaRE-MANET (Situation-aware Robot Enhanced Mobile Ad-hoc Network) routing protocol that adopt the online reinforcement learning technique to supervise the mobility of multi-robots as well as handle the practical uncertainty and complexity of harsh environment. Firstly, a set of mission oriented metrics has been introduced to describe the interrelation between network quality and multi-robots' mobility. Then, a distributed multi-agent reinforcement learning algorithm has been proposed. SaRE-MANET routing protocol as well as the mobility of multi-robots are optimized online by using the practical mission oriented metrics. The effectiveness of the proposed design has been demonstrated through both computer-aid simulation and experiments.