MARLAS: Multi Agent Reinforcement Learning for cooperated Adaptive Sampling

Abstract

The multi-robot adaptive sampling problem aims at finding trajectories for a team of robots to efficiently sample the phenomenon of interest within a given endurance budget of the robots. In this paper, we propose a robust and scalable approach using decentralized Multi-Agent Reinforcement Learning for cooperated Adaptive Sampling (MARLAS) of quasi-static environmental processes. Given a prior on the field being sampled, the proposed method learns decentralized policies for a team of robots to sample high-utility regions within a fixed budget. The multirobot adaptive sampling problem requires the robots to coordinate with each other to avoid overlapping sampling trajectories. Therefore, we encode the estimates of neighbor positions and intermittent communication between robots into the learning process. We evaluated MARLAS over multiple performance metrics and found it to outperform other baseline multi-robot sampling techniques. We further demonstrate robustness to communication failures and scalability with both the size of the robot team and the size of the region being sampled. The experimental evaluations are conducted both in simulations on real data and in real robot experiments on demo environmental setup 1 .

Energy-Efficient Blockchain-Enabled Multi-Robot Coordination for Information Gathering: Theory and Experiments

“…Wei and Zheng [28] have proposed an independent learning technique with credit assignment to solve this notoriously difficult problem. Pan, Manjanna, and Hsieh [29] have recently proposed a policy gradient-based DRL for multi-robot information sampling. Unlike the prior studies, they do not use GP as the underlying information inference tool.…”

Bioinspired Multipurpose Approach to the Sampling Problem

“…This class of approaches loads the communication channel to an even greater extent, but this disadvantage is compensated by the lower dependence of the problem solving process on the imposed maximum message transmission range constraints. Thus, in [23], a decentralized adaptive sampling approach based on reinforcement learning during the problem solving process is proposed. The strategy has been compared with the centralized DARP area division algorithm [24], taking into account the different numbers of mobile robots used and the communication range constraints.…”