Research Objective:

We want robot teams to monitor unknown dynamic environments. This requires assigning robots to sampling locations, a variant of the Multi Robot Task Allocation (MRTA) problem. Existing methods reduce the MRTA problem to a resource assignment problem and do not have the necessary scalability or fleixbility. In contrast, macroscopic ensemble methods have the scalability but lack the necessary flexibility to adapt to environment changes. This project aims to improve macroscopic ensemble methods in two critical ways. Firstly, we will use environment model feedback to inform desired distributions of robots performing spatially distributed environmental monitoring tasks. Secondly, we leverage the recent success of controlling higher order macroscopic ensemble moments to allow for small robot team size. Together we show improved team-wide flexibility with evaluation done in simulation and experimentally.

We experimentally verify our methods using the miniature Autonomous Surface Vehicles (mASVs) which are built in house at the ScalAR lab. Our testing environment is a 4m x 3m x 1.5m tank, equipped with 13 OptiTrack cameras, and the capability to make gyres in the tank. At any given time we can have upwards of 10 mASV in the tank performing different control strategies. Within the tank we have demonstrated heterogeneous teams using Crazyflie robots and mASV simultaneously.

Relevant Publications:

The original derivation of moments can be found in (Silva et al., 2022). We have an extended abstract which discusses a distributed approach to the adaptive macroscopic ensemble allocation framework (Edwards & Hsieh, 2025). Finally, we are currently preparing a journal manuscript to introduced adaptive macroscopic ensemble allocation which takes in environment feedback to inform robot task selection.

This work is done in collaboration with Dr. Thales C. Silva and Dr. M. Ani Hsieh at the University of Pennsylvania GRASP lab.