- Coverage Control on the Special Euclidean Groups
We developed an optimal coverage strategy for a multi-robot system across a domain of interest for applications involving pose-sensitive event services, e.g., elder care robots lift fallen people in certain poses to try to achieve the best user experience, and mobile charging robots transmit power to another team of robots that are designed to execute specific tasks and need to be recharged in certain poses. The multi-robot system has heterogeneous relative mobility, which can be naturally encoded by the Riemannian metrics.
R. Lin and M. Egerstedt, “Coverage Control on the Special Euclidean Groups,” in 2023 American Control Conference (ACC). IEEE, 2023, pp. 1972–1979.
- Dynamic Multi-Target Tracking Using Heterogeneous Coverage Control
We developed a coverage-based collaboration strategy for a multi-robot system with heterogeneous perception and mobility characteristics to simultaneously track and estimate the states of multiple targets governed by stochastic dynamics. The heterogeneous robots leverage complementary capabilities due to specific task requirements and inherent constraints to safely and collaboratively accomplish certain tasks.
R. Lin and M. Egerstedt, “Dynamic Multi-Target Tracking Using Heterogeneous Coverage Control,” in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2023, pp. 11103–11110.
- Mutualistic Interactions in Heterogeneous Multi-Agent Systems
This paper aims to answer two fundamental questions: “Why should agents collaborate?” and “How should agents collaborate?”. In this work, agents with different capabilities can work together within a shared workspace to accomplish their tasks. Moreover, the collaboration between agents is inspired by the ecological concept known as mutualism, i.e., an interaction between two or more species that benefits everyone involved. The collaborative arrangements between the agents are made possible through the composition of barrier functions, which expand and contract the agents’ safe sets. We also establish a framework for enabling collaborative interactions in a multi-agent setting. Lastly, we demonstrate our collaboration framework in two case studies requiring the agents to work together to complete their respective tasks.
A. A. Nguyen, F. Jabbari, and M. Egerstedt, “Mutualistic interactions in heterogeneous multi-agent systems,” in IEEE Conference on Decision and Control, 2023, pp. 411–418.
- Predator-Prey Interactions through Heterogeneous Coverage Control Using Reaction-Diffusion Processes
We developed a decentralized predator-prey interaction scheme based on a heterogeneous cooperative control strategy driven by reaction-diffusion processes, where the heterogeneity is understood along five modalities in terms of the dynamics of predators and prey. The proposed scheme can be applied to diverse applications, e.g., deployments of teams of: mobile robots for eroded beach monitoring or to clean oil spills, underwater robots for halobios/ocean floor/coral reefs monitoring, and aerial robots to put out (ideally modeled) wildfires.
R. Lin and M. Egerstedt, “Predator-Prey Interactions through Heterogeneous Coverage Control Using Reaction-Diffusion Processes,” in 2023 IEEE Conference on Decision and Control (CDC), Singapore, Dec. 2023.
