Centralized Multi-Robot Systems

Introduction

Multi-robot systems combine heterogeneous robotic platforms to perform complex tasks that exceed the capabilities of individual robots. By coordinating the complementary strengths of aerial and ground vehicles, these systems can efficiently acquire information, allocate resources, and adapt to dynamic environments where uncertainty and changing mission conditions are inherent.

Research at CIMLab develops methodologies for centralized decision-making, mission planning, task allocation, and cooperative robot coordination. Particular emphasis is placed on autonomous search-and-rescue, where teams of unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) collaborate to locate and identify untrackable targets in large, complex environments. These systems integrate probabilistic target estimation, adaptive mission planning, heterogeneous robot coordination, and wireless sensing technologies to improve search effectiveness while making efficient use of limited robotic resources.

Our research combines optimization, probabilistic reasoning, trajectory planning, and multi-robot coordination into experimentally validated robotic systems. Methodologies are evaluated through simulation and physical experiments using heterogeneous UAV and UGV platforms together with wireless sensor networks, enabling robust operation in realistic outdoor environments characterized by uncertainty, limited information, and dynamic mission requirements.

Representative Research: An Adaptive Approach to Optimal Sparse Mobile-Target Search Planning Using Heterogeneous Agents – Zendai Kashino (PhD, 2020)

Locating missing persons in large wilderness environments is a challenging problem because search resources are often limited while the search area continues to expand over time. This research developed adaptive search-planning methodologies that coordinate heterogeneous teams of unmanned aerial vehicles (UAVs), unmanned ground vehicles (UGVs), and wireless sensor networks to maximize the probability of locating untrackable targets. The principal contributions include novel strategies for dynamic sensor-network deployment, hybrid aerial–ground search planning, and probabilistic mission planning based on time-dependent target-location likelihood. Multi-robot search and rescue system Developed primarily for wilderness search and rescue, these methodologies were extensively validated through simulated and miniature-scale physical experiments, demonstrating significant improvements in search efficiency and probability of success. Beyond search and rescue, the underlying planning and coordination methodologies are applicable to a broad range of environmental monitoring, surveillance, and mobile-target search applications.

Selected Publications:

Wang, Z., & Benhabib, B. (2025). Concurrent Multi-Robot Search of Multiple Missing Persons in Urban Environments. Robotics, 14(11), 157.
DOI: 10.3390/robotics14110157
Publisher: https://doi.org/10.3390/robotics14110157

Kashino, Z., Nejat, G., & Benhabib, B. (2020). Aerial Wilderness Search and Rescue with Ground Support. Journal of Intelligent & Robotic Systems, 99(1), 147–163.
DOI: 10.1007/s10846-019-01105-y
Publisher: https://doi.org/10.1007/s10846-019-01105-y

Kashino, Z., Nejat, G., & Benhabib, B. (2020). A Hybrid Strategy for Target Search Using Static and Mobile Sensors. IEEE Transactions on Cybernetics, 50(2), 856–868.
DOI: 10.1109/TCYB.2018.2875625
Publisher: https://doi.org/10.1109/TCYB.2018.2875625

Shin, J. C. L., Kashino, Z., Nejat, G., & Benhabib, B. (2019). A Sensor-Network-Supported Mobile-Agent Search Strategy for Wilderness Rescue. Robotics, 8(3), 61.
DOI: 10.3390/robotics8030061
Publisher: https://doi.org/10.3390/robotics8030061