MAD Games workshop on Multi-Agent Dynamic Games at ICRA 2024 was organized by Rahul Mangharam, Hongrui Zheng, Shuo Yang, Johannes Betz and Venkat Krovi. icra2024-madgames.f1tenth.org/
In decentralized multi-robot navigation, ensuring safe and efficient movement with limited environmental awareness remains a challenge. Traditional navigation, based on local observations, often falters in complex environments. A potential solution is to enhance the understanding of the world through inter-agent communication; however, mere information broadcasting lacks efficiency. This work addresses the challenge by simultaneously learning decentralized multi-robot collision avoidance and selective inter-agent communication. Our method leverages a visual transformer and self-attention mechanism to encode local occupancy maps and the state information of neighboring agents into fixed-length encodings, facilitating the handling of an arbitrary number of neighbors for collision-free navigation. We focus on improving navigation performance through selective communication, casting the communication selection as a link prediction problem. The network determines the necessity of establishing a communication link with specific neighbors based on observable state information. We consider communicating the agent's goal information with their relevant neighbors. Additionally, we also explore what to communicate by encoding the agent’s state and observation information into a fixed-length message vector, communicated in lieu of only goal information. The communicated information enhances the neighbor's observation and aids in selecting an appropriate navigation plan. By training the network end-to-end, we concurrently learn the optimal weights for the observation encoder, communication selection, and navigation components. Our approach demonstrates significant benefits by achieving safe and efficient navigation among multiple robots in dense and challenging environments. Comparative evaluations against various learning-based and model-based baselines demonstrate our superior navigation performance, resulting in an impressive improvement of up to 24% in success rates within complex evaluation scenarios. Additionally, we evaluate our method against state-of-the-art baselines in complex scenarios, including narrow corridors and environments with multiple agents, observing considerable improvements in navigation performance. This includes up to approximately a 2x improvement in navigation success rates and a reduction of up to approximately 20% in path length.
Dinesh Manocha is Paul Chrisman-Iribe Chair in Computer Science & ECE and Distinguished University Professor at University of Maryland College Park. His research interests include virtual environments, physically-based modeling, and robotics. His group has developed a number of software packages that are standard and licensed to 60+ commercial vendors. He has published more than 750 papers & supervised 50 PhD dissertations. He is a Fellow of AAAI, AAAS, ACM, IEEE, and NAI and member of ACM SIGGRAPH and IEEE VR Academies, and Bézier Award from Solid Modeling Association. He received the Distinguished Alumni Award from IIT Delhi the Distinguished Career in Computer Science Award from Washington Academy of Sciences. He was a co-founder of Impulsonic, a developer of physics-based audio simulation technologies, which was acquired by Valve Inc in November 2016.