To enable effective data acquisition and processing in open environments, sensing systems must feature behaviors that are taskable, cognizant, and collaborative: the autonomous agents that carry the sensors need to be able to translate high-level human instructions to executable strategies, exhibit high-level awareness of their own and each other’s capabilities and limitations in both sensing and actuation domains for information collection, and be able to self-organize into meta-sensors to fuse multi-modal data collected from different viewpoints. Focusing on the last two characteristics (i.e., cognizant and collaborative), the goal of the proposed work is to develop and evaluate a computationally efficient and rigorous approach for collaborative information gathering and fusion in open environments.
Mori–Zwanzig Formalism based Collaborative Estimation for Dynamic Obstacle Avoidance
We present a novel learning-based approach for dynamic obstacle avoidance in multi-robot systems using the Mori-Zwanzig (M-Z) formalism. The key innovation lies in developing a method that enables an ego robot to predict the trajectory of a dynamic obstacle that is outside its sensing range, solely by observing the historical trajectory of an ally robot. We apply M-Z formalism to rigorously justify the use of sequential data in predicting obstacle behavior, and provide theoretical bounds on the prediction accuracy. Simulation results demonstrate that our method reduces collision rates compared to scenarios without obstacle trajectory prediction. Here is the video demonstrating our proposed algorithm
Publications
- Xiaoran Zha and Mengxue Hou, “A Mori–Zwanzig Formalism based Estimation Approach for Dynamic Obstacle Avoidance”, in 22nd International Conference on Ubiquitous Robots, College Station, TX, USA, Jun. 2025, pp. 464-469, DOI: 10.1109/UR65550.2025.11078068.