Active sensing and inference for robotic environment monitoring

Both hazardous contaminations and anthropogenic emissions are closely intertwined with public health and long-term climate change. Effective and prompt pollutant dispersion monitoring is of paramount importance in assessing emission sources, understanding contamination dispersions, and predicting the whereabouts of hazardous airborne materials.

This project aims to develop autonomous mobile robots to rapidly respond to such scenarios by providing real-time, high spatiotemporal awareness of pollutants, hazardous materials or greenhouse gases for a targeted event/region at site-specific level. The robotic systems are often required to operate in GPS-denied environments without a prior map or reliable communication for teleoperation. Therefore, the focus of this project is to develop an active sensing and inference framework which would allow the robot to establish a better understanding of the environment by autonomously exploring the area and actively collecting environmental data, while maintaining reliable simultaneous localization and mapping (SLAM). The overall system will be deployed on real robotic systems and tested in realistic operational environments.