Informative path planning for exploration and mapping of unknown environments using multiple robots

Using robots to intelligently explore and map an unknown environment has been an active research area in the robotics community over the last decades. Such a capability can play an important role in many applications ranging from automated vacuum clearing to search and rescue and emergency response. While existing methods mainly focus on learning the geometric map of the environment, there is another need of collecting other critical information of interest, such as chemical, biological and radiological contamination, during the exploration of the unknown area. The aim of this project therefore is to develop a multi-robots-based framework and algorithms to autonomously and efficiently gather information in an unknown environment.

The project will first design informative path planning algorithms not only for simultaneous localisation and mapping using traditional sensors like LiDAR and/or camera, but also for gathering useful information based on other sensor cues. Then, the collaboration strategy among a team of robots will be developed, so that they efficiently solve an exploration task. Last but not least, the project will investigate learning techniques to make inference about the incident of interest (e.g. finding the source of a release) based on collected information to support decision-making process.

It is envisaged that this project may provide effective solutions to various aspects of the autonomous map learning and information gathering problem and has the potential to be deployed in real world applications.