Coalition formation by multiple robots for task completion
Due to the complex nature of the real-world tasks, multiple robots need to work together towards accomplishing them. We are looking at the problem of partitioning a set of n robots into m coalitions so that these coalitions can be assigned to m tasks. The objective of the robots is to minimize the cost of moving from their initial locations to the positions, where the tasks are located.
- Dutta, A., Ufimtsev, V., & Asaithambi, A. (2018). Correlation Clustering Based Coalition Formation For Multi-Robot Task Allocation. (paper accepted at ACM SAC 2019; arXiv link)
Multi-robot informative path planning
A group of autonomous robots can be used for collecting useful information from an environment. The information can range from simple image/video to the measurement of moisture level in the soil. We are looking at the problem of dispersing the robots in a way such that maximum information from the environment can be collected. Our proposed approach will also handle real-world limitations of the robots such as communication and power constraints.
- Dutta, A., & Dasgupta, P. Bipartite graph matching-based coordination mechanism for multi-robot path planning under communication constraints. In 2017 IEEE International Conference on Robotics and Automation (ICRA), (pp. 857-862).
Modular Robotics: Configuration Formation & Locomotion Learning
During my Ph.D., I have worked on a NASA-sponsored project called ModRED (Modular Robot for Exploration and Discovery). My dissertation on modular robotics can be found here. I have primarily looked at three fundamental problems in modular robotics:
- how multiple robotic modules make decisions and form coalitions for task completion
- how the modules plan to form user-defined shapes or configurations
- how a modular robot of random shape consisting of multiple robotic modules learns to move from one point to the other.