Advisor Information
Prithviraj Dasgupta
Location
UNO Criss Library, Room 249
Presentation Type
Oral Presentation
Start Date
3-3-2017 9:00 AM
End Date
3-3-2017 9:15 AM
Abstract
Autonomous exploration using multiple robots is an important area of research with applications to extraterrestrial exploration. The use of robots to explore environments can reduce the dangers to human explorers in unstructured environments, such as after a disaster or on extraterrestrial planetary surfaces. In these scenarios, the environment is unknown or only coarsely known beforehand, and there is no pre-existing communications infrastructure to use. We propose an algorithm that balances the goals of communicating collected samples back to a base station, collecting samples from areas of the environment which are less known, and spending as little energy to do this. Our proposed algorithm uses Gaussian Processes to model communications and distribution of information in the environment. Points are selected from the Gaussian Processes and selected for travel for the robot based on a utility function that weights the three goals of the robot.
Multi-Robot Informed Path Planning Under Communication Constraints
UNO Criss Library, Room 249
Autonomous exploration using multiple robots is an important area of research with applications to extraterrestrial exploration. The use of robots to explore environments can reduce the dangers to human explorers in unstructured environments, such as after a disaster or on extraterrestrial planetary surfaces. In these scenarios, the environment is unknown or only coarsely known beforehand, and there is no pre-existing communications infrastructure to use. We propose an algorithm that balances the goals of communicating collected samples back to a base station, collecting samples from areas of the environment which are less known, and spending as little energy to do this. Our proposed algorithm uses Gaussian Processes to model communications and distribution of information in the environment. Points are selected from the Gaussian Processes and selected for travel for the robot based on a utility function that weights the three goals of the robot.