The University of Western Ontario

Published purpose

When rovers explore other planets or celestial objects such as the Moon to collect scientific data, they are programmed to find rocks and other materials based on basic features like shape and texture. During these missions, the connection between rovers and scientists on Earth is often limited by data bandwidth. Mission success would be improved if rovers could be trained to make scientific decisions autonomously and be less reliant on this data connection. This project will use laboratory studies and field trials to test “neural networks” that help rovers learn how to carry out scientific investigations independently. This includes teaching rovers to make decisions about where to explore, what information to collect, and what data to share with scientists on Earth. The project will position Canada as a leader in autonomous scientific investigation systems. It will also provide technology that can be used in mining and farming on Earth.