The purpose of our lab's research program is to advance visual navigation of mobile robots. Our work finds application in transportation, planetary exploration, mining, warehouses, offices, and military scenarios.

Over the last 20 years, our lab has spent a lot of time building and testing different navigation approaches in the field. Much of our work is focused on a navigation stack we pioneered called teach and repeat (T&R). T&R has been particularly interesting in that it allows a robot to repeat a long (several kilometre) route that was taught manually, using only a single core sensor (stereo camera, lidar, radar) for feedback (no GPS needed). T&R has been successful because it avoids the need to construct a map of the world in a single privileged coordinate frame and instead utilizes a topometric map. We also spent a lot of time improving the robustness of visual localization in the presence of lighting and seasonal change.

Today we are continuing to expand the possibilities for T&R-style navigation as well as the underlying state estimation tools upon which it relies. We are looking at the idea of cross-modal teach and repeat where one type of vehicle or agent teaches a route for another; for example we might fly a drone over an environment then expect a ground vehicle to execute the route. We are also looking at multi-robot teach and repeat where several vehicles cooperatively teach and then repeat routes in a coordinated fashion; this has applications in both space exploration (coordinated cargo transport) and defence. We are exploring how machine learning and foundation models can be used to improve our perception, primarily with an eye towards improving assessment of driveable surfaces around a robot. We continue to investigate the benefits of different sensing modalities including Doppler sensors such as FMCW radar but also FMCW lidar, both of which can natively produce velocity measurements. On the state estimation front, we are always interested in fun new mathematical ideas that advance our core localization and mapping tools; some current topics include certifiably optimal estimation, data-driven Koopman estimation, and advanced Lie groups. This blurb is inevitably already out of date so please refer to the papers below for the latest!