Rey Wiyatno - Lifelong Topological Visual Navigation

Rey Wiyatno (MSc, U. de Montréal) Supervision : Liam Paull

The ability for a robot to navigate with only the use of vision is appealing due to its simplicity. Traditional vision-based navigation approaches required a prior metric map-building step that was arduous and inefficient. An alternative is to build a topological map which still preserves some loose connectivity of the space, but is prone to failure due to spurious edges in the graph. Newer learning-based navigation techniques directly learn policies from image inputs, but tend to struggle in long-distance navigation tasks. Here, we propose a learning-based topological visual navigation method with graph update strategies that improves lifelong navigation performance.