Multimodal navigation

Here we demonstrate how a high level Reinforcement Learning trained agent can avoid a pilar using vision while walking toward an odour source.