Benchmarking Augmentation Methods for Learning Robust Navigation Agents: the Winning Entry of the 2021 iGibson Challenge
This work addresses the challenge of improving embodied AI navigation in dynamic settings, such as with moving obstacles, which is an incremental advancement over static environment methods.
The study tackled the problem of training robust navigation agents for dynamic environments by benchmarking augmentation techniques, showing that adding dynamic obstacles during training significantly improved test-time generalization and achieved first place in the 2021 iGibson Challenge.
Recent advances in deep reinforcement learning and scalable photorealistic simulation have led to increasingly mature embodied AI for various visual tasks, including navigation. However, while impressive progress has been made for teaching embodied agents to navigate static environments, much less progress has been made on more dynamic environments that may include moving pedestrians or movable obstacles. In this study, we aim to benchmark different augmentation techniques for improving the agent's performance in these challenging environments. We show that adding several dynamic obstacles into the scene during training confers significant improvements in test-time generalization, achieving much higher success rates than baseline agents. We find that this approach can also be combined with image augmentation methods to achieve even higher success rates. Additionally, we show that this approach is also more robust to sim-to-sim transfer than image augmentation methods. Finally, we demonstrate the effectiveness of this dynamic obstacle augmentation approach by using it to train an agent for the 2021 iGibson Challenge at CVPR, where it achieved 1st place for Interactive Navigation. Video link: https://www.youtube.com/watch?v=HxUX2HeOSE4