Towards Target-Driven Visual Navigation in Indoor Scenes via Generative Imitation Learning
This addresses the problem of enabling robots to navigate autonomously in indoor environments for applications like home assistance or logistics, though it is incremental as it builds on existing imitation learning methods with specific enhancements.
The paper tackles target-driven visual navigation in indoor scenes without maps, odometry, or GPS, by proposing a generative imitation learning system with three key designs that improve training efficiency, collision avoidance, and generalization, resulting in a model successfully tested on a TurtleBot in real-world settings.
We present a target-driven navigation system to improve mapless visual navigation in indoor scenes. Our method takes a multi-view observation of a robot and a target as inputs at each time step to provide a sequence of actions that move the robot to the target without relying on odometry or GPS at runtime. The system is learned by optimizing a combinational objective encompassing three key designs. First, we propose that an agent conceives the next observation before making an action decision. This is achieved by learning a variational generative module from expert demonstrations. We then propose predicting static collision in advance, as an auxiliary task to improve safety during navigation. Moreover, to alleviate the training data imbalance problem of termination action prediction, we also introduce a target checking module to differentiate from augmenting navigation policy with a termination action. The three proposed designs all contribute to the improved training data efficiency, static collision avoidance, and navigation generalization performance, resulting in a novel target-driven mapless navigation system. Through experiments on a TurtleBot, we provide evidence that our model can be integrated into a robotic system and navigate in the real world. Videos and models can be found in the supplementary material.