Obstacle Avoidance through Deep Networks based Intermediate Perception
This addresses the problem of robust obstacle avoidance in textureless environments for robotics, though it is incremental as it builds on existing learning-based methods with a two-stage approach.
The paper tackles obstacle avoidance for robots from monocular images by proposing a method that predicts trajectories through intermediate perception of depth maps and surface normal, increasing accuracy by 20% compared to direct prediction and demonstrating generalization to other datasets and robot flights.
Obstacle avoidance from monocular images is a challenging problem for robots. Though multi-view structure-from-motion could build 3D maps, it is not robust in textureless environments. Some learning based methods exploit human demonstration to predict a steering command directly from a single image. However, this method is usually biased towards certain tasks or demonstration scenarios and also biased by human understanding. In this paper, we propose a new method to predict a trajectory from images. We train our system on more diverse NYUv2 dataset. The ground truth trajectory is computed from the designed cost functions automatically. The Convolutional Neural Network perception is divided into two stages: first, predict depth map and surface normal from RGB images, which are two important geometric properties related to 3D obstacle representation. Second, predict the trajectory from the depth and normal. Results show that our intermediate perception increases the accuracy by 20% than the direct prediction. Our model generalizes well to other public indoor datasets and is also demonstrated for robot flights in simulation and experiments.