Journal
IEEE ROBOTICS AND AUTOMATION LETTERS
Volume 7, Issue 4, Pages 8667-8674Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/LRA.2022.3184779
Keywords
Representation learning; deep learning for visual perception
Categories
Funding
- NVIDIA
- Swiss National Science Foundation (SNSF) [188596]
- National Centre of Competence in Research Robotics (NCCR Robotics)
- European Union [780883]
Ask authors/readers for more resources
We propose a learning-based method for reconstructing local terrain for mobile robots in urban environments. The algorithm utilizes depth measurements and the robot's trajectory to estimate the topography in the robot's vicinity. The method is able to accurately reconstruct the terrain despite noisy measurements and missing data.
We propose a learning-based method to reconstruct the local terrain for locomotion with a mobile robot traversing urban environments. Using a stream of depth measurements from the onboard cameras and the robot's trajectory, the algorithm estimates the topography in the robot's vicinity. The raw measurements from these cameras are noisy and only provide partial and occluded observations that in many cases do not show the terrain the robot stands on. Therefore, we propose a 3D reconstruction model that faithfully reconstructs the scene, despite the noisy measurements and large amounts of missing data coming from the blind spots of the camera arrangement. The model consists of a 4D fully convolutional network on point clouds that learns the geometric priors to complete the scene from the context and an auto-regressive feedback to leverage spatio-temporal consistency and use evidence from the past. The network can be solely trained with synthetic data, and due to extensive augmentation, it is robust in the real world, as shown in the validation on a quadrupedal robot, ANYmal, traversing challenging settings. We run the pipeline on the robot's onboard low-power computer using an efficient sparse tensor implementation and show that the proposed method outperforms classical map representations.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available