Monocular SLAM refers to using a single camera to estimate robot ego motion while building a map of the environment. While Monocular SLAM is a well studied problem, automating Monocular SLAM by integrating it with trajectory planning frameworks is particularly challenging. This paper presents a novel formulation based on Reinforcement Learning (RL) that generates fail safe trajectories wherein the SLAM generated outputs do not deviate largely from their true values. Quintessentially, the RL framework successfully learns the otherwise complex relation between perceptual inputs and motor actions and uses this knowledge to generate trajectories that do not cause failure of SLAM. We show systematically in simulations how the quality of the SLAM dramatically improves when trajectories are computed using RL. Our method scales effectively across Monocular SLAM frameworks in both simulation and in real world experiments with a mobile robot.
@article{arxiv.1607.07558,
title = {Learning to Prevent Monocular SLAM Failure using Reinforcement Learning},
author = {Vignesh Prasad and Karmesh Yadav and Rohitashva Singh Saurabh and Swapnil Daga and Nahas Pareekutty and K. Madhava Krishna and Balaraman Ravindran and Brojeshwar Bhowmick},
journal= {arXiv preprint arXiv:1607.07558},
year = {2020}
}
Comments
Accepted at the 11th Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP) 2018 More info can be found at the project page at https://robotics.iiit.ac.in/people/vignesh.prasad/SLAMSafePlanner.html and the supplementary video can be found at https://www.youtube.com/watch?v=420QmM_Z8vo