Abstract
Abstract. In this research, we investigate the use of Reinforcement Learning (RL) for an effective and robust solution for exploring unknown and indoor environments and reconstructing their maps. We benefit from a Simultaneous Localization and Mapping (SLAM) algorithm for real-time robot localization and mapping. Three different reward functions are compared and tested in different environments with growing complexity. The performances of the three different RL-based path planners are assessed not only on the training environments, but also on an a priori unseen environment to test the generalization properties of the policies. The results indicate that RL-based planners trained to maximize the coverage of the map are able to consistently explore and construct the maps of different indoor environments.
Highlights
Simultaneous Localization and Mapping (SLAM) is the given name for the mathematical methods aiming at simultaneously building the map of the environment and estimating the pose of the given sensor within it
One of the first implementation can be attribute to (Smith et al, 1990) and, only later, SLAM algorithms started to be developed for mobile robotics (Montemerlo et al, 2002)
When the SLAM algorithm is used in real-time for actively planning robot paths while simultaneously building the environment map, the SLAM algorithm is named as Active SLAM (Trivun et al, 2015)
Summary
Simultaneous Localization and Mapping (SLAM) is the given name for the mathematical methods aiming at simultaneously building the map of the environment and estimating the pose of the given sensor within it. We look at the control signal generation and state transition of SLAM as solving a RL problem, where the agent has to learn the best sequence of actions (the best sequence of control signals), i.e the sequence that maximizes the total cumulative reward In this context, the goal is to explore and complete the map without collisions in the minimum time. RELATED WORK 2.1 Robot Navigation and Active SLAM with RL In recent years, RL has been used to learn path planning skills for autonomous robot navigation in unknown environments These methods don’t build and use any map for navigating in the different environments, e.g. I=0 where M is the total number of grid cells in the map
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
