Autonomously explore unknown environments and building maps is an essential skill for mobile robots. However, when the environment is unknown a priori, it is challenging to decide which target locations has to be explored first. Many approaches and algorithms have been proposed in literature to solve this problem, however, we will focus on improving the approach proposed in [1]. In this assignment, we focus on solutions involving Extended Kalman Filter and Reinforcement Learning.
References:
[1] Trajectory Optimization using Reinforcement Learning for Map Exploration, Kollar T. and Roy N
Learning to explore and map with EKF SLAM and RL
Finished: 2021-08-25
MSc assignment