Autonomous Navigation Through Crowds
Jonathon Volkmar · 2011
This research topic of this paper addresses the goal of enabling a mobile robot to navigate autonomously in densely populated environments. The main contribution of this report is the presentation of a framework for feeding real world data to a maximum entropy inverse reinforcement learning algorithm, for modelling socially acceptable navigation behavior, and the presentation of some preliminary experimental results. The data collection is done using a low cost Kinect RGB-D camera.