Evaluation of Optimization for Pedestrian Route Guidance in Real-world Crowded Scene

Shusuke Shigenaka (Artificial Intelligence Research Center AIST)

Abstract

In this paper, we proposed evaluation index considering safety of pedestrian. This evaluation index, it is possible to evaluate all of pedestrian traveling time, unfairness and congestion degree. We also confirm that the guidance control method optimized by CMA-ES can realized better than real guidance.