Collecte of Data Stemming from the Fine Trajectory of the Pedestrians

Abstract

Walking, the oldest means of transport and practiced used by everyone, represents the “heart of life itself” aroundwhich any othermeans of transport should be subordinately organized. It is necessary to detect, locate, identity, and eventually track pedestrians to understand how they interact with each other. This article concerns the detection of pedestrians to using a laser sensor. This sensor is located approximately 18cm above the ground and collects information about in-motion feet distances of pedestrians and the presence of objects within 20m radius. A walking model based on the typical foot movements is defined in order to exclusively extract data with the in-motion foot hallmarks. Each laser sensor is con- trolled by a client computer that sends the information to a server computer, where they are spatially and temporarily integrated in a global database.Atrackingmethod is developed using theMCMCDA(MarkovChainMonteCarloDataAssociation) to monitor pedestrians’ trajectories. The system is evaluated by computer simulation and a full-scale experiment is being developed. The simulation is conducted in a school entrance hall where eight laser sensors cover an area of about 1,500m2.The position and direction changes of students during the break where the rush hour, are analyzed. Computer simulation is carried out to examine quantitatively the system performance with respect to the density of the crowd.