A research group consisting of Fujitsu Laboratories and Professor Shingo Takahashi of Waseda University's Faculty of Science and Engineering has developed a technology to automatically analyze the causes of congestion from the results of human behavior simulation in a short time.With the introduction of this technology, congestion mitigation measures at event venues, commercial facilities, transportation terminals, etc. can be quickly discovered.
According to Fujitsu Laboratories, human behavior simulations that model human behavior are used for predicting evacuation behavior in emergencies and confirming movement lines during city planning.Of these, in the simulation of congestion prediction, since experts verify the causes one by one from a large amount of results, it takes a lot of time and oversight of the causes has become a problem.
Therefore, the research group has developed a technology to extract features by grouping by common items and combining a small number of items, instead of listing thousands to tens of thousands of simulation results as items one by one.
For example, if two stores A and B of a commercial facility are crowded, it is assumed that the reason for the congestion of store A is the effect of attracting customers on the information board, and that store B is caused by a large number of restaurant users.In this case, it can be seen that it is effective to install a guide board to guide customers to another destination in store A, and to improve the response speed by increasing the number of staff in store B.
When this technology was applied to human behavior simulation developed in 2015 for the purpose of alleviating congestion at airports, it was found that the cause of congestion was about four times that of expert analysis.The analysis time can be significantly reduced from months to minutes.