Application of Indoor-movement-path prediction to controlling communication systems by online analysis

摘要

As a result of the continuing increase in the popularity of smartphones as well as devices connected by the "Internet-of-Things" (IoT), sometime in the future, there is the risk that it will not be possible to secure sufficient communication resources for the number of connected communication devices. To effectively utilize limited communication resources, it is useful to predict the movement of people. As for many conventional technologies, to predict movement of people, it is necessary to obtain log data in advance. However, it is a problem that acquiring movement logs takes time and incurs a cost. Accordingly, in the present report, a system for predicting movement from logs acquirable online and utilizing the prediction result for communication control is proposed. On the basis of online logs, the proposed system assigns a score to the interest a target person has for particular categories and uses that score to calculate the expected utility obtainable by that person at each location they visit. The movement path that maximizes the sum of the calculated expected utilities is taken as the prediction result, and the degree of congestion along the link between each location is calculated. That degree of congestion is then utilized for controlling communication resources. In this report, the results of calculating expected utilities for each location by questionnaire (instead of online logs) and predicting movements of target persons are presented. The effectiveness of calculating degree of congestion along a link between locations on the basis of the movement-prediction results and controlling communication resources by predicting peoples' movements on the basis of online logs is demonstrated.