Best Paper Award

  In order to improve event quality and to maintain visitor safety in an outdoor events, it is important to manage the number of visitors in event sites. On the other hand, a low-cost doppler sensor has become available these days to detect the approach and distance of people.
  Therefore, in this paper, we propose a new population monitoring system utilizing the doppler sensor so that administrators of events can know the number of visitors in real time. In order to measure the degree of congestion in event sites, we define a new metric of population scale which indicates not only the number of people but also the amount. The population scale can be calculated from the sensor data of the doppler sensor by approximating the relationship between the population scale and the sensor data based on a Gompertzian curve.
  In order to clarify the effectiveness of the proposed system, we performed the experimental evaluation of population monitoring in Echigo Hillside National Government Park in Niigata prefecture when several events (e.g., illumination, flower viewing) were held. In the illumination event, three feature quantities (i.e., the most crowded time, crowded period and the maximum population scale) were extracted by fitting the sensor data to the mixed normal distribution by using the EM algorithm. Although the most crowded time and crowded period did not change during the event period, it was clarified that the maximum population scale was markedly different between days. In addition, there was negative correlation between the population scale and the lowest temperature, which means that many visitors can be expected in days when the temperature is low. Furthermore, we couldn’t find any correlation between the population scale and weather conditions in the flower festival, but the ability of the event to attract customers can be quantified by the proposed system because the population scale certainly decreased after the event was finished.
  As mentioned above, this paper proposes a new sensing system and an analysis method of sensor data for estimating population scale, and explains the results of experimental evaluation for clarifying effectiveness. This paper indicates a significant direction for performing practical research and development, hence it is highly deserving of an IEICE Best Paper Award.