Research On Group Identification And Monitoring Based On Mobile Phone Sensors

A pedestrian flock is a collection of people gathering together and walking along similar routes.In fact,people often walk in flocks which are widely present in various social activities.The analysis of flock relationships can help researchers understand the patterns of interaction among pedestrians,thereby further exploring the social relationships among pedestrians and applying them to various fields.For example,based on flocks,merchants can send more targeted promotional information to users.As for social events such as large gatherings,pedestrian flocks also provide a basis of decision making for emergency evacuation plans.With the popularity and development of smartphones,it is easier to obtain information reflecting characteristics of pedestrian position and movement by using built-in sensors.These data can help researchers better analyze the potential relationship between pedestrians and thus accelerate the development of the research in pedestrian flocks.In recent years,most of the work for flocks is to use the location information of pedestrians as a necessary condition,supplemented by relevant motion data to mine the hidden flocks,so as to achieve the identification and tracking of groups.GPS sensor built in the mobile phone is the first choice for researchers to get the necessary location data of the research because it can get location information of people in various scenes of daily life.However,the huge energy consumption caused by the long-term and continuous opening of sensors such as GPS is one of the challenges in this field.To address the challenges above,this paper proposes an energy-efficient pedestrian flock identification and monitoring method,which can largely reduce the power consumption of mobile phones without decreasing the accuracy of the existing flock detection method.This study first uses GPS sensors to collect latitude and longitude data and identify flocks among pedestrians.Once it is determined that certain pedestrians are stable in a flock,we immediately turn off GPS sensor of the mobile phones held by these people and simultaneously turn on two inertia sensors with low power consumption such as acceleration and gyroscope,using the movement characteristics of pedestrians within the flock to monitor group status.The main contributions of this paper are divided into the following three points.(1)As for the spatial-temporal clustering method used in group recognition stage,we propose a method to calculate the stability of group structure in continuous time period,and enrich the determination condition of group stability with sociological attribute to improve the recognition efficiency of stable group.(2)According to the acceleration and gyroscope sensor data,we propose a group monitoring method based on anomaly detection technology,which is the key point to reduce the power consumption of smartphones in this paper.(3)We build the data acquisition platform,design and test the relevant experiments.By analyzing the results of these experiments in many aspects,it is proved that the method proposed in this paper can not only get a slightly better accuracy of group detection compared with the existing research,but also highly reduce the power consumption of smartphones.