Early Prediction of Stampede In Assemblage

Abstract

Early prediction stampede techniques are essential in public assemblages or gatherings such as cricket match stadiums, auditoriums, football stadiums, religious crowds etc. It reduces the chance of stampedes in assemblages and eliminates the chances of casualties of large numbers of people. The video analytics researchers have mainly worked to detect the stampede effectively but have yet to perform exact detection when the headcounts are enormous. Thus, sensor-based technique plays a vital role in detecting the stampede during assemblages. The wristband prototype was randomly distributed among participants at crowded events, and data was gathered through the use of the LoRa communication protocol. The collected data was analyzed using the support vector machine (SVM) algorithm to determine the risk of a stampede. The results of the study demonstrate the effectiveness of the prototype in reducing the likelihood of stampedes in public gatherings, thereby providing organizers with valuable information to promote the safety of their attendees. In doing so, we use an Inertial Measurement Unit (IMU) sensor, heartbeats sensor, noise sensor, humidity sensor, and temperature sensor for collecting data and sending it to the server node. Server analysis of the data and predict the chances of a stampede in the public gathering. Server nodes monitor the activity dynamically using a trained machine-learning model to predict the chances of stampedes in public gatherings. The experimental analysis via simulation and developed prototype depicts the effectiveness of the proposed technique. . © 2023 IEEE.