A novel approach to provide safe indoor industrial environment
2016 International Conference on High Performance Computing and Simulation, HPCS 2016
Human tracking, Motion recognition, Safe workplace, Wireless Sensor Network
In this paper, a new method to have a safe place by human motion recognition and routing in an indoor environment, in presence of different objects, such as on a manufacturing shop floor, is presented. Due to the high cost and limitations of some of the available methods such as video surveillance, limitations in the indoor because of the presence of other objects, processing time, and limitation of data storage, a low cost, low power, and high accuracy approach using Wireless Sensor Network (WSN) is proposed. In this proposed method we collect data from indoor area using nodes equipped by sensors, and use them for processing. Features are extracted from signals which are sensed by sensors and applied for motion recognition. Subsequently, they are processed to recognize motion by Hidden Markov Model (HMM) classification. HMM are tested along with an investigation of their accuracy. Rules are extracted by the routing and motions. The presented method is evaluated for three distinct motion categories that are common in indoor environment work places. Experimental results show the high accuracy is obtained for the proposed approach for the case under study which was at 100%.
Anvaripour, Mohammad; Saif, Mehrdad; and Ahmadi, Majid. (2016). A novel approach to provide safe indoor industrial environment. 2016 International Conference on High Performance Computing and Simulation, HPCS 2016, 544-548.