Sensor networks for remote monitoring of public transport vehicles

ICT technologies provides non-invasive and low cost solutions to improve safety and security in public transport vehicles for persons and freights. In this paper we consider an intelligent heterogeneous sensor network installed on the vehicle to monitor its status. The network can be remotely interrogated by operators in the command and control centre (C&C) that, in turn, can be automatically notified by the same network when a critical condition is assessed. To reduce the signaling overload on the mobile network connecting vehicles to the C&C, an intelligent on-board controller analyzes data from sensors, infers on the vehicle status and reacts properly. In accordance to the severity of the event, it decides whether the situation can be resolved at local level or if it is necessary to alert the C&C. This strategy permits to significantly reduce the amount of information sent over the public radio network. We express the performance of the considered solution in terms of the blocking probability due to (unavoidable) false alarms generated by vehicles as a function of the false alarm rate characterizing the on-board network. Results are used to assess the maximum tolerated false alarm rate guaranteeing a desired level of blocking probability for a given number of mobile radio channels.