Development and Experimental Validation of a LoRa Wireless Sensor Network for Wildfire Surveillance

This paper describes and evaluates a surveillance system based on Wireless Sensor Networks (WSNs) and Autonomous Vehicles (AVs). Firstly, the main wireless communication technologies are surveyed, resulting in the choice of the Long Range (LoRa) technology to support the communication of the sensor nodes. In order to better deal with the characteristics of the LoRa communication, a protocol is designed and tested in realistic experiments in both rural and urban settings, which allows researchers to have a comparison between manufacturer values and realistic ones. Given the intended objective of having the surveillance system applied to remote areas to detect malicious people entering private property or forest fires, autonomous vehicles are used to further enhance the detection capabilities. The overall system is implemented in ROS packages and documented for future use. In experiments at the Institute for Systems and Robotics (ISR) flying arena, the system achieved a good average tracking Root-Mean-Square-Error (RMSE) of 0.0381 m and inspected all events that crossed the motion detection sensors with a triggering delay of 133 ms.