Accuracy improvement method for vehicle detection using optical sensors

Abstract

Deploying wireless sensor networks for detecting and counting vehicles is not a new concept. One disadvantage of most simple wireless sensor for detecting vehicles is wrong interpretation of complex vehicle such as a trailer truck with multiple containers. In this paper, we propose an improvement of a simple wireless sensor network for detecting vehicles developed by Chinrungrueng et al. This system consists of two optical sensors, which are able to detect small and medium sized vehicles. However, the system cannot distinguish a truck with several containers. If only two sensors are used without modification, the truck will be detected as multiple vehicles. In this paper, we present techniques for obtaining vehicle speed, length and gap length between two consecutive vehicles. These additional parameters can be utilized to distinguish a truck from multiple, medium sized vehicles. We propose a decision making rule to differentiate various type of vehicles and scenarios. The accuracy of the decision making rule depends on the thresholds of the parameters.Index Terms-wireless sensor networks, vehicle detection.