Development of a Self-Powered Traffic Data Colletion System

This thesis describes the development of a novel weigh-in-motion (WIM) system that utilizes piezoelectric elements for sensing load and powering an ultra-low power microcontroller unit (MCU) that serves as its data acquisition system. A system of four piezoelectric (PZT) stacks serves as the energy harvester, while load sensing is done via a set of four PZT elements connected in parallel. Two alternative MCUs were considered, with various data handling capabilities and power consumption requirements. These MCUs have very short "wake-up" times allowing vehicle sensing without the need for inductive loops commonly used in conventional WIM systems. Special electric circuits were developed for maximizing the power output and for conditioning/sensing the voltage output. Electromechanical models were fitted to describe the relationship between voltage output, load and loading frequency/vehicle speed. Software were developed implementing the electromechanical model selected for estimating vehicle speed, axle load, number of axles and their spacing as well as vehicle classification. The software was first implemented in Matlab® and then converted to C prior to loading on the MCU. The system was tested in the laboratory by applying loads through a UTM servo-hydraulic loading system. Accuracies in measuring load, speed and class were successfully compared to the tolerances prescribed by the ASTM standard E1318.