Developing a Portable Data Acquisition System to Study Road User Behavior

Abstract

Data on driver behavior under various circumstances are essential for understanding traffic and quantifying traffic safety. Traditional data collection approaches (such as using driving simulators and video-graphic methods) possess various drawbacks. Recent advances in technology make instrumented vehicles (IV) affordable. In such an experimental setup, a portable data acquisition system (PDAQS) is equipped with various sensors, including light detection and ranging (LiDAR) sensors, to capture drivers’ behavior when they are driving the vehicle as naturally as possible. A PDAQS helps to accurately observe instantaneous vehicle information and driver behavior for longer road sections and durations. A PDAQS can be used as a probe unit to capture drivers’ driving behavior around the IV. The data that can be obtained accurately at high frequency include changes in headways and speeds and interactions among road users. Because the installation is portable, the same system can be used among different vehicles and locations (such as urban, rural, and different states). In this project, we assembled a PDAQS and developed software to process the collected data. The tasks included developing the hardware and software programs to extract valuable data. On the software side, we developed a modularized framework capable of efficiently processing LiDAR and video data and detecting objects in two and three dimensions. Our main results were an alignment tool that adjusts and tests point transformation parameters, LiDAR/camera timing synchronization in real time, and the integration of point cloud detection into our framework. Although lots of progress was made, a additional future work must be done before the product can be applied to real-time projects.