This project was aimed at developing and implementing a ready-to-commercialize prototype to obtain automated turning movement counts at intersections, regardless of lane configuration. In a previous IDEA project (NCHRP-177), a vehicle trajectory classification algorithm was developed to demonstrate the feasibility of automatically classifying vehicle trajectories into movements regardless of the lane configuration. In this follow-on project, the previously developed algorithm has been improved to increase the accuracy of the vehicle movements’ classification and to make the algorithm ready for commercialization. Improvements to the algorithm focused on two areas: first, improving the vehicle count by more reliably recognizing the vehicles that crossed the stop bar of an approach, and; second, by using the lane used by vehicles and lateral displacement to assign a movement to vehicle trajectories. A data collection device was developed, capable of logging vehicle trajectories from intersections instrumented with a commercially available radar-based vehicle detection system. The device can be installed inside a signal cabinet and is independent of the controller platform. The commercialization partner released an initial version of the data collection and analysis product in the Summer of 2019. The data collection device implements the key noise removal techniques described in this report and can be a platform for future performance monitoring techniques. Furthermore, it implements procedures to breakdown vehicle volume at an intersection approach by lane. As a result, adding classification by movement based on a streamlined version of procedures developed as part of the previous IDEA project (NCHRP-177) can be accomplished via a software update. Once the necessary commercialization steps with the licensing arm of the University of Wisconsin-Madison are cleared, the commercialization partner plans to integrate a streamlined version of the classification algorithm into their product via a software update. Adding the classification algorithm to the commercialized data collection product as a software update will be possible because of the joint effort between the commercialization partner and the research team to define a strong and flexible underlying software architecture and data storage techniques.