Software-Based Vehicle Re-Identification

Standard loop detector controller cards sample traffic at a rate of 60Hz. However, they frequently do not push detailed information back to the traffic management center (TMC). Instead, they calculate and then aggregate values such as volumes and occupancies at up to 30-second intervals before sending this data back to the central processing system. Past approaches to loop-based vehicle re-identification have focused on creating more sophisticated controller card hardware. These approaches aim to create higher-frequency controller cards that sample at rates of 20kHz to 100kHz. While these approaches can create high resolution signatures of vehicles for matching, they are extremely difficult to deploy at large scales because they require the replacement of controller cards with new, expensive hardware at large scales. This hardware deployment in and of itself may make a loop-based vehicle re-identification scheme more expensive than a Bluetooth-based solution. This research proposes to use 60Hz samples and data fusion techniques to create a software-based signature-matching algorithm. The core strength of this approach is that it creates a clear path toward widespread deployment: it can be implemented in software, at a centralized location. Thus, agencies will be able to leverage their existing loop and controller infrastructure for re-identification.