Using Connected Vehicle Data to Evaluate Traffic Signal Performance and Driver Behavior after Changing Left-turns Phasing

There are three types of left-turn traffic signal phasing: protected, permitted, and protected-permitted. Using protected-permitted, the traffic signal displays a sequence of a static green arrow followed by a flashing yellow arrow. With protected phasing, only a static green arrow is shown. Traditionally, crash data, traffic volumes, and engineering judgement are used to select the type of left-turn operation. This approach often takes several years to accumulate statistically significant crash data, and engineering judgement can vary considerably. This study utilizes connected vehicle data to analyze the change on vehicle delay, progression, and hard-acceleration at a location where a protected-permitted left-turn was changed to a protected-only. Over 7,000 vehicle trajectories (172,000 GPS points) were analyzed. The analysis found that hard-accelerations after the stop bar decreased by 14% following the change on the left-turn phasing. However, control delay increased by 28 seconds, corresponding to a decrease in level of service from C to E for the left-turning vehicles. This paper contributes a methodology that can assess efficiency and safety surrogate measures that can be scaled in a systematic manner. This scalable technique is particularly valuable for agencies that want to assess operations and/or prioritize improvements across hundreds or thousands of signals. In this study, the effects on vehicle behavior and signal performance after changing the eastbound (EB) left-turn fixed-time phasing at the intersection of US-40 and German Church Rd., located east of Indianapolis (Fig. 1, callout ii), are analyzed.