Search Content

Flaggers protect workers by providing temporary traffic control and maintaining traffic flow through a work zone. They are often the first line of defense to stop distracted, inattentive, or aggressive motorists from intruding into the work area. This project aims to develop an automated intrusion detection system to alert drivers who are unsafely approaching or entering a flagger-controlled work zone. A human factors user needs assessment found maintenance workers preferred a modified traffic signal to feature the alert system due to flagger risks of being in the roadway and drivers failing to stop and remain stopped when presented with the STOP side of the flagger sign. A modified traffic signal that could be operated using a handheld remote was developed. The low-cost embedded electronics on the traffic signal enabled it to track trajectories of nearby vehicles, detect potential intrusions, and trigger audio-visual warnings to alert the intruding driver. Usability testing in a simulated driving test found poor expectancies and stopping rates of the traffic signal-based alarm system compared to a traditional flagger but did demonstrate evidence that drivers may be less likely to stop and remain stopped with the flagger STOP sign than the red ball indicator of the traffic signal. Furthermore, some drivers corrected their initial stopping error after triggering the auditory alarm of the traffic signal. A follow up test found improved performance with the alert system incorporated into an audiovisual enhanced STOP/SLOW flagger paddle. Testing of the developed sensor system found the system capable of simultaneous multi-vehicle tracking (including estimation of vehicle position, velocity, and heading) with a range of up to 60 meters and angular azimuth range of 120 degrees and correctly detecting all test intruding vehicles.

Pedestrian deaths are at a 30-year high nationally, accounting for 16% of total deaths in 2018 and far exceeding the previous decade of 12%, a trend mirrored in Minnesota. Previous research found an increase in local and citywide yielding at unsignalized crosswalks following an engineering and high-visibility enforcement program in Saint Paul, Minnesota. This study examined a modified engineering-focused (i.e., without enforcement) program expanded to both unsignalized and signalized intersections across the Twin Cities. The six-month study found modest improvements in yielding from baseline to treatment end (48.1% to 65.5% in Saint Paul and 19.8% to 38.8% in Minneapolis) at unsignalized engineering treatment sites but no improvements at generalization sites. No significant improvements in left- or right-turning yielding by drivers in Saint Paul were found at treated signalized intersections, but given that yielding was significantly worse at generalization sites over time, there may be some evidence that treatments mitigated performance declines among Saint Paul drivers during the study period. Yielding improvements at signalized treatment sites were more pronounced for only right-turning drivers in Minneapolis, but generalization sites showed no improvement or even worsened over time. Overall, study results suggested no shift in driving culture in either city, as found with the previous study using police enforcement, but found some evidence of local, site-specific changes in driver yielding behavior at treatment locations.

As more automated vehicles enter shared roadways, an essential aspect of automated vehicle (AV) safety is understanding the interactions between these vehicles and other road users. Anecdotal incidents about aggressive following and overtaking behaviors at crosswalks near the Med City Mover (MCM), a low-speed automated shuttle (LSAV) pilot demonstration in Rochester, MN, suggested the need for a scientific study of the behaviors of drivers of manual vehicles near the LSAV. In this report, the research team conducted a series of laboratory and field studies aimed at better understanding the safety relationship between LSAVs and the humans with whom they share the road. Overall, the studies found an increased risk of overtaking and multiple threat passing near the MCM, which may increase the risk of pedestrian-involved crashes, sideswipe crashes, and rear-end crashes. Study findings suggested that poor human-machine interfaces, exceptionally slow vehicle speeds, and resultant large queues behind the MCM contribute to these risks. Improved communication interfaces, speeds more consistent with the surrounding traffic, and smaller queue size will be important factors that AV developers and future pilot demonstrations must consider to better promote pedestrian safety near AVs.