Blind spots created by the design of heavy vehicles make it difficult for operators to detect nearby vulnerable road users (VRUs). Visibility detection technologies are designed to reduce and mitigate crashes between heavy vehicles and VRUs by increasing the visibility of VRUs. Sensors installed around the vehicle monitor blind spots and provide audible and/or visual warnings to notify drivers if a VRU is detected nearby.
Measure the VRU detection system performance on extended use under real operating conditions in Canada and evaluate driver acceptance and experience using the system.
Transport Canada sought to address the risk posed to VRUs around heavy vehicles due to blind spots with an assessment of the safety potential of new VRU detection systems, the most promising countermeasure identified by collisions investigations. In 2018, Transport Canada conducted a year-long field operational test in Edmonton, Hamilton, Toronto, Ottawa, and Montreal. Fourteen vehicles were equipped with different camera/sensor systems and driver-vehicle data loggers to capture the performance of the detection systems under natural driving conditions. Vehicles differed in size and design and were driven on different road types in a variety of weather, traffic and operations.
The Mobileye Shield+ is a smart camera VRU detection system that uses a two-staged alert to increase driver awareness. The two-staged system was designed to provide drivers with more time to respond once the system issued either an audio or visual alert to the driver depending on the proximity of the VRU.
Transport Canada evaluated the acceptance and experience of the VRU detection system by analyzing data collected during the field trial, data collected from self-report questionnaires and responses received from the follow-up debriefing.
Data and lessons learned
Lau et al. (2020) found that system data showed promise when it came to successfully alerting operators of VRUs in proximity of the vehicle. Although the warnings from the VRU detection system were abundant, the operators’ experiences using the system reported on the questionnaire and debriefing gave further insight in the disadvantages of the system.
As noted by the operators’ responses, auditory alerts issued by the system were loud and annoying. It was determined that failure to address the alert and annoyance when designing for auditory signals can increase workload, become distracting, or even cause drivers to disable the warnings altogether. According to the final report, these consequences were all evident in the field operational test.
Although the general impression of using the detection system was negative, there were some added advantages noted from the responses. Drivers felt that the system was effective at capturing their attention (via audio and light) when warnings were issued and thus raised their awareness to their surroundings. However, this benefit was outweighed by the increasing number of false alarms issued. The high rate of false alarms left a negative impression of the system as a whole, making it hard for drivers to accept the technology.
The debriefing made it possible to understand that acceptance of a detection system relied heavily on its ability to be customizable for the vehicle and operational tasks at hand. For example, operating heavy vehicles at night did not receive any of the benefits provided by the detection system because it was not designed for this circumstance.
In conclusion, the study found that distraction, system unreliability, and failure to apply the system to different vehicles and tasks were all common factors that decreased operator acceptance of the VRU detection system. The challenge was that the system itself could not determine which alarms it issued were false positives or true detections. Future field studies could record confirmation data from the operators to verify the accuracy of every alarm.
Prior to the field trial, Charlebois et al. (2019) conducted in-depth assessments to identify suitable systems for more extensive long-term testing. Five different systems were subjected to a series of simulated “urban environment” scenarios on a closed track to evaluate the accuracy and reliability of alerts to drivers based on a time-to-collision warning, as well as the number of false positive detections. These test scenarios were based on the most common real world VRU-truck collisions identified from collisions investigations reports. The tests revealed that:
- A VRU detection and warning system which provided a two-staged alert tended to be more effective.
- The smart camera system, composed of three cameras located around the vehicle providing audio and visual signals to the driver, performed best overall.
- The smart camera warning system helped to increase driver awareness and provided additional time for drivers to assess the situation and respond.
- Alerting the operator too frequently, or when it might not be needed, as demonstrated with the false positive scenarios, was an important factor when collecting information about a system. The smart camera system was selected for more extensive testing on different vehicles in the multi-city field operational test due to its capabilities to select when or not to warn the operator.
The report indicates that as a look to the future, a system that works with the operator to identify risk constantly and in a predictable fashion may better complement the current mirrors and training provided to vehicle operators.
Transport Canada will continue to conduct research with Advance Connectivity and Automation in the Transportation System (ACATS) program. Phase II of the ACATS program took place in 2020.