EU-funded study pilots Intelligent Transport Systems (ITS) to improve the safety and mobility of vulnerable road users (VRU) through ITS applications.
Helmond, Netherlands
Improving the safety and mobility of vulnerable road users through ITS applications: Deliverable D5.3 Evaluation Results
Summary Information
The VRUITS project, which is sponsored by the European Commission DG MOVE, places the Vulnerable Road User (VRU) in the center, assesses the impact of current and upcoming ITS applications on the safety and mobility of VRUs, identifies how the usability and efficiency of ITS applications can be improved, and recommends which actions have to be taken at a policy level to improve ITS safety and mobility. By applying a multi-disciplinary approach, the VRUITS project aims at developing tools to evaluate, field-test and subsequently improve ITS for vulnerable road users.
A trial in Helmond, Netherlands tested the Intersection Safety (INS) system, combined with a Cooperative Autonomous Emergency Braking (CAEB) system. In the trial, cyclists were detected by a roadside unit (RSU), which sent Cooperative Intelligent Transport System (C-ITS) messages to a vehicle. If a high risk for a collision between the vehicle and cyclist was detected, the driver would be warned and, if needed, the system would automatically brake to avoid a collision. The cyclist had an intelligent bicycle and received warnings on a display in addition to handlebars that vibrated.
For safety reasons, two use cases were defined. In use case 1 (UC-1), volunteer co-drivers / car passengers were used, together with a cyclist dummy system to evaluate the CAEB system (warning to driver and automatic braking of the car). In the second use case (UC-2), volunteer cyclists were used. For testing, an instructed and experienced driver drove the equipped car.
Testing occurred during March 2015 with 23 treatments rides for UC-1 and 16 pairs of baseline and treatment rides for UC-2.
Evaluation
For UC-1 (dummy bicyclist), the evaluation compared the detection rate of the RSU, to manually-detected bicycles and powered two-wheelers. For UC-2 (real bicyclist) aspects measured included the behavior of the cyclist and the cyclist’s perception of safety, usefulness and acceptance of the system.
Both groups of test subjects completed several questionnaires. They filled in a Van Der Laan Acceptance Scale before and after the test. Additional questions were also asked related to mobility (to measure stated expected changes in trips by bike due to the system), comfort (perceived safety, usefulness, acceptance, and timeliness of warnings) and willingness to buy such a system for a bike.
Results
UC-1: The performance of the CAEB system was consistent during testing: the CAEB system consistently warned the driver when the TTC was estimated to be between 4 and 1.8 seconds. It consistently initiated soft automated braking with -3m/s2 for a TTC between 1.8 seconds and 0.8 seconds and consistently initiated braking with -6m/s2 for a TTC of less than 0.8 seconds.
UC-2: During the volunteer cyclist tests (UC-2) the RSU correctly identified the paths of 29 cyclists and failed to do so for 7 cyclists which resulted in a detection ratio of 81 percent.
In all but one of the use cases, the RSU started transmitting earlier than 4 seconds before the collision, i.e. when the driver should be warned and the distance was larger than the minimum detection distance for warning.
Questionnaire findings
Bicyclists:
- To avoid a critical situation with a car, 60 percent of the respondents (strongly) agreed that the warnings of the system were effective.
- In terms of stress reduction, 40 percent of the respondents were neutral about the statement, "The warnings of the system reduce stress during cycling," while 33 percent agreed and 27 percent disagreed.
- The majority (53 percent) of the respondents indicated that they were willing to pay for this equipment for a bicycle.
Car Passengers:
- 96 percent of respondents (strongly) agreed that the probability of an accident was reduced by using this system.
- The majority of respondents (83 percent) would want to have the system.
- Less than half of the respondents (44 percent) trust the system even though they do not know exactly how it works.