https://www.greencarcongress.com/2021/1 ... 6-aaa.html
AAA study finds automatic emergency braking and lane keeping assistance performance can be impeded by rain
A new study from AAA finds that moderate to heavy rain affects a vehicle safety system’s ability to “see”, which may result in performance issues. During closed course testing, AAA simulated rainfall and found that test vehicles equipped with automatic emergency braking (AEB) traveling at 35 mph collided with a stopped vehicle one third (33%) of the time. Lane keeping assistance test vehicles departed their lane 69% of the time.
Depending on the type of sensor, external influences such as weather and sensor cleanliness can have various effects on operation. Specifically, radar sensors are minimally affected by rain, snow, and fog relative to other sensor types and function equally well in lighting conditions ranging from complete darkness to blinding sun. Additionally, radar sensors tend to be less affected by bugs and dirt because they are frequently placed behind plastic bumper covers. In cases where radar sensors are exposed, emitted radio energy can penetrate these particles with minimal attenuation.
However, systems such as lane keeping assistance or lane departure warning systems require integration of additional sensors because radar is not effective at discerning object detail and cannot detect variations on a flat surface, such as lane markings. Image sensors (cameras) are currently utilized for object classification and to track lane markings. Depending on the type of electromagnetic radiation (visible, near-infrared, medium wave infrared, etc.) detected by the camera, the sensitivity to weather and dirt varies. Cameras that detect energy in the visible spectrum are most affected by weather, lighting conditions, and bugs/dirt relative to cameras specific to the infrared spectrum.
Automatic emergency braking systems utilize front-facing radar and/or camera(s) to obtain kinematic data pertaining to surrounding vehicles and objects. Lane keeping assistance systems currently rely on one or more cameras to track the position of lane markers. Lane keeping assistance and automatic emergency braking systems effect sustained lateral and temporary longitudinal motion control, respectively.
—“Effect Of Environmental Factors On ADAS Sensor Performance”
Advanced driver assistance systems (ADAS) are typically evaluated in ideal operating conditions. However, AAA believes testing standards must incorporate real-world conditions that drivers normally encounter. . . .
AAA selected crossover vehicles for testing because of their continuing popularity in the United States. In 2020, sales of crossovers and utility vehicles accounted for 50% of the new vehicle market share. Additionally, the following criteria were utilized for vehicle selection:
Inclusion of domestic and import OEMs including European and Asian manufacturers
Variety of manufacturers (only one vehicle per manufacturer will be tested)
Based on those requirements, AAA selected the following vehicles for testing:
2020 Buick Enclave Avenir with Automatic Emergency Braking and Lane Keep Assist
2020 Hyundai Santa Fe with Forward Collision Avoidance Assist and Lane Keeping Assist
2020 Toyota RAV4 with Pre-Collision System and Lane Tracing Assist
2020 Volkswagen Tiguan with Front Assist and Lane Assist
Rain has the biggest effect on vehicle safety systems. AAA, in collaboration with the Automobile Club of Southern California’s Automotive Research Center (ARC), simulated rain and other environmental conditions (bugs and dirt) to measure impact on the performance of ADAS-like automatic emergency braking and lane keeping assistance. Generally, both systems struggled with simulated moderate to heavy rain, with results showing:
Automatic emergency braking engaged while approaching a stopped vehicle in the lane ahead
In aggregate, testing conducted at 25 mph resulted in a collision for 17% of test runs
In aggregate, testing conducted at 35 mph resulted in a collision for 33% of test runs
Lane keeping assistance engaged to maintain the vehicle’s lane position
In aggregate, veered outside of the lane markers 69% of the time
During testing with a simulated dirty windshield (stamped with a concentration of bugs, dirt and water), minor differences were noted, however, performance was not negatively impacted. While AAA’s testing found that overall system performance was not affected, ADAS cameras can still be influenced by a dirty windshield. It is important drivers keep their windshields clean for their own visibility and to ensure their ADAS system camera is not obstructed.
Also, some systems may provide an alert or deactivate in extreme situations, however, the conditions AAA tested under provided no such alert or warning. . . .
Previous AAA testing of vehicle safety systems in both closed-course and real-world settings show that performance is greatly impacted by driving scenarios, road conditions and vehicle design, finding issues such as the following:
Struggling to stay within in a marked lane in moderate traffic, on curved roadways and on streets with busy intersections.
Failing to stop for pedestrians in common scenarios such as crossing in front of a vehicle, a child darting out between two parked vehicles, or walking at night.
Colliding with a simulated disabled vehicle and instances of coming too close to other vehicles or guardrails.
AAA’s research continues to show that vehicle safety system performance varies widely, reinforcing that they are not a replacement for a fully engaged driver.
AAA recognizes these systems have the ability to lessen the chance of a crash and improve the overall safety of driving. Fine-tuning their performance and providing drivers with a more consistent experience will go a long way in unlocking their true potential.
https://www.greencarcongress.com/2021/1 ... onyai.html
Pony.ai to start driverless tests on public roads in Beijing
. . . In Beijing, Pony.ai is now authorized to conduct driverless testing in an area of around 20 square kilometers in a pilot zone for autonomous driving vehicles. The area covers major subway stations, residential areas, and tech parks, which allows Pony.ai to test its autonomous technology in the most challenging road conditions.
The company was authorized in June to conduct driverless tests in California and Guangzhou, China.