Add LiDAR to ADAS for Pedestrian Safety

Most new vehicles sold in the United States today include Advanced Driver Assistance Systems (ADAS) equipped with Pedestrian Automatic Emergency Braking (PAEB) as either a standard or optional feature. While state-of-the-art ADAS has improved freeway and high-speed driving, the basic need for collision mitigation with pedestrians and bicyclists has mostly remained unaddressed.

Over 6000 pedestrians are killed every year in traffic-related accidents in the United States, reported the National Highway Traffic Safety Administration (NHTSA). A Governors Highway Safety Association report noted that 75 percent of these fatalities occur at night. Utilizing ADAS based on camera and radar technologies, has proven inadequate to addressing this challenge. Independent tests conducted by NHTSA and the American Automobile Association (AAA) reveal that PAEB systems frequently fail to protect pedestrians in dark conditions.

Testing PAEB in Dark Conditions

Figure 2. Images show vehicle with LiDAR-based PAEB stopping before adult target at 50% overlap (top) and vehicle with camera and radar-based PAEB crashing into adult target (bottom). (Image courtesy of Velodyne LiDAR)

PAEB systems could save thousands of lives annually by improving performance in dark conditions. Between 2009 and 2018, pedestrian fatalities increased by 53 percent. Of this increase, 90 percent were caused by nighttime crashes, according to the Insurance Institute for Highway Safety (IIHS) and Governors Highway Safety Association reports. NHTSA has reported that in 2018, 76 percent of the 6283 pedestrian crash fatalities in the US occurred in dark conditions.

Assessments by AAA and NHTSA have shown that PAEB features frequently fail to avoid crashes at night. The AAA report stated “there is little to no publicly available information regarding the performance of pedestrian detection systems in low-light conditions. Based on vehicle/pedestrian crash statistics, this environment is especially critical to evaluate.” Adding that testing PAEB in dark conditions would fill that void. “While this parameter is very challenging, it is nonetheless a reasonable test scenario considering the lack of lighting in many naturalistic environments.”

AAA found that in a test of four vehicles, none alerted the driver or automatically slowed down for a pedestrian crossing the road in dark conditions. Based on these results, AAA advises that “drivers must not rely on assistance from current pedestrian detection systems during nighttime driving or other environments with reduced visibility.” Concluding its analysis, AAA’s report noted that even though “the owner’s manual of each test vehicle states that the integrated pedestrian detection system may not discern pedestrians at night or in adverse weather such as rain, snow, sleet or fog…it is irrefutable that assistance from a pedestrian detection system would be of benefit during nighttime conditions and could possibly be the time of greatest need.”

However, examining current protocols employed by NHTSA’s New Car Assessment Program (NCAP), European New Car Assessment Programme (Euro NCAP), and IIHS demonstrates that performance under nighttime conditions is only rarely tested as a basis for vehicle safety ratings and awards.

Velodyne LiDAR (San Jose, CA) therefore proposes that assessment organizations include in their PAEB test protocols, scenarios conducted in dark conditions. Taking this approach will ensure that consumers of vehicles equipped with PAEB features understand the limitations of their vehicle’s performance and encourage automakers to improve ADAS features that can save thousands of lives annually. More precisely, these tests should be conducted with less than one lux ambient illuminance, using the test vehicle’s low-beam headlights, and without the aid of streetlights.

Comparing LiDAR and Camera/Radar PAEB Systems

LiDAR-based PAEB solutions have inherent strengths compared to camera and radar combinations.

Cameras can have very high resolution, but they typically require multiple modules and additional processing to calculate objects’ distances from a vehicle. Also, like the human eye, cameras perform relatively poorly in dark conditions. A 2018 Department of Transportation (DOT) report states, “Vision-based systems are better able [than radar-based systems] to detect stationary people but are limited to daylight operation in well-lit environments.”

In comparison to cameras, the DOT report noted, radar functions well at night and can provide distance measurements. But radar does not have high enough resolution to perceive the precise location of an object or distinguish between multiple objects that are close to each other. Radar can also fail to detect stationary or slow-moving objects. As a result of these shortcomings, camera and radar based PAEB features struggle to protect pedestrians in nighttime conditions.

Thermal imaging has occasionally been proposed as a potential supplement to camera and radar in PAEB applications. However, this technology presents its own weaknesses and does not adequately address those of current camera and radar systems. As with optical cameras, a vehicle system’s ability to detect objects with thermal sensor data depends on the sensor’s ability to correctly perceive and transmit the contrast between an object and its surroundings. As a result, both sensing modalities can miss objects that blend in with their backgrounds. In the case of thermal imaging this would result from the blending of similar heat characteristics, rather than similar colors or optical illusions, as with cameras.

In contrast, LiDAR does not suffer from any of these characteristic drawbacks of camera, radar, and thermal sensors. LiDAR acts as its own light source so it performs well both in darkness and daylight. It also provides rapid and accurate measurement data with high enough resolution for precise real-time free-space detection while tracking multiple objects within a scene.

Affordable LiDAR-based PAEB solutions that are currently available would significantly improve performance under all lighting conditions. This would be confirmed if regulatory and testing agencies add dark testing to their assessment protocols.

Testing LiDAR-based PAEB System Against Camera- and Radar-based Technology

To demonstrate that improved nighttime PAEB performance can be achieved by implementing readily available technologies, Velodyne tested its LiDAR-based PAEB system against a highly-rated PAEB system built around camera and radar. The tests were conducted at a driver-controlled speed targeting 30 mph on a straight track, one hour after sunset, with less than one lux ambient lighting.

The two test vehicles each had their low beam headlights on during the trials. The stationary child and adult pedestrian dummy targets utilized in the tests were compatible with current testing protocols prescribed by organizations such as IIHS and Euro NCAP.

Figure 3. Test results comparing LiDAR versus camera + radar sensors at night in various scenarios. (Image courtesy of Velodyne LiDAR)

The scenarios in which vehicles were evaluated included:

  1. Crossing adult at 50 percent overlap (at the center of the test vehicle’s width)

  2. Crossing adult at 25 percent overlap

  3. Crossing child at 50 percent overlap

  4. Crossing child at 25 percent, adult @ 75 percent, 10 ft behind child

  5. Crossing adult at driver-side corner

  6. Fallen adult at 50 percent overlap

Velodyne tested both vehicles in each scenario five times, or until the vehicle collided with the target three times, to minimize damage to the targets and vehicles.

The results of the nighttime tests support the findings of AAA and NHTSA that camera and radar-based

PAEB systems frequently fail in dark conditions. The testing found the failure-rate to be especially evident in scenarios involving a child, more than one pedestrian, an adult at the corner of the vehicle, or an adult fallen down in front of the vehicle.

In contrast, the Velodyne LiDAR-based PAEB system, equipped with Velodyne’s Velarray H800 sensor and Vella™ software, successfully stopped in time to avoid a crash five out of five times for every scenario tested. These hardware and software components are being developed to comply with the automotive functional safety requirements defined within the ISO-26262 standard. Combining Vella and Velarray, Velodyne’s PAEB solution is designed for urban, suburban, and freeway driving up to 80 mph.

Conclusion

Since dark nighttime conditions are shown to be dangerous for pedestrians, Velodyne proposes that vehicle assessment organizations expand PAEB testing to include ambient light conditions of less than one lux. Tests performed by AAA and NHTSA demonstrate that nighttime performance represents a major opportunity for improvement in current PAEB systems. Testing shows that a LiDAR-based solution is effective and ready for implementation.

This article was written by David Hall, Founder and Executive Chairman of the Board, Velodyne LiDAR (San Jose, CA). For more information, contact Mr. Hall at This email address is being protected from spambots. You need JavaScript enabled to view it. or visit here .