Gated Sensor in ADAS: Improved Vehicle Safety

Introduction

In recent years, considerable effort has been devoted to developing Advanced Driver-Assistance Systems (ADAS) for level 2-4 autonomous vehicles (refer Exhibit 1), especially in terms of improving perception systems in low visibility, adverse weather conditions, and during night driving. Success criteria for any perception system are defined by its performance in difficult driving conditions. Although night-time driving accounts for merely 25% of all driving, the number of pedestrian fatalities at night is actually higher than during the day. According to the National Highway Traffic Safety Administration (NHTSA) report in 2010, 68% of fatalities in the US occurred during the night.

Thus, the selection of sensors suit plays a crucial role in perception systems to accurately detect environmental conditions. Several tier-I suppliers and start-ups worldwide are working on introducing new detection technologies based on depth perception, Time-of-Flight (ToF), stereo-vision, etc., with low production costs and high-performance capabilities. Gated sensor technology, based on the ToF principle, may fill the gaps in the current camera offerings, thereby enhancing performance in adverse weather conditions, especially during low visibility.

Gated Imaging

Gated imaging is a class of Time-Of-Flight imaging technologies, wherein a camera with controlled opening and closing of a shutter is used in conjunction with a high-power pulsed light source for imaging. The camera sensor “opens the gate” (i.e., starts the exposure) after a certain time delay for a very short period of time. Therefore, the sensor is not affected by scattered photons or parasitic light sources (refer Exhibit 2). Only the photons that arrive within the right period of time contribute to the resulting image. The time delay or gating delay determines the position of the range gate in the scene, and the camera gating time or exposure time will define the depth of view. Therefore, the resulting image consists of information only from reflected photons at the distance of interest. The images obtained with this technology contains target reflectivity as well as distance information.

Time-Of-Flight:

The Time-of-Flight (ToF) principle is a method for measuring the distance between a sensor and an object, based on the time difference between the emission of a signal and its return to the sensor, after being reflected by an object.

A snowstorm illustrates this technique (refer Exhibit 3): instead of collecting all the rays reflected by each snowflake, the camera is set up to record only the rays coming back at the distance of the targeted object. The quality of the final image is then highly improved.

Gated Sensors in ADAS Application

The primary goal of gated imaging technology is to improve the contrast of underwater photography by avoiding backscatter. Using this technique, it is possible to detect sea mines and to see through clouds when observing territories from a plane. Military research has further driven the development of this technology for additional applications, such as target identification, night vision, and three-dimensional imaging. In the automotive sector, Brightway Vision paved the way for the use of gated sensors in ADAS applications by introducing cost-sensitive components, such as CMOS imager and laser diodes.

KOITO and Magenta venture partners invested $25 million in BrightWay Vision.

• BrightEye, the flagship offering of BrightWay Vision, is a patented active gated day and night vision safety ADAS, which utilizes a repetitive pulsed NIR extended laser source, located in vehicle headlights, coupled with a gated high-resolution image sensor to provide an active-gated image and ADAS camera-based functions.
• It allows self-driving cars to drive under adverse weather by suppressing backscatter very efficiently.
• Gated images not only can provide intensity images but also can generate perfectly aligned depth information.
  

  Forward Facing ADAS Systems and Comparisons

  Optical Sensing Technologies:

  An inherent problem in optical perception systems is the impact that inclement weather conditions, such as humidity, haze, fog, mist, smoke, and rain, have on the image produced. Particles or substances in the atmosphere may further deteriorate the image produced. For example, haze results from aerosols in the air, creating an obstruction between the observation system and the target to be observed. Gated sensor technology seems to perform better in these conditions compared with other technologies, as indicated in Table 1.

  Night Vision Technologies:

  Active Gated Imaging System (AGIS) utilizes a repetitive pulsed-NIR beam (with a wavelength above 800 nm) to illuminate the scene, and then capture the reflected pulsed light by a synchronized gated image sensor.

  There are two additional types of night vision technologies in the market: Long Wave Infra-Red (LWIR) and NIR systems (refer Exhibit 4). LWIR systems are passive, detecting the thermal radiation at wavelengths in the interval 8-14 μm, typically with a low-resolution image sensor. NIR systems use a continuous light source in wavelengths of above 0.8 μm to illuminate the scene and continuously collect the reflected light using a High Dynamic Range (HDR) image sensor. However, gated sensors seem to outperform them, especially in terms of image quality and commercial aspects, as indicated in Table 2.

• 
  

  Future of Gated Sensors

  As industry players continue to research on highly robust imaging systems, for all light and weather conditions and with a cost-sensitive approach, gated sensors seem to resolve several limitations of the current systems. The key advantage of low-cost CMOS imaging and laser illumination technologies, coupled with computer vision capabilities, would enhance existing standards and future autonomous capabilities in vehicles during the day, night, and harsh weather conditions. As more and more research projects (e.g., the DENSE project shown below) are involved in the development of reliable gated imaging technology, it is expected that many tier-I suppliers will begin exploiting this technology in the near future.

  References

  1. http://roboticsandautomationnews.com/2017/07/01/adas-features-of-advanced-driver-assistance-systems/13194/
  2. https://vision4allseasons.files.wordpress.com/2019/06/abstract_gated.pdf
  3. https://new-imaging-technologies.com/wp-content/uploads/2019/07/Gated-imaging-application-note.pdf
  4. https://www.brightwayvision.com/technology/
  5. https://www.spiedigitallibrary.org/conference-proceedings-of-spie/9407/94070F/Active-gated-imaging-for-automotive-safety-applications/10.1117/12.2078169.short?SSO=1&tab=ArticleLinkFigureTable
  6. https://dense247.eu/technology/