Cat's eye retro-reflectors extend the range of lane detection when using automated vehicle camera technology at night, especially in inclement weather.

A presentation on automated vehicle camera technology and performance considerations for Advanced Driver Assist Systems (ADAS) applications.

Date Posted
10/16/2019
TwitterLinkedInFacebook
Identifier
2019-L00911

ADAS / AV Camera Technology Introduction

Summary Information

Automated vehicle (AV) camera technology can detect and extrapolate location of lane markings and identify a wide variety of MUTCD traffic signage at vehicle speeds of up to 155 mi/h, with lateral offsets up to approximately 60 feet and at heights up to approximately 20 feet dependent on light levels and weather conditions (e.g. fog, snow, rain). Current AV camera systems can compare and process a visual field (1 to 2 mega pixels) up to 35 times per second. Delta information can be used to estimate movement and paths of other objects e.g. pedestrian approaching roadway. Vision software requires an object to have a minimum number of pixels width for recognition (typically around 20). Newer systems continue to improve frame rates for faster detection of objects (e.g. signs) and lane markings, and faster calculation of desired vehicle paths.

The following ADAS applications are enabled by AV camera technology:
  • Lane Departure Warning
  • Lane Keeping Assist
  • Distance Warning
  • Pre-crash Vehicle Preparation
  • Collision Warning
  • Automatic Emergency Braking
  • Basic Adaptive Cruise Control
  • Traffic Sign Recognition
  • Auto Head Lamp (on/off)
  • High Beam Control.

Lessons Learned

Recommendations for infrastructure upgrades to improve function ADAS and AV Camera technology applications.

  • Very wide or narrow lane marks are uncommon and should be discouraged; 12 to 15 cm is the ideal width as they are most common; lane marks under 10 cm or over 25 cm wide should not be permitted
  • Very long dashes can be mistakenly detected as solid; dashes should be limited to a maximum length of 5 meters
  • Markings with higher contrast are easier to detect; in some scenarios the combination of lighting and concrete causes yellow marks to be less visible than white marks
    • White road markings are better than yellow
  • Under certain conditions (angle of the sun or wet roads at night) tar seams can appear as lane marks
    • Recommend using less reflective material
  • Road construction: overlap of old and new lane markings can be confusing; hiding irrelevant marks can also be useful for humans
  • Cat’s eye retro-reflectors can extend the range of lane detection at night, particularly in inclement weather conditions
  • Botts’ dots can make detection difficult, especially if they are dirty or have worn-away paint. The study suggested using dots with retro-reflectors to improve visibility.
Goal Areas
System Engineering Elements

Keywords Taxonomy: