Active Safety Systems with Accurate Ranging Can Reduce Traffic Fatalities

The past 2 years have been extremely challenging in many ways. More than 5 million people around the world have died from COVID-19. Also, there have been countless secondary effects of the pandemic—among them, a surge in traffic fatalities. Despite a significant reduction in vehicle miles traveled during lockdowns in the first half of 2020, US traffic fatalities still rose by 7.2% compared with 2019, and they are up another 18.4% through the first half of 2021. In addition to changing driver behaviors, more robust and reliable active safety systems are needed to address this chronic issue.

The surge in fatalities is likely a direct result of people driving faster and more recklessly when congestion thinned out during the early part of the pandemic. However, even before the pandemic there was a troubling upward trend in fatality numbers, with the biggest increases among vulnerable road users (VRUs), pedestrians, and cyclists. Improvements in vehicle occupant protection over the past several decades have dramatically improved the likelihood of survival and reduced injury when a crash happens. But that doesn’t help those outside the vehicle.

Proactive Safety

Now we need much more emphasis on proactive safety that prevents crashes in the first place. The share of new vehicles equipped with advanced driver assistance systems has increased substantially in the past 5 years. Although these systems can be helpful—especially the radar based blind spot monitors and cross traffic alert systems at the rear of the vehicle—forward collision prevention still lags. Many vehicles are still relying on monovision cameras and, in some cases, a low resolution forward radar sensor.

Radar is great for measuring distances and closing speed to other vehicles even in poor visibility but current sensors are insufficient, especially for VRUs. Cameras are inexpensive and extremely useful for enabling object classification through machine vision. Unfortunately, the single cameras used for most applications today must rely on inference for distance and speed measurements, and that is far too prone to error.

Active sensors such as high resolution radar and lidar can accurately complete those measurements but they are much more costly than cameras. However, arrays of multiple forward-facing cameras can be used for accurate and high resolution distance measurement based on geometry rather than error-prone inference. The removal of infrared cut filters from these camera sensors even allows improved low light and visibility performance.

Making Safety Accessible to All

Used in this way, multi-camera sensing can be implemented at a relatively low cost, making it accessible to vehicles at all price points. Even as more vehicles begin to incorporate more advanced partial and high automation driving capabilities over the coming decade, such camera arrays still need to be part of the sensing suite to enable the fail-operational capabilities needed when a human driver is not available as a backup.

The Guidehouse Insights white paper, Accurate Ranging Perception for Assisted and Automated Driving, delves into how the different types of sensors work and the benefits and challenges of each. Join me, Light CEO Dave Grannan, and Light chief product officer Prashant Velagaleti for a discussion of the importance of accurate ranging perception during a Guidehouse Insights webinar on November 16, 2021 at 2 p.m. EST.