Skip to content

Beyond human vision: AVs require new perception baseline

Metrics based on human vision are not optimised for machine perception, writes Megan Lampinen

The automotive industry has historically used human vision as the metric by which it develops traffic control and vehicle perception systems. Today, all photometric and road measurements for design, performance, and standard compliance are based on the CIE reference observer for photometry, which is representative of human visual performance (i.e. a biological reference observer). But computers are now taking over much more of the sensing from humans, and camera, radar and LiDAR systems ‘see’ the world differently. What works for human drivers may not work so well for a machine. As the industry heads towards SAE Level 3 and 4 capabilities, it’s time to develop a new frame of reference representative of automated driving and ADAS.

It’s time to log in (or subscribe).

Not a member? Subscribe now and let us help you understand the future of mobility.

Pro
£495/year
or £49.50/month
1 user
News
yes
Magazine
yes
Articles
yes
Special Reports
yes
Research
no
OEM Tracker
no
OEM Model Plans
no
OEM Production Data
no
OEM Sales Data
no
Pro+
£1,950/year
or £195/month
1 user
News
yes
Magazine
yes
Articles
yes
Special Reports
yes
Research
yes
OEM Tracker
yes
OEM Model Plans
yes
OEM Production Data
yes
OEM Sales Data
yes
Pro+ Team
£3,950/year
or £395/month
Up to 5 users
News
yes
Magazine
yes
Articles
yes
Special Reports
yes
Research
yes
OEM Tracker
yes
OEM Model Plans
yes
OEM Production Data
yes
OEM Sales Data
yes
Pro+ Enterprise
Unlimited
News
yes
Magazine
yes
Articles
yes
Special Reports
yes
Research
yes
OEM Tracker
yes
OEM Model Plans
yes
OEM Production Data
yes
OEM Sales Data
yes

Welcome back , to continue browsing the site, please click here