Advancements in imaging and digital processing capabilities are enabling auto manufacturers to provide enhanced design features and regulated safety solutions via a single platform — transforming costly, low-return safety mandates into revenue-generating differentiators.
Technology forecasts predictthat four to 10 cameras will be built into each car by 2015. Some of these will focus on comfort and convenience, while others will aid predictive alerts and safety features. All of these automotive imagers fit into two potential use types: scene viewing and scene processing.
Consumers are most familiar with scene-viewing applications, which display the scene onto a screen for the driver or passenger to react to. These include rear vision and blind spot assistance cameras — which may eventually help replace vehicle mirrors — and side-view cameras mounted on the front and rear bumpers, which help the driver see around blind corners. Scene-viewing applications like these can be a great aid for drivers when backing up, parallel parking or attempting to pull out into busy traffic.
Scene-processing cameras, on the other hand, send video to the car's data-processing system, which in turn feeds data to the vehicle's control systems. Scene-processing applications can include occupant detection and positioning for airbag deployment; drowsy driver detection; automatic mirror, seat or steering wheel and pedal adjustment; driver awareness aids; automated windshield wiper control; automated high beam adjustment; active cruise control; lane-departure warning; pedestrian detection and protection; collision warning, avoidance and mitigation; and imminent collision vehicle adaptation.
Some of the safety applications listed above are now or will soon be regulated, increasing costs for automotive manufacturers Mandatory features provide no room for differentiation, and consumers won't pay the increased cost for regulated technologies. Designers may choose from numerous technologies for any one of these applications, but vision-based systems are a strong contender because the same hardware platform can have multiple uses — a huge benefit. What if the same platform that detected occupant position for smart airbags could also find the driver's eyes and adjust the mirrors or reposition the seat, steering wheel and pedals to move the driver into the best and safest driving position? People would pay for that. What if that same platform could also detect when the driver was not paying attention to the road and issue an alert? Parents of young drivers would pay for that.
Thousands of children are in-jured each year in vehicle back-over accidents — it may not be long before a rearward obstacle warning solution is mandated. Several technologies could be used for this, but a camera could aid in other ways as well, such as showing the best route to back out of a long, narrow driveway, the best way to back into a parking spot, or the best path for parallel parking. People would pay for those features.
Pedestrian collision protection regulations may make exterior airbags standard equipment — a feature the consumer wouldn't pay extra for. If, however, that same hardware platform could also be used to track lane markers, manufacturers would have a key component of drowsy driver alert, active cruise control, “and imminent collision warning/adapt-ation systems. Again, those are the types of features people would pay for.
So these types of mandated safety systems — which typically eat into manufacturers' profits — now have the potential to become a real differentiation source with significant revenue generation. Low-cost, high-performance imagers and processors specifically designed for the stringent environment of automotive applications will help make tomorrow's cars safer and more profitable.
ABOUT THE AUTHOR
Paul Gallagher is the director of technical marketing for the Micron Imaging Group. He has more than 20 years of experience in image sensors and image processing, and has worked on applications as diverse as DNA sequencing imagers, the Barbie cam, and autonomous targeting algorithms. He has worked with CCDs, NMOS and PMOS imagers, including CMOS imagers.