Infineon launches first ISO26262-compliant, automotive 3D image sensor

3D depth sensors play a key role in monitoring systems in a vehicle’s cabin, allowing for innovative car cockpits and seamless connectivity with improved passive safety. These sensors are becoming essential for meeting regulations and NCAP ratings as drivers become more like passengers.

To this end, Infineon Technologies, in collaboration with 3D time-of-flight system specialist pmdtechnologies ag, has developed the second generation of the REAL3 automotive image sensor — an ISO26262-compliant, high-resolution, 3D image sensor.

“Leveraging our leadership in 3D sensors for consumer applications, we are now offering high resolution with a tiny image circle to the automotive world,” said Christian Herzum, VP 3D Sensing at Infineon. “This enables cars with functions from the consumer world, while maintaining automotive standards and even improving passive safety. For instance, reliable and secure facial authentication allows seamless connectivity for any type of service that requires authentication such as payment, battery charging, or accessing private data.”

In addition, the same camera meets all requirements for driver monitoring to detect driver distraction and fatigue. This enables the offering of a driver-monitoring system with secure 3D facial recognition using only one ToF camera.

“From the start we were focussed on improving the robustness of the underlying pmd-based ToF Technology against external influences such as sunlight or other disturbing light sources. For this reason, the new imager shows excellent and cutting edge performance even under harsh conditions”, said Bernd Buxbaum, CEO pmdtechnologies ag.

The sensor comes in a 9 x 9 mm² plastic BGA package and offers a VGA system resolution of 640 x 480 pixels with a tiny image circle of 4 mm. This allows lens sizes similar to those known from smartphones, but now also for automotive applications. The high resolution of the REAL3 sensor also makes it suitable for camera applications with a wide field of view, such as complete front-row occupant monitoring systems. The resulting 3D body models enable accurate estimates of occupant size and weight, as well as highly precise passenger and seat position data, which is key information for intelligent airbag deployment and restraint systems.

In addition to its small size, the single-chip solution is qualified to AEC-Q100 Grade 2 and is the first of its kind being developed according to the ISO26262 (ASIL-B) standard.

Besides safety critical applications, the 3D data allows for comfort features such as gesture control or intuitive interior lighting that follows passengers’ movements. Additionally, the imager might be used in environmental perception scenarios as a Flash-LiDAR in the automotive space and — based on the safety conformity — it’s also an ideal candidate for adjacent applications in mobile robotics, drones, and other autonomous use cases in general where reliability and safety is a must to protect operators.

The time-of-flight technology of the depth sensor is the basis for a wide range of use cases. It provides highly reliable distance information and at the same time a 2D grayscale image of the entire scene under all ambient light conditions — at night, in sunlight, and under heavily changing lighting conditions.