|
Sensor chip both sees and measures
|
01/04/2007
|
| |
Object recognition and distance measurement almost always involves the use of separate technologies – for example, a camera and radar – that have to be engineered to work together in a ultra-reliable, writes Roger Bishop.
Systems are therefore inherently expensive and take time to bring to market.
The ideal is a low-cost system that provides both ranging and recognition functions in a single package, flexible enough to cover applications including park-assist, airbag out-of-position recognition, lane departure warning and pre-crash warning.
Canesta, a research company in San Jose, California has invented a suitable low cost ‘electronic perception’ technology. It includes new chip-based image sensors – similar in size, complexity and cost to video camera chips – that are uniquely able to resolve the three-dimensional features of a scene, combining ranging and recognition in a single low-cost sensor.
CanestaVision produces the distance from the sensor to the scene at every single pixel in the image, in real time. The technology is implemented in a single CMOS chip.
Canesta believes technology in such a tiny, low-cost format means new safety features will be deployed at a more rapid pace and become available in an even wider selection of vehicles.
The technology – and the company behind it – has some impressive backers. Honda has been quietly supporting the development of the 3D chips to the tune of more than $5 million over the past three or four years. Toshinori Arita, head of the car-maker’s Strategic Venturing, confirmed that the new sensors could “help us provide advanced safety systems in our vehicles”.
Development of the chip started more than seven years ago and is now covered by 17 patents.
What happens inside the CMOS
There are three components of electronic perception technology: the CMOS sensor itself, imaging software and an IR light source. Within the CMOS technology, every pixel is designed as an individual LIDAR (light detection and ranging) element. As every pixel determines range, the resulting ‘image’ is a 3D picture or depth map. As light bounces off the target into the lens, the reflected signal is subjected to a phase delay. If this can be determined, so can the distance travelled by the light. The phase delay is determined by the semiconductor material though what Canesta calls its “differential demodulation structure”. Each pixel has two gates. Synchronised with the light source, gate 1 is polarised high when light is ‘on’ and gate 2 is polarised when light is ‘off’. So if the differential voltage between gate 1 and gate 2 is measured, the phase delay is calculated.
|
| |
Author
Roger Bishop
|
| |
| Email this article |
| |
|
|
|
|