An Advanced Driving Assistance System (ADAS), capable of improving road safety and convenience, has been provided in many new cars as standard. In order for subsequent interpretation and action, ADAS must use high-resolution image sensors to take images outside and inside the car, making high-resolution image sensors a key component of ADAS. This article will show you the development and trends of ADAS and the high-resolution image sensors launched by onsemi.
ADAS has become standard on new cars
In the past, ADAS was a basic or optional component for high-end cars, but, with the rise of the awareness of driving safety in recent years, coupled with the promotion of mandatory laws and regulations in Europe, the United States, and other countries, it is required to integrate ADAS into new low-end cars. In a common driving assistance system, the car navigation system, adaptive cruise control system, lane departure warning system, automatic lane change system, collision avoidance warning system, pedestrian detection, and other functions are incorporated. Some of the more advanced systems are moving in the direction of autonomous driving, with the expectation of achieving a level 5 autonomous driving capability in the future, allowing drivers to carry people or goods to the designated destination without even touching the steering wheel (the car is not even equipped with a steering wheel).
Among the factors that promote the operation of ADAS, in addition to the artificial intelligence system behind it that performs data interpretation and action, its front-end data collection appears to be more important. ADAS cannot function properly without accurate car condition data. Car condition data is primarily collected by image sensors, millimeter wave radar, LiDAR, and other technologies, among which image sensors are the most widespread technology with relatively low cost at present.
The image sensor used by ADAS must have high resolution to improve the system's capabilities to interpret images. The high dynamic range (HDR) feature must also be available to provide the optimal image quality in high- or low-light environment to cope with the changing driving environment and road conditions.
Sharper images are possible with HDR technology
The so-called HDR refers to a technology used in computer graphics and cinematography to achieve a larger dynamic exposure range (i.e., contrast) than normal digital imaging technology. High dynamic range imaging is designed to correctly display sharp images in the real world over a wide range of brightness, from the brightest area of direct sunlight to the darkest shadow.
High dynamic range imaging was initially applied to purely computer-generated images, and subsequent methods were developed to generate high dynamic range images from photographs with different exposure ranges. For example, in repeated exposure image synthesis technology, images with different exposure levels are obtained by adjusting the exposure levels up and down, and then relatively clear partial images are obtained from the continuous images respectively. These partial images are synthesized to obtain the most complete and clearest high dynamic range images.
Previously, HDR technology performed image synthesis through DSP or microprocessors at the back end. onsemi now offers high-resolution image sensors that support on-sensor HDR features to ease the workload of the back-end processor while increasing processing speed and efficiency.
Diverse application demands met through a complete product line
onsemi offers a wide portfolio of image sensors with resolutions ranging from 300,000 to 10 megapixels and frame rates ranging from 4 fps to 815 fps, providing camera manufacturers with a wide variety of options to meet the requirements of all possible end applications, from wearables and consumer electronics to harsh industrial and automotive applications. onsemi image sensors offer configuration flexibility coupled with optimal performance characteristics such as high speed, high sensitivity, and high image quality to meet specific application requirements. onsemi also provides existing customers with a simple upgrade path that allows manufacturers to support multiple products with a single camera design, speeding time to market.
For the application of ADAS in vehicles, image sensors must have high resolution and multiple ADAS-specific features, including advanced HDR features with flexible exposure ratio control, configurable ASIL-B support, black level control, frame-to-frame context switching of multi-function systems, and multi-camera synchronization to assist vehicle developers who are committed to achieving optimal performance to accelerate the development of advanced and fully-integrated driving assistance systems.
Digital image sensors designed for ADAS applications
The AR0820AT, released by onsemi, is a 1/2-inch CMOS digital image sensor with a 3848 H x 2168 V active-pixel array and high resolution of 8.3 megapixels. This advanced automobile sensor is capable of capturing images in linear or high-dynamic range mode and rolling shutter readout. The AR0820AT, with 2.1 µm DR−Pix automotive grade backside illuminated (BSI) pixels and on-sensor 140 dB HDR capture capability, is ideal for low brightness and challenging high dynamic range scenes.
The AR0820AT image sensor incorporates diversified advanced features such as in-pixel binning, windowing, and both video and single frame modes to provide a flexible region of interest (ROI) and specific resolution, enhancing performance in extremely low brightness conditions. The AR0820AT's complex sensor fault detection features and embedded data enable camera ASIL-B compliance. This device is programmable through a simple two-wire serial interface that supports the MIPI output interface.
The advanced on-sensor HDR reconstruction feature with flexible exposure ratio control of the AR0820AT image sensor supports fast full-resolution video capture at 3840×2160, up to 40 fps in 3-exposure HDR and 30 fps in 4-exposure HDR. It supports line interlaced T1/T2/T3/T4 outputs and ASIL-B compliant sensor fault detection, allowing for 2×2 in-pixel binning mode and color binning mode. The 1.8 Gbps/lane, 4-lane MIPI CSI-2 data interface and optional automatic or user-controlled black level control allow the sensor to switch frame to frame among up to 4 contexts to enable a multifunctional system. The sensor, which is Pb-free, supports multi-camera synchronization and offers multiple color filtering matrix (CFA) options, including RGB, RCCC, and RCCB.
The AR0820AT image sensor is found mainly in the field of front view cameras (ADAS) and autonomous driving (AD), with end products including front view ADAS, surround sensing camera, in-cabin camera, robot taxi, robot delivery, autonomous trucks, and other products, making it fit the various requirements of these advanced applications.
onsemi also launched the AR0820AT-iBGA95 evaluation board, which is designed to demonstrate the features of onsemi's image sensor products. The headboard is intended to plug directly into the Demo3 system. Test points and jumpers on the board provide access to the clocks, I/Os, and other miscellaneous signals.
AR0820AT-iBGA95 evaluation board supports clock input, built-in 27 MHz crystal oscillator, and optional Demo3 control MClk. It has two-wire serial interface and selectable base address, provides 4-lane MIPI interface, and complies with ROHS specification.
Conclusion
ADAS has evolved into standard equipment in many new cars and is currently showing rapid market growth. Image recognition technology is conductive to increasing the application and value of ADAS. The current application of driving assistance technology can avoid accidents and reduce collisions. With the continuous improvement of technologies, it is expected that more ADAS applications will be developed in the future and transition to the field of autonomous driving, making it very promising and profitable. onsemi’s image sensors, designed for ADAS applications, can reduce the burden of ADAS developers and are worth your further understanding and adoption.