Beyond Braking: Advanced Driver Assistance Systems (ADAS) for a Safer Driving Experience

In the history of automotive development, aside from the evolution of power output, the most significant investment in research and development has been dedicated to providing drivers and passengers with more comprehensive safety protection. Initially, this focused on passive safety features such as airbags, seatbelts, and crumple zone design. With the rapid advancement from microcomputers to automotive computers, the focus shifted to active safety design, giving rise to the concept of Advanced Driver Assistance Systems (ADAS).

While ADAS features like Anti-lock Braking Systems (ABS) have been around for a longer time, the concept of driver assistance has been applied for even longer. As early as 1995, BMW introduced the Park Distance Control (PDC) system in the E38 model, which used ultrasonic signals to help drivers understand the distance between the vehicle and front and rear obstacles during parking, preventing collisions. This marked the beginning of the development of systems that actively influence driver control based on detected vehicle environment data.

Modern ADAS systems are generally divided into three parts: sensors, processors, and actuators. First, sensors such as cameras, radar, ultrasonic sensors, and even more expensive LiDAR are used to measure the distribution of the vehicle’s surrounding environment using light, images, and sound waves. These sensors are typically installed on the front and rear bumpers, windshield, and side mirrors, with the number of sensors depending on the vehicle’s ADAS configuration. According to the SAE International standard, the lowest level, Level 0, has no automated driving functions and only provides driving assistance information. For example, the commonly used Around View Monitoring (AVM) system only requires the installation of cameras and short-range radar.

ADAS is not synonymous with autonomous driving but is a prerequisite for achieving it. Processors and actuators play a crucial role in self-driving technology at Level 1 and above. In addition to displaying the sensor data for driver reference, the data is also transmitted to the processor, where the Electronic Control Unit (ECU) analyzes it and makes decisions, which are then transmitted to the actuators to control the throttle and steering wheel, enabling partial or fully automated vehicle control. For example, the Level 2 Automatic Parking Assist (APA) system only requires the driver to control the throttle, while the actuator operates the steering wheel.

Market sales data and public information show that Level 1 and Level 2 self-driving technologies will remain the mainstream for the next five years. Cameras and ultrasonic sensors are the primary components. Although cameras rely on software to simulate human visual perception to judge the environment, they are susceptible to weather and light conditions. However, compared to radar and LiDAR, cameras are less expensive and can be installed at multiple angles on the vehicle to obtain a complete view of the environment, making them the current mainstream configuration for ADAS.

Cameras provide image information and are therefore called image sensors (CIS: CMOS Image Sensor). The captured color image is divided into red, green, and blue pixels by a filter and enters the CIS. The CIS, based on its resolution and pixel design, has a corresponding matrix. Each pixel is a photodiode. When a photodiode receives light, it is converted into an electrical signal, which is transmitted through CMOS switches as an analog signal to the next component, the analog front-end (AFE), where it is converted into a digital signal and then processed by an image signal processor (ISP) to generate an image, which is then transmitted to the processor SoC (System on Chip) for calculation, judgment, prediction, and even learning. The conversion from analog to digital signals requires extremely fast speed and low interference, allowing the SoC to accurately interpret the data in real-time and provide the necessary parameters for the backend actuator to operate, assisting the vehicle in steering and power system driving.

CMOS Image Sensor and Signal Processing

With the rapid development of ADAS, there is a need for stable and fast electronic signal transmission. As the amount of information transmitted increases, more power is required. Since ADAS has multiple processors, such as PCIE Switch, Security ECU, etc., to avoid interference between different controllers when problems occur, each processor requires a separate power supply. Additionally, to meet the requirements of down-conversion and reducing electromagnetic interference (EMI) within a limited space, power supply circuits often use Power Management ICs (PMICs), which integrate DC/DC converters and low dropout regulators (LDOs) and have a smaller footprint. For example, as mentioned earlier, if the CMOS image sensor encounters problems with image acquisition due to freezing at low temperatures, a heater is used to heat the CMOS sensor, and the power for the heater needs to be provided by a boost converter. The boost converter uses MOSFETs as switching elements.