Applications of sensors in autonomous driving

Binapart Electronics GroupAbout BinapartBinapart Electronics GroupAbout BinapartBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart

Home
News
Power management
Applications of sensors in autonomous driving

May 30 , 2023 0

Information perception is the foundation and source of autonomous driving algorithm. Comprehensive recognition and judgment of the surrounding environment is a prerequisite to ensure the safety of intelligent vehicles. Currently, cameras, millimeter-wave radar, ultrasonic radar and lidar are all commonly used sensors in autonomous driving. The development and progress of sensor technology is very important for the improvement of vehicle perception ability in complex driving environment.

Cameras are the most indispensable sensors in today's smart cars. It has high resolution and can be used for model calibration, target tracking and obstacle identification through deep neural network learning, which can be used for road environment perception outside the vehicle and monitoring of drivers and passengers inside the vehicle. Visual perception solutions are widely used by vehicle manufacturers to implement Level 2 and above assisted driving features such as forward collision warning (FCW), Lane Departure warning (LDW), adaptive cruise control (ACC), and autonomous emergency braking (AEB).

The number of cameras installed on bikes has increased in recent years as the installation rate of ADAS (Advanced Assisted Driving) has increased. Tesla, for example, has eight cameras for its autopilot hardware system, while Nio's ET7 has 11 cameras. In 2019, there were an average of 2.2 cameras per car worldwide, but Omdia expects that number to double by 2026.

As the number of cameras in use increases, so does the demand for camera performance. The adoption of more resolution cameras is a market trend, such as the Nio et7 Smart Drive with 11 8-megapixel cameras. In high-contrast scenes, the camera sensor chip requires a higher dynamic range to capture high-quality images. In-car cameras with higher frame rates provide smoother video. The global shutter technology of the sensor chip helps to avoid the "jelly effect" of images in high speed motion.

In addition, car cameras have some unique requirements, such as LED flicker suppression technology, to ensure that the car can correctly recognize the LED signal. With the deepening of autonomous driving technology, the on-board camera also needs to sense the depth information of the environment. For example, the driver fatigue monitoring function, and through face recognition to determine the driver identity and other functions. Therefore, 3D sensing technology such as ToF will attract more and more attention from the market.

Millimeter wave radar

Millimeter wave radar is a kind of equipment that uses millimeter wave to detect the distance of obstacles. According to the detection distance, millimeter wave radar can be divided into short range radar (SRR), medium range radar (MRR) and long range radar (LRR). With the continuous improvement of autonomous driving Level, millimeter wave radar will be widely used in Level 2 and above autonomous vehicles due to its cost advantage and stable working performance. It works 24/7 and is a necessary complement to the camera.

Millimeter wave radar is developing towards smaller size, higher accuracy and longer detection distance. Due to increasing safety requirements from regulations for autonomous driving, 24GHz millimeter-wave radars will gradually be replaced by 77GHz products, as higher frequencies mean higher performance, wider bandwidth and better resolution.

Omdia expects to have an average of five to eight millimeter-wave radars on autopilot systems at Level 3 and above to enable blind spot detection (BSD), lane change assistance (LCA) and rear collision warning (RCA).

Ultrasonic radar

Ultrasonic wave is one of the most commonly used sensors in conventional cars. It is a device that transmits, receives and processes ultrasonic signals and obtains information about the distance and position of the target. Ultrasonic sensors perform well in harsh weather conditions, such as fog, rain, snow and low-light conditions, and are relatively inexpensive. As a result, it is widely used in astern radar and automatic parking systems.

The disadvantages of ultrasonic radar include long response time, limited field of view, and low accuracy. In addition, ultrasonic sensors have difficulty detecting small, fast-moving objects or multiple objects. Omdia believes that in the long term, some of the ultrasonic radars used in more advanced autonomous vehicles will be replaced by better millimeter-wave radars.

Lidar

Lidar is a system that transmits laser beams and receives echoes to obtain information about the three dimensions and velocities of a target. Lidar is considered the most important sensor for Level 3 and above because it has a longer detection range than cameras and can identify static targets more accurately than millimeter-wave radar. However, lidar can be ineffective in heavy rain or fog. In addition, due to its high cost, the vehicles equipped with lidar at this stage will still be mainly high-end models.

In terms of technology trends, semi-solid lidar solutions are now a mainstream choice due to their maturity and ease of passing vehicle regulations. The future of Lidar will be towards miniaturization, high resolution (64 channels, 128 channels, even 200 channels, etc.) and low cost solid-state solutions. In addition, the solution of integrating lidar into a single chip will be a long-term research and development direction.

Tesla has long promoted a "pure visual" perception scheme, based on its algorithmic strengths and data reserves, without the use of lidar. But its latest self-driving hardware HW4.0 incorporates 4D millimeter wave radar. 4D millimeter wave radar can realize information perception in four dimensions (3D+ altitude) and is cheaper than lidar. It is expected to replace short-range lidar in the future and facilitate further penetration of advanced autonomous driving.

As each type of sensor has its limitations and cannot provide complete information about the vehicle's surroundings, there is an increasing demand for sensor fusion technology. Sensor fusion technology refers to the integration of information obtained by various sensors (such as millimeter wave radar, camera, laser radar or ultrasonic radar), comprehensive analysis and decision-making, so as to obtain a more complete and reliable description of the environment, so as to improve the confidence of autonomous driving decision-making.

Omdia believes that the need for sensors of all types will be amplified in the coming decades as smart cars replace conventional fuel vehicles. The advancement of autonomous driving will also promote continuous innovation in sensor technology and push sensor performance to a higher level.