Understanding the Autonomous Driving Sensors

0
73
Autonomous Driving Sensors

With automatic electricity, intelligence, sharing development, autonomous driving prototypes have gradually emerged. Although autonomous driving cars have not been fully commercialized, many car companies have begun to layout L3 or even L4-level autonomous driving vehicles.

Six levels of autonomous driving

The US National Highway Traffic Safety Authority (NHTSA) divides autonomous driving into six levels. That is according to the control of the vehicle occupied by the driving behavior.

The L0 is not automatic, the driver needs to master all the machines. The vehicle has only an alarm device, providing a safety warning for the driver during the vehicle driving. And the vehicle does not participate in any driving behavior process.

The L1 level is operated by the driver. The vehicle can have a function of a vehicle anti-lock system, such as a vehicle anti-lock system. These functions are only auxiliary driving safety, and the control of the vehicle is still in the hands of the driver.

The L2 level is still operated by the driver. The vehicle has a fixed-speed cruise system, blind spot detection system can reduce the driver’s operational burden.

The L3 is in some driving scenarios. The vehicle is able to control itself. There is no need for the driver’s participation. When the vehicle detects the road condition that requires the driver to control the vehicle, the driver will immediately carry out subsequent supervision.

L4 is a vehicle that can complete autonomous driving in a conditional case. Generally, it does not require the driver’s intervention. The vehicle can send passengers to the destination in accordance with the set procedure, but in a harsh climate or accident, the vehicle will remind the driver and take over the driver enough to take over. And the driver still needs to supervise the operation of the vehicle at the time of the L4 level.

When the L5-level autonomous driving, the driver does not need to be in the cab. The vehicle does not require driver monitoring. It can independently complete the autonomous driving behavior in all road environments. The vehicle can perform optimal road planning and decision on the target point.

In-vehicle cameras

The car camera is the basis for achieving many warnings, identifying advanced assist driving functions. For drivers, the lane deviates from warning, forward collision warning, traffic logo identification, lane maintain assist, pedestrian collision warning, panorama Advanced auxiliary driving functions such as cars, driver fatigue warning, requires the assistance of a car camera.

The main hardware of the car camera includes an optical lens, an image sensor, an image signal processor ISP, a serializer, a connector, or the like, an optical lens, a filter, and a protective film, or the like. The optical lens is responsible for focusing the light, projection of objects in the field of view into the surface of the imaging dielectric. It depends on the requirements of the imaging effect, may require multiple layers of optical lenses, the filter can filter out the light wave segment that the person is not seen, leaving only Visible optical bands of actual scenes within the human field of view.

The image sensor can use the photoelectric conversion function of the photoelectric device to convert the light image on the photosensitive surface to the electrical signal of the corresponding proportional relationship, mainly divided into CCD and CMOS. The ISP image processor mainly uses the hardware structure to complete the pre-processing of image video source RAW format data input by the image sensor, converted to YCBCR, etc… It can also complete image scaling, automatic exposure, automatic white balance, auto-focusing, etc. The serializer transmits the processed image data, which can transmit a variety of image data species such as RGB, YUV. The connector is primarily used to connect the fixed camera.

Tesla Front Camera Module

Tesla Front Camera Module

Car cameras are also higher than industrial cameras and commercial cameras in terms of the manufacturing process and reliability requirements. The working environment of the car has changed. Sometimes work in harsh environments. The car camera can be divided into a front-view camera, a comic camera, a rear view camera, a side camera, and an inner view camera depending on the installation position.

Millimeter-wave radar

Millimeter-wave radar refers to electromagnetic waves between 1 mm to 10 mm. The wavelength of millimeter waves is between centimeters and optical waves. Therefore, millimeter waves have the advantages of microblogging guidance and photoelectric guidance. Millimeter-wave radar is a small size, easy integration, and spatial resolution relative to cm. Millimeter-wave radar can work all day, in extreme weather and night can play a role. The millimeter-wave radar ranging is also far, but the resolution is low, it is difficult to image and cannot identify images.

The millimeter-wave radar is relatively mature. It uses relatively popular in advanced auxiliary driving. Millimeter-wave radar is an important part of the perceived link. However, the resolution is low and cannot be used as an alternative to laser radar. At present, the resolution of automobile millimeter-wave radar is divided into 24 GHz, 77 GHz, and 79 GHz. 24 GHz is also known as short-band, the sensing distance from 0.15 meters to about 30 meters, mainly used as a parking assist. 77 GHz sensing distance from 1 meter to about 100 meters, mainly used to detect blind spots. The 79 GHz sensing range can reach 250 meters. It mainly uses the active cruise system and the automobile to collide alarm system, so that the owner has enough time to brake or dodge.

4D high-wellness imaging mm-wave Radar increases the detection and analysis of radar on target pitch height data can realize the pitch angle, time, distance, and azimuth information infection, 4D high-precision imaging millimeter-wave increase time dimension information, it can be effective Data such as parsing the behavior, size profiles, and categories of the target. 4D high-profile imaging millimeter-wave radar by increasing the virtual channel, greatly enhanced resolution and target detection confidence and detection range (such as distance and fov), and evolving the same high-density point cloud-like laser radar, can bring rich perception Enhanced application.

Lidar

Lidar, also known as optical radar, is an abbreviation of laser detection and ranging system. It analyzes the propagation distance between the sensor transmitter and the target object, analyzes the amplitude, frequency of reflection spectrum, frequency, and Phase information, etc., thereby presents three-dimensional structural information of the target.

lidar map

lidar map

Lidar is a highly progressive radar technology for military transit. It has excellent performance that achieves extremely high angles, distance, and speed resolution. Usually, the angular resolution of lidar is not less than 0.1mard. And it can track multiple targets simultaneously. The distance resolution can reach 0.1m. The speed resolution can reach 10m / s. Such high distances, speed resolution means that lidar can be greatly clearly available with Doppler imaging technology.

Lidar is linearly propagated. The direction is good, the beam is very narrow, and the diffusion is very low. Therefore, the lidar has high precision. Unlike millimeter-wave radar, there is not much signal source that can interfere with lidar in nature. Therefore, the anti-interference ability of lidar is particularly strong. Most lidar systems are composed of 4 parts of lasers, optical scanners, photoelectric detectors, and navigation systems.

Depending on the action, lidar can be divided into one-dimensional lidar, two-dimensional lidar, three-dimensional lidar, etc… One-dimensional lidar is mainly used for ranging speed. Two-dimensional lidar is mainly used for contour measurement, object identification, regional monitoring. Three-dimensional lidar can achieve space three-dimensional modeling. Lidar installed in the autonomous driving car is three-dimensional lidar. Through high-speed rotation, the point cloud data around the vehicle can be obtained, thereby plotting a three-dimensional surrounding vehicle’s Space map. Lidar can also measure information such as surrounding vehicles, pedestrians, speed, acceleration, angular velocity, etc., and then combined with GPS maps to calculate the location information of surrounding vehicles, pedestrians, to automatically drive a driving judgment.

Ultrasonic radar

Ultrasonic radar’s receiving probe receives ultrasonic waves reflected back through an obstacle, calculating the distance between the obstacle according to the time difference received by the ultrasonic reflection. The ultrasonic sensor is cheap. Its detection distance is near and the accuracy is high and is not affected by light conditions. It is not affected by waterproof and dustproof. Even if there is a small amount of sediment occlusion, it is used in the parking system.

However, ultrasonic radar has certain limitations in the case where the speed is high, as the ultrasonic radar is largely affected by the weather. In different weather conditions, the transmission speed of ultrasonic waves is different. When the transmission speed is slow and the speed is high, Ultrasonic ranging cannot keep up with changes in automotive speed. That will cause greater errors. Since the ultrasonic scattering angle is large, the directionality is poor. The echo signal will be greatly weakened, affecting the accuracy of the measurement when the measurement is long.

Ultrasonic generators have three types of operating frequencies. In general, the higher the frequency, the higher the sensitivity. But the smaller the horizontal and vertical direction will be, 40 kHz Working frequency, common ultrasonic radar has two types. The first is on the bumper before and after the car. That is, the reversing radar used to measure the obstacle before and after the car. This radar industry is called UPA. The second is installed on the side of the car. It’s used to measure the side disorder Ultrasonic radar of the material, and is called APA. At present, most cars use reversing radar UPA. With the continuous improvement of autonomous driving technology, automatic parking function based on ultrasonic radar gradually enters the public vision. And the APA market will gradually open.

Read more about technology:

Why has 5G Become One Of The Much-Awaited Technologies In 2021?

LEAVE A REPLY

Please enter your comment!
Please enter your name here