A Beginner's Guide to Self-Driving Car Technology

Many people think that autonomous driving technology is a technology that can realize vehicle driving without manual control. Is this true? When cars no longer need drivers, how much do you know about it? What is car autonomous driving, and how can we understand its functions and principles in a simple and easy-to-understand way? To understand the functions and principles of autonomous driving technology, we must first talk about what autonomous driving technology is. In fact, as long as it relies on computers and sensors to replace people to perform some of the driver's operations, it can be called autonomous driving technology.

But when it comes to autonomous driving technology, most people think it is a technology that can achieve vehicle driving without manual control. Of course, this is true in some cases. This is the most mature technology of autonomous driving technology, that is, the most perfect autonomous driving technology. But in fact, the current level of technology and concept limitations are far from achieving the above goals, and complete autonomous driving will not appear in the lives of our generation.

Since it is impossible to achieve full autonomous driving, there are naturally different levels of autonomous driving. Currently, autonomous driving is divided into five stages according to the degree of machine intervention. The overall technology is currently in the transition stage from L2 to L3. After the L3 stage, the machine will take over the driver's perception and control of the vehicle.

For example, the already mature Adaptive Cruise Control (ACC) technology has been widely used in cars priced between 100,000 yuan and 200,000 yuan. So what is an adaptive cruise system? It is a system that automatically maintains a stable vehicle speed. In complex traffic conditions, the ACC system can automatically control the vehicle's forward speed. However, this can only belong to the second stage of autonomous driving, that is, the driving assistance stage. In more popular terms, it is to replace various parts of a person with different types of sensors, motors, and processors.

Back to the topic, talking about the principles of autonomous driving technology, most universities and intelligent driving companies currently have autonomous driving divided into three modules: environmental perception, intelligent decision-making, and control execution. Environmental perception is similar to human eyes and ears, receiving information from the outside world; intelligent decision-making is equivalent to the human brain, processing information based on the senses; control execution is similar to human feet and hands, deciding how to turn, accelerate or decelerate. The figure below shows the structure of autonomous driving technology.

The environmental perception layer generally includes various vehicle status sensors and industrial computers to achieve vehicle and environmental information fusion. Commonly used vehicle status sensors include: radar (mainly laser radar and millimeter wave radar), image acquisition system, camera, inertial navigation, GPS, etc., which are used to locate the vehicle position and perceive the objects in front and their distance. Various types of information need to be fused through industrial computers to finally capture environmental information and pass the information to the intelligent decision-making layer.

Specific implementation method: Through the on-board environmental perception technology and satellite positioning technology (RTK), the static and dynamic information of the vehicle's own attributes and the vehicle's external attributes (such as road vehicles and pedestrians, etc.) are extracted and collected, and the industrial computer completes the data screening and fusion of multiple sensors, and identifies roads, vehicles, pedestrians, traffic signs and traffic signals, etc., and transmits necessary information such as road conditions, traffic signs, vehicle status, pedestrians and vehicles to the intelligent decision-making layer through CAN signals.

Among all kinds of sensors, radar is equivalent to the human eye, which undertakes most of the functions of positioning and receiving information. Other sensors are similar to the ears, touch, smell and other perception systems, which assist radar in completing the detection work. The full name of millimeter wave radar RADAR is Radio Detection And Ranging, which uses radio waves to detect and locate objects. As shown in the figure below, the combination of short-range radar and long-range radar can achieve the best detection effect.

What is LIDAR? As mentioned before, it is the abbreviation of laser detection and ranging technology. In addition to the laser transmitter, this system also requires a high-precision receiver. Because it can provide a three-dimensional image of the detected object in a unique way, LIDAR is mainly used to measure the distance to fixed or moving objects. Here is a common sensor type and its parameters included in the perception layer, which is convenient for understanding the types of different sensors and their general functions.

The intelligent decision-making layer includes hardware and software control programs. Based on comprehensive considerations of safety, economy, and comfort, the intelligent decision-making layer will eventually produce a real-time processing signal, which is transmitted to the control execution layer. The intelligent decision-making layer includes the vehicle controller at the hardware level and the vehicle control program at the software level. The specific implementation method is: receiving the information fused by the environmental perception layer, judging the vehicle driving mode and the operations to be performed based on the vehicle's own status and the vehicle's surrounding environment, and at the same time transmitting instructions to the control execution layer and calling related subsystems such as adaptive cruise control (ACC), automatic emergency braking (AEB) and lane keeping (LKA).

The control execution layer mainly controls the electric drive system and the electronically controlled chassis. The control execution layer completes the final driving and control actions through the coordination of the electric drive system, the dynamic coordination of the braking system, and the dynamic coordination of energy resource management, and regulates the air conditioning, thermal management, motor, and hydraulic brake. The specific implementation method is: control the longitudinal braking torque, accelerator pedal opening, gear position, and lateral steering wheel angle. The control of the braking part, steering angle, accelerator pedal opening, and gear position requires modification of each part and calibration. After the modification and calibration are completed, sending the corresponding CAN information to the corresponding interface of the car can complete the basic actions of steering, braking, acceleration, and shifting. And it can operate the vehicle and control the human-machine collaboration according to the instructions of the intelligent decision-making layer, and provide road traffic information, safety information, etc. for the networked car to ensure safe driving and comfortable driving of the car.

In summary, autonomous driving technology is an automated and intelligent technology that replaces humans in driving and other maneuvers. This is only the most basic function, which is based on the automation of basic driving behavior. With the development of technology, more and more intelligent technologies will be applied to the economy and comfort of automobiles. As a young person, you can imagine that in the future city, everyone can sit in a dedicated seat, the car will spread its wings and fly in the air, music is playing, coffee is warm, the people you like are around, and beautiful scenery is in front of you, and every journey is an enjoyment.