laitimes

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

Written by / Jinyang Shi (Institute of Modern Physics, Chinese Academy of Sciences) Chen Kun (Institute of Physics, Chinese Academy of Sciences)

This article is from the magazine Knowledge is Power

During the holidays, many people will choose to travel by car, but the traffic congestion is too good, if you can achieve automatic driving! How is autonomous driving in modern vehicles achieved? What role does radar play in this?

What is radar?

Autonomous driving is like a pedestrian walking, first using the eyes to observe and determine the route, and then the brain gives instructions to the body. For the car, various radar, cameras and other sensors are the eyes of the car, the electronic circuit is the central nervous system, and the control system such as calculation and analysis is like the brain, which ultimately determines the forward direction and speed of the car. In these organs, the "eye" plays a very critical role, in order to make the vehicle as keen as a human perceive the surrounding environmental information, we need to install a pair of eyes for the car - radar.

The word radar is derived from the English transliteration of Radar, which is an abbreviation for "radio detection and ranging" in English. As the name suggests, the role of radar is to find the target and determine the "position" of the target, which is very similar to the function of our eyes.

So how does radar work? As an example, when we are in a dark environment without lights, it is difficult for us to distinguish objects around us with the naked eye. At this point we turn on the flashlight, the surrounding becomes clear and the relative position of the surrounding objects can be more accurately located. This process of finding objects around can be broken down into the following steps:

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

The brain's process of finding objects (Courtesy of Shi Jinyang)

First of all, we use a flashlight to emit light, so that the flashlight shines on the surrounding objects, because the surface of the object can not fully absorb these lights, so it will reflect some light into the environment. The eye then receives these reflected lights and converts the light information received by the eye into a nerve impulse signal. Nerve impulse signals are conducted through nerves to reach the brain. Finally, the brain is judged to be aware of the presence of surrounding objects and determines the relative position of objects based on the angle of both eyes. Of course, the above process is completed in only a split second.

If the flashlight in the above process is replaced by a radar transmitter, the eye receiving the light is replaced by the radar receiver, and the radar signal processing system is equivalent to the brain, we can visually understand the operation principle of the radar system. Radar determines the distance from the target by the interval between the pulse echo and the transmitted pulse. The propagation speed of the electromagnetic wave in the local environment is then measured experimentally, and then the flight time of the electromagnetic wave can be determined by determining the interval between the emitted pulse and the reflected wave. When we have two parameters, velocity and time, we can use half the product of velocity and time to calculate the distance between the two.

24/7 "Service Butler"

- Millimeter-wave radar

Millimeter wave refers to electromagnetic waves with wavelengths between 1 and 10 nanometers, and millimeter wave radar refers to radar operating in the millimeter wave band. Millimeter waves have wavelengths between centimeter waves and light waves, so millimeter waves have the advantages of both microwave guidance and photoelectric guidance. Millimeter-wave radar benefits from a special high-gain antenna (the higher the antenna gain, the better the directionality) design, which is extremely directional and can accurately identify the relative position of obstacles. If when the directionality of the radar is relatively poor, it will illuminate a large range, then if there is an obstacle, it is difficult to accurately measure its relative position, only to determine that it is near a certain location in this range. When the directionality of the radar is better, once the object is detected, the position of the object can be more accurately framed in this smaller range.

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

Driverless car equipped with radar (Courtesy photo/Shi Jinyang)

Because millimeter wave has good penetration, millimeter wave radar can still accurately perceive the surrounding environment in the smoke and dust environment, and has the characteristics of all-weather all day. These characteristics are extremely important for autonomous driving, which can distinguish between passing vehicles, people, low roadblocks and small obstacles under poor visibility conditions, avoiding vehicles from making wrong decisions due to lack of perception of surrounding information, thus providing effective protection for passenger safety.

The "fragrant food" of the driving world

- Lidar

In order to achieve assisted driving of vehicles or higher-level autonomous driving, vehicles must be able to perceive their own environment, which is a new challenge for the automotive industry and the most critical step to achieve automatic driving.

Lidar can also be called Lidar, which is a collection of laser and radar. In fact, everyone is not unfamiliar with it, because when many driverless cars are experimenting on the road, there will be a "small jar" on the roof that looks similar to the camera, but will always rotate within a certain angle, which is usually lidar.

Lidar is an active detection method that uses light waves to make measurements. The active detection method refers to the detection system by receiving the signal echo emitted by itself to measure, which is different from the passive detection method of obtaining the signal by receiving ambient light, such as a camera. Lidar calculates the distance of an obstacle by measuring the time it takes for the laser to be reflected through the obstacle to be received by the sensor.

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

Scanning perception map of lidar (courtesy photo/Chen Kun)

For example, millimeter-wave radar can spot roadside obstacles but only "see" blurred shapes, while centimeter-level accuracy lidar can clearly distinguish in a very short time whether the obstacle is a shoulder or a slope. If the driverless car concludes that it is a slope, it can make the decision to drive safely into the lane. This precision can be closer to extreme safety for driverless cars on the road.

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

Lidar ranging renderings

Since the laser wavelength is 3 orders of magnitude shorter than millimeter wave, the resolution accuracy is much stronger than that of millimeter wave, which can clearly identify surrounding vehicles, pedestrians and obstacles. However, due to the fact that most objects have absorption and reflection in the laser band, the propagation distance is also much shorter than that of millimeter waves, and in rainy days, the laser may be absorbed or reflected by raindrops, interfering with driving safety. Since vehicle-road collaboration is the trend of autonomous driving in the future, lidar can also be applied to the end of the road, which is scanned in real time by the end of the road and pushed to nearby vehicles in synchronization.

Lidar and millimeter-wave radar are the right and left arms of the car to explore the way. LiDAR handles high-precision modeling in close proximity to ensure driving safety on complex streets. Millimeter-wave radar handles low-precision modeling at a distance to ensure driving safety in complex weather.

It is expected that the driverless cars of the future can be equipped with low-cost lidar and millimeter-wave radar to comprehensively cope with complex road conditions and make travel safer and traffic smoother.

Knowledge development

Lidar "Big Anatomy"

Lidar is mainly composed of a laser transmitting module, a receiving module, a scanner, a lens antenna, and a signal processing module.

The laser transmission module mainly emits laser pulses and regulates the laser beam; the laser receiving module mainly receives the returned laser and generates the signal; the information processing module mainly analyzes and calculates the signal amplification and modeling. The main role of the scanner is to drive the radar in a certain form to scan the surrounding situation in one or more planes, and the scanner is generally divided into mechanical, rotary mirror, micro-mirror and flood surface array.

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

The mechanical scanner is to install the radar scanner on the roof of the car, mechanically rotate the entire radar scanner to achieve scanning (courtesy photo/ Chen Kun)

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

The rotary mirror scanner is a mechanically rotatable planar mirror that controls the beam discharge angle to achieve scanning, similar to the mechanical type (Courtesy photo/Chen Kun)

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

Micro-mirror scanner by adjusting the micro-mirror emitted electronically controlled beam to achieve scanning, the disadvantage is that it can not scan 360 degrees (courtesy photo / Chen Kun)

From self-drive to self-driving! How is autonomous driving achieved? What role does radar play in this?

The flood area array scanner has a built-in beam expanding mirror, which can quickly obtain data in a single direction by firing a divergent laser beam forward, with the disadvantage of low detection distance and low imaging accuracy (courtesy photo/Chen Kun)

Now ordered for 2022

Knowledge is Power magazine

Only 300 yuan

Read on