You are still racing on the road late at night, in a hurry, your electronic recorder is almost full, and your truck is seriously overloaded. The rain is getting bigger and bigger. You think it is too fast, but the surrounding cars are not slow at all.
Suddenly, a burst of alarm makes you refreshed. You instinctively observe the front through the headlights-you can't see anything. But your car has started emergency braking. You glance at the display on the dashboard, where a red image is getting bigger and bigger, an overturned truck blocking the lane around the bend. Your truck stopped by itself, and the headlights illuminated the rescuers who were trying to rescue the truck driver.
Welcome to the world of radar-assisted driving.
In the next few years, radar will play an important role in the development of advanced assisted driving systems (ADAS). With the expansion of its role, the functions of radar transceiver, signal processing and automatic hedging will make the vehicle-mounted ADAS system more and more like the tactical system in combat aircraft, which will have a great impact on the basis of automotive system design.
Why is radar?
Most of the discussion on ADAS has focused on passive vision systems using visible wavelength cameras. And Ralf Reuter, a radar system engineer at Freescale Semiconductor, persuasively demonstrated the role of 77-GHz radar. Reuter talked in a recent interview: "The important point is that [ADAS sensor technology] only radar is not related to the weather. The camera has an advantage in identifying targets, and the radar is better at detecting objects and measuring their speed." Reuter explained that for these reasons, many early systems that value detection and risk classification assessment will choose radar. He pointed out: "In Europe, there is a great demand for truck emergency braking systems. It is based on radar."
The radar system will activate a simple mid-range system, looking straight ahead at the entire road. However, it will soon develop into a multi-sensor system that includes long-distance forward looking and short-distance 360-degree risk assessment functions, as shown in Figure 1.
Figure 1. The radar system can perform a forward search and observe the surroundings of the vehicle.
The picture is authorized by Freescale Semiconductor
Although the optical vision system is very mature, the advantages of the radar system will make it more perfect. Reuter predicts that in the near future, multi-camera complex systems with target recognition will be able to fuse video and radar data to build a dynamic model of the surrounding world.
Acquire the Signal
Understanding the impact of radar on automotive system design starts with understanding sensors. Most automotive radar designs are willing to use the 24 or 77-GHz free frequency band. The transmitter is generally a frequency-modulated continuous wave (FMCW) design, which emits "chirp": rapid changes in the frequency domain, as shown in Figure 2.
Figure 2. The onboard radar will use the chirped FMCW.
The picture is authorized by Freescale Semiconductor
Reuter explained: "The biggest advantage of FMCW is that it simplifies the understanding of the reflected signal. You can directly read the target range from the reflected frequency and derive the speed from the Doppler frequency shift. Compared to the pulse modulation scheme, the CW is generated Not too complicated, easy to understand, and reliable. This is the most concern of vehicle manufacturers, who feel that every euro spent on improving ADAS comes directly from corporate profits.
Receiving signals also requires low-cost novel designs. To collect reflected signals by collecting azimuth information requires a mechanical scanning antenna or a phased array antenna, combined with a digital beamforming algorithm. Behind the antenna is usually a homodyne receiver with many channels, which is required for antenna design-for simple rotation of the antenna, there are 16 in an array.
Reuter said: "The output of the receiver is a baseband signal in the DC-20-MHz band." A system designed to achieve good azimuth resolution will have 8 to 16 channels and require 8 to 16 high-speed analog-to-digital converters (ADCs). .
Extract information from chirp
The digital baseband signal from each channel flows into the fast Fourier transform (FFT) module to achieve a transform of up to 2K samples in length. Reuter said: "In the past, it took a lot of FPGA to perform FFT. Today, the trend is to integrate a floating-point DSP accelerator 32-bit microcontroller." The beamforming system can use two FFTs to extract range and speed data from the signal ,As shown in Figure 3.
Figure 3. FFT configuration for bunching and range and speed estimation.
The picture is authorized by Freescale Semiconductor
In fact, this front-end processing unit processes multiple input FMCW analog channels to form a digital stream of azimuth / range / speed arrays. This data stream enters one or more CPU cores, and the software supported by other accelerators will infer whether there are objects around the vehicle, their location and attributes.
Reuter explained: "You need to identify the targets, separate them from the background, and choose the most critical one. You may need to process 200 targets, so the calculation will be very complicated, especially to extract the angle information.
The physical configuration of the system is also getting more and more complicated. Reuter said that currently there is very little processing work on the sensor itself. In contrast, ADCs have dedicated analog interfaces, dedicated digital interfaces for signal processing hardware are used for FFTs, and microcontrollers have other interfaces to extract targets and classify them. The target information is input to the vehicle control area network (CAN) or FlexRay bus, and the central CPU cluster interprets and analyzes it.
The entire pipeline has great bandwidth and delay requirements. Reuter said that the CPU's interpretation of the data is generally presented to the driver in a graphical display, that is, the front multi-function display that he can observe through the windshield. This hybrid display requires a maximum update interval of 50-ms and a more challenging 50-ms maximum delay. Otherwise, the image will be very unstable, and the image through the windshield will lag, which may cause the driver to misjudge.
As the system evolves from one sensor to multiple sensors that support the beamforming function to achieve camera data fusion, the interconnection architecture has also changed. Reuter said: "There is a need to use Ethernet to reduce costs." However, the system still requires real-time work, which brings the problem of how to ensure the real-time performance of Ethernet.
The future of interference is everywhere
As long as no one around uses radar, vehicle-mounted radar is generally more reliable. However, this has obvious problems: more and more vehicles use radar, so interference between the equipment will inevitably occur. Reuter said that you can change the modulation rate to reduce interference. Eventually, the coded frequency hopping pattern is used to replace simple frequency changes, so each car can recognize its own chirp. This change can maintain the architecture of the early system and most of the hardware unchanged, but to achieve complex extraction and analysis functions, each system is required to be able to recognize the reflected signal from its antenna.
Moreover, there is another type of problem that is not easy to solve: interference caused by fixed equipment. Reuter reminded: "The immediate problem is that European tunnels use high-power radar to identify vehicles. Their transmitters can cause vehicle-mounted radars to not work."
A more scientific and sensitive issue is that the key part of the astronomical object's radiation spectrum is located in the 77-GHz band. In densely populated areas, more and more vehicle-mounted radars will strongly interfere with astronomical RF signals.
Reuter reminded: "There are already regulations in Japan that require the transmitters within 1,000 kilometers of the RF astronomical telescope to be turned off. This may cover the entire country."
Chirped modulators, digital beamforming, target recognition, hazard assessment, protective frequency hopping, jamming, etc. — these all sound like new combat aircraft, not trucks and cars. In fact, ADAS inherits many technologies of military systems. This is not surprising. When the weather is getting worse and the cars are getting more and more jammed, it really looks like a war.
Cordless Floor Cleaner is a new and smart vacuum mopping cleaner. recharge floor cleaner very small,free,simple use. Latest floor wipe cleaner can vacuum ,mopping,wipe home,hotel,office clean area.recharge mopping cleaner very good design,use simple and freely,
Wireless hand floor cleaner advantage:
Cordless hand floor cleaner with special design and best quality
Vacuum ,mopping function
Fast and silent and smart floor hand cleaner
Cordless floor cleaner facotry in shenzhen,china,welcome to visit facotry,we are 15 years manufacture experience.
Cordless Floor Cleaner
Cordless Floor Cleaner,Robot Vacuum Cleaner,Cordless Hard Floor Cleaner,Cordless Floor Cleaning Machine
Zhengzhou Bangmi Smart Technology Co., Ltd. , https://www.cleanwindow-robot.com