| Autonomous Driving, the Ultimate Goal of Automotive OEMs
| Level 3 autonomous vehicles, no commercial models yet
| Radar, excellent object recognition even in adverse weather conditions

Autonomous vehicles are the ultimate goal of automotive OEMs.

Existing mechanical components are not capable of autonomous driving. A car with only machines can move, but it cannot determine direction and speed. Driver intervention is absolutely necessary. Electronic components are needed to reduce driver intervention by even a single point.

Nowadays, cars are not just moving machines, but moving home appliances. The battlefield, ADAS, in-vehicle infotainment, and autonomous vehicle markets are becoming the next battleground for the electronics industry.

As the automotive semiconductor market grows by over 10% annually, the 'Automotive Innovation Day 2019' was held on the 3rd at the The K Hotel in Seocho, Seoul, sharing the latest automotive technology trends.

With many experts from academia and industry filling the session, Kim Myung-gil, head of the Automotive Business Unit at Infineon Technologies, the second largest automotive semiconductor company, took the time to explain the principles of radar, a technology for autonomous driving.


Definition of autonomous driving
What is autonomous driving? In 2016, the Society of Automotive Engineers (SAE) rated autonomous driving from Level 0 to Level 5. Level 0, “Warning,” is a state where there is no autonomous driving technology at all. The sensors on board the vehicle only warn the driver. All decisions and actions while driving are entirely up to the driver.

Level 1 'Assistance' applies the brakes in case of an emergency. Again, all decisions and actions that need to be made while driving are entirely up to the driver. Only emergency braking is possible. Level 2 'Partial Autonomy' means the car steers for you. Sometimes. The driver is still afraid to take his hands off the steering wheel.

True autonomous driving starts at level 3, “conditional automation.” Up to levels 0, 1, and 2, the driver is responsible for monitoring the car. From level 3, it becomes the system’s responsibility. At level 3, autonomous driving is only possible when the environment is appropriate. In addition, as we move from Level 4 'high automation' to Level 5 'full automation', there will be no need for drivers at all, but this is still a story for the future and it is not yet time to discuss commercialization.

The industry also calls the transition from Level 2 to Level 3 a “quantum jump.” There are no commercial Level 3 cars yet. It’s not because the industry lacks technology. It’s because responsibility for autonomous driving accidents is unclear and the infrastructure is lacking. That’s why auto OEMs and auto parts manufacturers are focusing on developing ADAS technology that can be commercialized while waiting for autonomous vehicle technology to mature and infrastructure to be built.


The first step to autonomous driving: ADAS and radar
ADAS (Advanced Driver Assistance System) is a system that assists drivers in driving, reducing driver fatigue and helping to ensure safe driving. BSD, one of the key technologies of ADAS, is a device that obtains and provides information about cars located in the driver's blind spot using radar. A 24㎓ radar sensor that covers a short distance is mainly used.

In particular, Adaptive Cruise Control (ACC) is a system that automatically maintains an appropriate distance from the car in front using radar mounted on the front of the vehicle, without using the accelerator pedal or brake pedal. A vehicle equipped with ACC accelerates the car only to the set speed when it detects that there are no other cars on the driving path. Radar is also key in ACC. A 77㎓ radar sensor, which is longer than BSD, is used.

The radar module implementing ACC consists of a single ECU (Electronic Control Unit). In detail, it includes an RF front-end, a signal processing section, and an MCU that processes AAC logic.

The RF front-end extracts the distance, speed, and angle of a specific target by shooting and receiving 77㎓ FMCW radar, and targets the target. The signal processing section determines whether to accelerate or decelerate based on this, and sends instructions to the throttle actuator and brake actuator through CAN (Controller Area Network). Recently, as the number of radar sensors installed in automobiles increases, the domain controller is trying to take the judgment logic.


Why 77㎓?
But why use the 77㎓ radar band? This is related to the antenna size. The wavelength of the 77㎓ radio wave is 3.9㎜, which is advantageous for making the module small. Accordingly, in April 2001, Korea allocated a 1㎓ bandwidth of 76~77㎓ as a vehicle radar frequency in accordance with Article 9 of the Radio Act.

The 77㎓ radar band belongs to the millimeter wave (mmWave). mmWave is a radio wave with a wavelength of 1 to 10 mm and a frequency of 30 to 300 GHz. mmWave allows antennas and transmitters and receivers to be made smaller and lighter. It also has good directivity and basically uses electricity, so it has little impact on the human body.

The biggest advantage is that it has relatively little error even under various weather conditions. This is the most contrasting feature with cameras that can obtain the most information but are vulnerable to weather conditions such as snow, rain, and fog.


Principles of automotive radar
Automotive radar transmits radio signals from the driver's vehicle, receives radio waves reflected from other objects, and estimates the distance and relative speed between the radar and the other object by using the time difference and Doppler frequency shift between the two signals.

Frequency Modulated Continuous Wave (FMCW) radar is a radar that transmits radio waves as a continuous wave, but the frequency of the radio waves continuously changes over time. Due to the characteristics of mmWave, it can be operated with low power and has superior distance resolution to objects compared to other radar systems.

The FMCW radar transmission signal is generated as a linear radar transmission signal with a center frequency using a voltage controlled oscillator (VCO), and the signal that passes through the transmission antenna is transmitted toward the target.

Depending on the distance between the radar and the target, a time-delayed reflected signal is received through the receiving antenna, and the signal generates a sine wave with a beat frequency component through dechirping.

The beat frequency signal passes through a multi-stage amplifier with a gain of approximately 60 to 70 dB and a low-pass frequency filter with a band of several tens of MHz, and is sampled through an analog-to-digital converter (ADC) and converted into a digital reception signal.

The converted digital signal has its frequency components extracted through a spectrum analysis algorithm, and the relative distance between the reflected signal and the radar can be determined through the extracted frequency. In addition, the speed and angle of the target can be extracted through the phase difference between each received signal.

Due to the nature of frequency-modulated continuous wave radar, unlike pulse Doppler radar, it samples only the bit signal bandwidth rather than the frequency bandwidth, which reduces hardware costs.


An irresistible flow
Autonomous vehicles are the ultimate goal of almost all automotive OEMs. However, autonomous driving is not possible with existing mechanical components. Electronic components are needed. The biggest role that a driver has to play while driving is monitoring.

Drivers must monitor their surroundings and make appropriate decisions at every moment to ensure smooth driving. Combining electronic components makes it possible to monitor the vehicle's surroundings and respond accordingly. Sensors are a key component of monitoring.

Radar, which has been used since 1935, is an excellent monitoring tool that has been recognized for its usefulness over a long period of time. Radar is now being actively adopted into vehicles, helping the industry on its journey toward autonomous driving.