Starting from Lv0 warning level in 2005, ADAS stages have progressed beyond current Lv2 integrated driving assistance to reach Lv3 partial autonomous driving. In fact, German automotive company AUDI launched the A8 vehicle at the first Lv3 autonomous driving level last year, and vehicles are currently being supplied according to regional traffic regulations.
David Hall, Senior Marketing Manager at NI, states that the development of autonomous vehicle technology is creating numerous challenges in the testing field. The radar system is one of the most readily anticipated areas.

Currently, the radar measures values and transmits them to the driver, who confirms and applies them to driving. However, in the future, the radar will be used to inform the ECU what action is needed, allowing the ECU to make decisions accordingly.

This means that from a tester's perspective, the focus shifts from simply measuring whether the radar correctly detects objects, to simulating whether object detection and the ECU's response speed and response methods are accurate.

We listened to a detailed explanation from David Hall, Senior Marketing Manager at NI, regarding methods for this comprehensive radar testing.

The focus of radar systems we currently observe has rapidly shifted from a mechanism that notifies the driver of potential problems during driving to a method that embeds more functionality within the ECU itself. In the past, the radar informed the driver of issues, but now it is evolving toward a direction where the ECU makes decisions through communication between the radar and ECU.

Radar testing aligned with this evolution of autonomous vehicles proceeds in three major approaches.
The first is basic simulation, which simply models radar operation using mathematical methods. This is the most fundamental approach desirable when performing engine ECU testing, where mathematical models of devices or signals can potentially be used to verify how the ECU operates in specific situations.

Another technique is to emulate the radar's own environment. This technique involves reproducing the signature of electromagnetic signals that can occur as a result of radar stimulation that bounces off objects and reflects back.

The final technique is to simply deploy radar devices in the field and conduct actual testing. This technique is very important and highly effective. However, it has the problem of requiring significant time and cost. Additionally, one of the issues encountered in field testing is the difficulty in simulating the most extreme scenarios. In reality, a radar system may operate with 99% accuracy, but the probability of not functioning correctly as intended for testing and identification is around 1% or 0.1%.

Therefore, considering the difficulty in reproducing such scenarios in field testing, it is important to use a combination of field testing and emulation to identify as many bugs as possible not only in device design but also in the device validation phase performed later.

Radar Emulation Technique: The Fastest Advancing Method in ADAS System Testing

The radar emulation concept is to reproduce the electronic signature of the environment surrounding the radar system. Regarding the operation method, a radar system using frequency modulated continuous wave (FMCW) effectively sends signals that bounce off and reflect from surrounding objects.

Once the signal bounces and reflects, the reflected signal returns to the radar system at a specific frequency and delay time according to the device's velocity and distance.
For example, if radar stimulation reflects off a rapidly moving or high-speed object, the frequency shift of the returned signal becomes more significant.

The radar system can determine both range and velocity by measuring the phase and frequency difference between the transmitted and received signals at the receiver side. There are several methods to test this capability of the radar system, and it is effective when using two different types of object simulation.
The first is simple passive object simulation, which effectively transmits signals through delay lines. According to the distance of the delay line, the time required to propagate the signal through the actual delay line can be predicted. By applying basic physics, you can determine how long it should take to transmit the signal matching the delay line distance. Additionally, you can determine how accurate the radar system's range detection is.

Next is to use a method called active object simulation. In active object simulation, there is an object simulator or RF control unit that effectively digitalizes the signal received from the radar system. Then, using signal processing mathematically, a delay combination is applied to simulate distance. Simultaneously, frequency shift is applied to simulate the velocity or Doppler of a moving object.

When a reflected signal is generated that represents how an actual object would reflect, you can use analog-to-digital and RF converters to convert the signal back to the analog domain and retransmit it to the radar system.

Using object simulation has several advantages. First, since all objects can be modeled mathematically, multiple objects can be simulated simultaneously.
Additionally, you can simulate broader movements in terms of actual object movement and more complex scenarios accordingly. However, active simulation techniques have some issues regarding minimum distance.

This is because there are delays associated with digitalization using ADC, some signal processing, and retransmission through digital-to-analog converters. In reality, many engineers are using a combination of the two techniques.

NI Alliance Partners provide solutions that use a combination of passive and active techniques to simulate radar object behavior for radar system testing purposes.

KONRAD System Example Applied to AUDI
As can be seen in the upper right of the figure below, KONRAD creates a software package that allows users to draw device movement, and users can draw object movement relative to the radar sensor.

After loading this diagram into software and running the simulation, the object simulator reproduces the electronic signature of the device moving along that path.
You can then verify that the device is operating properly by comparing the distance reported by sensor velocity with the distance of simulation velocity. One of the customers using this type of technology is AUDI.

Nils Koch, who is responsible for radar components at AUDI, mentions that the broad bandwidth of the PXI system—specifically the combination of KONRAD's software and NI's vector signal transceiver—was the ultimate factor that enabled identifying bugs using software that was previously unknown.

There are several techniques for radar testing. How these techniques are suitable for various situations, integrated vehicle testing, or integrated system testing can be explained in the figure below.
Through this radar test simulator concept, what has just been described is actually hardware testing of the actual radar device, and potentially embedded software testing.

This radar device is used together with an Engine Control Module (ECM) or Engine Control Unit (ECU). Therefore, the ECU and radar system are interconnected. Historically, engineers have tested the ECU and ECM using Hardware-In-The-Loop (HIL) techniques across a wide range of scenarios.

Today, one of the benefits of applying techniques such as PXI's radar target emulation is that some techniques can be combined with common HIL test systems. And it means that in the past, engineers could test the ECU in a simple manner by modeling the I/O they would use and replacing that I/O with a test system capable of scenario reproduction.

Today, the advantages of PXI for HIL simulation include the fact that with PXI, you can simulate not only ECU load and open/short circuit operation but also some of the RF characteristics and synchronize the signals accordingly.

This means you can test the entire ECU or ECM through the entire RF and bus interface simulated using PXI. In fact, the ideal scenario is to test the operation of multiple interconnected ECUs, and such scenarios can be reused with HIL and PXI.

NI is currently releasing solutions that can reduce the cost and time of automotive testing and secure high-level measurement accuracy. Additionally, through NI's platform-based approach, the company provides flexible and forward-looking systems with overall timing and synchronization capabilities for the entire system, capable of meeting complex automotive test requirements including ADAS, V2X, infotainment, powertrain, body, and chassis.