What radar systems does FORVIA HELLA develop and what do you work on there?

Dr. Andreas von Rhein: FORVIA HELLA has been developing and producing radar sensors for automotive applications for more than two decades – now in its fifth generation. The radar sensor records speed, size, and distance information from objects in the vehicle’s enironment. The sensor technology is suitable for the now established standard functions such as adaptive cruise control (ACC), emergency brake assist (AEB), blind spot detection (BSD), and rear traffic alert (RTA).

We are currently working on the preseries of the sixth and development of the seventh generation of radar sensors. I am involved in every task that this entails: from the basic idea to the selection and definition of the hardware and hardware components to the details of the powerful software implementation. This also includes simulation and testing.

What characterizes your radar sensors?

Dr. Andreas von Rhein: Our sensors are state of the art with a good price-performance ratio. They provide 360° vehicle environment detection with a scalable platform approach from Entrance to High Resolution variants. The radars are usually located behind the bumper. This makes us a “hidden champion” in the truest sense of the word. As a design element of the car, the covers are not necessarily optimized for radar applications. But even in this challenging environment, our sensors do a very good job. For example, you can detect an object 200 meters away with centimeter-level accuracy over a wide field of +/- 80° azimuth.

What sets us apart is our ability to implement customer requests quickly. Feedback is immediately incorporated into development so that we can supply each customer with sensors tailored to their individual needs.

How does dSPACE help you with your development activities?

Dr. Andreas von Rhein: In order to examine the quality of our high-frequency signals, we needed powerful test hardware for performance and signal analysis. We initially built a target simulator ourselves for testing and learned a lot during the implementation phase. With our growing success in the radar business, we also had to expand and multiply our test systems. As this is not our core business, we looked for a strong partner for the development to series maturity, the rollout. This is where dSPACE came into play with DARTS.

Why did you choose dSPACE?

Dr. Andreas von Rhein: We have also worked together in other areas and have had very good experiences with dSPACE. We value the open, transparent, and trusting cooperation. Requests for changes, improvements, or extensions are gratefully accepted and, where possible, implemented relatively quickly or, if desired, discussed at length in order to find a joint solution.

How do you use DARTS in the development of your radar sensors?

Dr. Andreas von Rhein: In the course of development, we have to ensure that the specifications are adhered to at every stage of software and hardware development and that we do not overlook any errors. We use DARTS to validate that the radar sensor meets the specifications in the data sheet. This includes the maximum distance at which objects with different radar cross section (RCS) values can still be detected, precision of distance and speed measurement, and also accuracy. To do this, we use a test chamber with a robot to move the radar sensor and two DARTS that allow us to determine the properties of the sensor. In addition to precision, another important point is the separability in the various dimensions of distance, speed, and azimuth. To facilitate these tests, we purchased the DARTS 9040-G and DARTS 9042-G, as this allows us to simulate two targets that can be adjusted independently of each other in all these dimensions. But as I have seen, you now also have an improved DARTS, which can simulate two such targets on its own and thus enables such separability tests with just one device.

Yes, that’s right. Thanks to the feedback from FORVIA HELLA in particular, we have developed a solution that enables separability testing of two targets from a single DARTS – the 9040-GT.

What do you see as the greatest advantages of DARTS for your fields of application?

Dr. Andreas von Rhein: DARTS is easy to use and has a fast and versatile application options with all the necessary information. In addition, our radar sensors must be able to accurately detect very close and distant targets. With DARTS, you can test exactly this – and generate targets at distances of three to three hundred meters. In addition, DARTS’ high instantaneous bandwidth of 5 GHz is a key element for us. Our radar sensors use a bandwidth of 2 GHz in the field and 4 GHz for the test software. In addition, the center frequency of our radar sensors jumps from the lower 76 GHz band to the upper 79 GHz band within one cycle. For this reason, we need a target simulator that covers the entire band instantaneously and enables this frequency hopping. We can run the tests at different frequencies without having to adjust the DARTS.

Another advantage is the easily adjustable/manageable (RCS) amplitude over this high current bandwidth. The accuracy with which the radar cross section (RCS) – the area equivalent of a target – is detected over a large frequency band is very important. Customers are becoming increasingly sensitive to small differences in the signal-to-noise ratio (SNR) and, therefore, in the detection range. DARTS gives us all that.

You also use AURELION. What advantage do you see in using both validation methods?

Dr. Andreas von Rhein: We use AURELION to calculate the beam propagation (raytracer) in conjunction with our own sensor model that simulates the behavior of the radar front ends. The radar waves simulated with AURELION are the input variables with which we then calculate the ADC values. The output of the ADC is then fed into our resimulation, in which it is then no longer possible to distinguish whether the simulated test data or data from a real test vehicle is being used. Object tracking and ADAS functions can then be tested with synthetic data. Debug data is output, just like a radar in a real vehicle setup. The calculation is carried out with a powerful graphics chip. With this setup, we can reproduce test scenarios exactly and thus verify changes in the sensor software. AURELION also enables us to simulate future radar sensors under realistic environmental conditions before we convert them into hardware. We can also simulate the OEMs’ test cases in a multi-radar sensor setup. AURELION is therefore used for the final analysis of the object tracking and DARTS for the investigation of hardware and software effects on the radar detections. With AURELION and DARTS, we have two optimal test systems for our use cases.

Where do you see the future challenges for the development of radar test systems?

Dr. Andreas von Rhein: Radar target simulators must simulate the physical characteristics of real targets. The first steps have been taken, with point targets with a Doppler offset of 10 m/s and a range of 100 m, for example. This gives us synthetic targets with predefined parameters. Of course, we can change these and modify targets based on expected movement profiles. However, these targets cannot yet simulate the complex micro-Doppler signature of “real” targets. The simulation of interference is also still in its infancy. In addition, the new radar sensors have a higher distance resolution, which results in a higher scattering of the distance peak of an object. This should also be simulated by test hardware in the future.

Dr. von Rhein, thank you for talking to us.