Evaluating 77 to 79 GHz Automotive Radar Radome Emblems
Advanced driver assistance systems (ADAS) in cars assist the driver and increase road safety, and these systems are widely deployed in many modern vehicles and car types. Currently, autonomous driving is a major focus of the automotive industry, and R&D institutes are making headlines with fully automated cars driving hands-off along the highway—even in cities with dense traffic. It is certain that autonomous driving will become reality in the near future.
Camera, LiDAR and radar sensors are key enabling technologies in this area. Millions of automotive radars are produced every year, and car manufacturers are starting to deploy them as standard equipment in all higher-class cars. Today, automotive radar sensors are mainly used to increase driving comfort and prevent crashes. Most automotive radar sensors that enable adaptive cruise control (ACC) operate in the conventional 76 to 77 GHz frequency range to sense other cars and objects far ahead. Advanced radar capabilities, however, demand larger bandwidths, with coverage up to 81 GHz to enable 360 degree radar vision around the vehicle. This is required for ADAS functions such as lane change assistance and blind spot detection, where high resolution and a wide operating angle are essential. Additionally, extending the automotive frequency band to 81 GHz helps mitigate interference. Altogether, this puts pressure on radar system integration to be functional across a wider frequency band than in the past.
Automotive radars must be masked by a cover, known as a radome, which is constructed from a transparent RF material. Radomes can be an emblem or a car bumper, with the radar behind it as illustrated in Figure 1. Although an emblem may be underestimated and dismissed as a simple plastic cover, it is actually a sophisticated RF element that often degrades radar detection range and accuracy. Emblems need to satisfy the requirements of aesthetic appearance, often including the 3D shape of the car logo. However, this is usually in conflict with the RF performance needed for 76 to 81 GHz operation. On the other hand, car bumpers are typically coated with metallic paint, which is, of course, critical for the automotive radar frequencies. So it becomes essential to validate the material characteristics and examine their influence on the radar sensors. Considering the criticality of the applications where radar sensors are deployed, uncertainties are unacceptable. Consequently, engineers and manufacturers need new measurement capabilities to evaluate the effect of emblem radomes and bumpers on radar performance. This article explains a novel radome measurement method and discusses the radome’s influence on the accuracy of the angle of detection with advanced radars.
Automotive radar sensors transmit radio signals at 24 and 77 to 79 GHz. They mainly use frequency modulated continuous wave signals that have the advantages of low power, no “blind” range and low receiver bandwidth, meaning they can be manufactured more cost effectively than pulsed radar systems. The transmitted radio signals are reflected by other objects. Due to the propagation delay and Doppler frequency shift, the radar sensor can measure and resolve range and radial velocity for multi-target situations. Depending on the properties of the antenna array, it is also possible to measure and resolve the azimuth and elevation angles. After the detection process and tracking, signal processing generates a target list that contains values such as the positions and velocities of objects and estimates of type. This list is passed to the vehicle’s electronic control unit (ECU), where it is further processed to deliver real-time decisions for vehicle maneuvers. The accuracy and reliability of this data is extremely important for the safety of the vehicle and its passengers.
The accuracy of a radar depends on many factors, such as the hardware components, software processing and the radar echo signal itself. The parameters of signal echoes with lower signal-to-noise ratio (SNR) can be measured less accurately than signals with high SNR. The effects of multipath propagation and distortion due to radomes greatly impact measurement accuracy. Inaccuracies in the azimuth measurement cause the target to appear misplaced. This is shown in Figure 2, illustrating that an angular measurement error of only 1 degree at the radar sensor causes a target at 100 m range to appear misplaced by 1.75 m in azimuth. The target will be interpreted as located in the neighboring lane. In practice, angular accuracy for such far distances must be significantly less than 1 degree for reliable operation.
For a modern radar sensor with an antenna array receiver front-end, the azimuth (and sometimes the elevation angle) is estimated from phase and amplitude information using digital beamforming on the receive side. To get the best azimuth measurement accuracy, every radar sensor needs to be adjusted. The following procedure is typical for radar calibration. First, the radar sensor is mounted on a turntable in an anechoic chamber. A corner reflector in the far field at a known distance is commonly used as a reference target. The radar pattern is then measured and flashed into the radar sensor’s memory. Later, this information is used by the detection algorithm. This ensures high azimuth measurement accuracy when the radar sensor completes production. The car manufacturer integrates this calibrated radar sensor into the car, often behind an emblem or the bumper, where the radome material influences the RF attenuation. The radome’s transmission loss increases the two-way attenuation of the radar signal, which reduces the maximum detection range of the radar. The power level of a transmitted radar signal is reduced by the range, R, to the target and by R4 on the return. For a 77 GHz radar system with 3 W output power, 25 dBi antenna gain, a target with a 10 m2 radar cross-section and a minimum detectable signal of ‐90 dBm, the maximum range would be 109.4 m, according to the range equation. If the radome has 3 dB two-way attenuation, the maximum range of the same radar measuring the same target would decrease to 92.1 m. That is approximately 16 percent less range.
It is not only material attenuation that has a great impact. The material reflectivity and homogeneity also play important roles. Reflections and RF mismatch of the material cause direct signal reflections in close range to the radar. The signals are received and down-converted in the receiver chain, reducing the radar’s detection sensitivity. Many car manufacturers try to mitigate this effect by tilting the radomes—not only for design reasons—to reflect the transmitted radar signal somewhere other than directly back into the receiver front-end. This solution, of course, has mechanical limitations, not to mention the expected loss of RF energy from these parasitic reflections. Another problem comes from material inhomogeneities that disturb the echo signal wavefront used to estimate the azimuth value. Inhomogeneous material distorts the wavefront, which results in less accurate angular measurements.
The radar sensor calibration is no longer sufficient, since the previously calibrated radar can be mounted behind any indeterminate radome from a different manufacturer (see Figure 3).
Radome manufacturers typically test their units with a known or “golden” radar. For this test, several corner reflectors are mounted in front of the radar at predetermined ranges and azimuth positions (see Figure 4). A differential measurement is conducted with and without the radome, and these measurements are compared. For the radome to pass the test, the range/azimuth positions and echo signal levels must be within defined limits. However, this approach only tests certain azimuth angles and does not account for possible inhomogeneity or blind spots.
Another measurement method relies on a functional test. The radar sensor with the radome is mounted on a turntable and a corner reflector is placed in front of them. By turning the complete unit, every azimuth and elevation angle can be measured and compared to the radar-only standard. This method is as accurate as the positioning of the turntable; however, the test takes a long time and is not feasible for production tests.
Instead of testing transparent radar material with a golden device, a proposed novel measurement method combines a transmission measurement with three-dimensional, high resolution radar imaging in the 77 to 79 GHz frequency band used by the radar itself. This is done using the R&S QAR system (see Figure 5). A multi-antenna array consisting of several hundred transmit and receive antennas operating from 75 to 82 GHz is used. This measurement system can measure the range, azimuth and elevation with millimeter resolution. It operates in the same frequency band as the automotive radar and “sees” what the automotive radar would see if it also had hundreds of transmit and receive antennas. Thanks to the large aperture, the resolution of the test system is much higher than that of the automotive radar, and it can visualize the measurement as an image. The radome is placed in front of the test system, which performs a two-stage measurement.
First, a reflectivity measurement determines the amount of energy reflected by the radome material; this is energy that does not pass through the radome and contributes to performance degradation. Reflected signals decrease the performance of the radar and can even interfere with the received signals. Areas with high reflectivity can have various causes, such as material defects, air gaps, undesired interaction between layers of material, excessive amounts of certain materials or foreign objects. The measurement method achieves a spatially resolved reflectivity measurement for a radome by linking the information collected by the distributed, coherent transmit and receive antennas. The receive signals are gated and processed for all receive antennas, which results in a high resolution 3D radar image. The resulting mmWave image enables intuitive as well as quantitative evaluation of the radome’s reflection behavior.
To illustrate the approach, a radome was manufactured where the R&S logo was milled with different thickness, as shown in Figure 6. The high resolution radar image in Figure 7a visualizes what an automotive radar sensor would perceive when covered by this radome. The color scale shows the reflectivity, where the dark color indicates minimal reflectivity and bright high reflectivity. Metal, which cannot be penetrated by automotive radar signals, appears as white (e.g., the screws in the four corners). The radome image indicates high reflectivity and the inhomogeneity of the logo, showing the increased thickness of 0.5 mm in the logo area is sufficient to cause major disturbances in radar performance on the street. In this example, calculating the mean reflectivity in the middle area, where the radar sensor is usually mounted, yields approximately ‐11 dB with a standard deviation of ‐17.7 dB. In many practical cases, this reflection is too high to maintain acceptable radar operation. In practice, the expected reflectivity depends on the sensitivity of the radar and the maximum detection distance to be covered.
In a second test, the frequency matching and attenuation of the radome material is measured. A transmitter module is located behind the radome on the table. The transmitter uses a frequency sweep to cover a selected frequency span. This allows the radome’s transmission frequency response to be measured. The frequency response delivers detailed information about the RF matching of the radome at the frequencies intended for radar operation. It is independent of the signal waveform used by the radar, which facilitates the testability and optimization of the radome itself.
The measured one-way attenuation versus frequency of the radome is shown in Figure 7b. Since automotive radars operate in the 76 to 81 GHz band, attenuation should be low across this range. Depending on the thickness of the material, its air gaps and RF matching, a good radome should maintain low attenuation across the desired frequencies. The logo example shows 0.64 dB one-way attenuation with better matching at 79 than at 76 GHz. A more sophisticated example for commercial radomes with a 3D design typically results in a transmission measurement, as shown in Figure 8. This radome would have various performance issues:
- Frequency matching is incorrectly placed around 71 instead of 76 GHz, which is often caused by increased thickness of some radome layers
- Significant increase in the standing wave ratio within the 79 GHz band, which identifies high reflections at the radome boundaries and a strong interference phenomenon
- Overall one-way attenuation is relatively high, which results in a significant reduction in the detection range.
Researchers are already driving autonomous cars on the highway and in city traffic. Due to the all-weather capability of radar sensors, their reliability and their price, these sensors are essential in automotive applications. Integration behind bumpers and emblems also makes them attractive for car designers. Without high-quality radomes, objects with low radar cross-sections, such as pedestrians, will likely be undetected and may even appear at erroneous azimuth angles. Even larger objects at far distances may be incorrectly identified and measured at different positions when the radar signal is distorted by the inhomogeneous radome material.
This article presented a novel measurement method that can be applied for any kind of automotive radome operating in the 75 to 82 GHz band, such as the ones used in emblems, bumpers or front grills that cover automotive radar sensors. Using a massive multistatic array, within a few seconds this method measures and calculates the mean reflectivity and standard deviation of a defined area (homogeneity) and the transmission loss over the complete frequency range. The visualization of the results as an image assists radome designers and R&D labs, while determining a pass/fail result based on the radome’s RF performance significantly speeds production, especially at end-of-line testing.
Watch a video demo of the Rohde & Schwarz QAR system at http://videos.microwavejournal.com/video/Automotive-Radar-Radome-Chara-2