Driving towards the automotive radar test

Driving towards the automotive radar test

Technology News |
By Julien Happich

Radar has the unique ability to instantaneously detect objects and measure their velocity via the Doppler shift of radar signatures and performs even in bad weather conditions such as rain, fog and snow. It enables applications like adaptive cruise control, lane change assist, autonomous emergency braking, and stop-and-go traffic jam systems to name a few.

These benefits are driving automakers to adopt radar in increasing numbers. Recently the National Highway Traffic Safety Administration (NHTSA) in the US reached an agreement with 20 automakers, to equip all production vehicles with low-speed Automatic Emergency Braking (AEB), including forward collision warning systems by 2022, a safety feature often enabled by radar. 

As vehicles evolve from Advanced Driver Assistance Systems (ADAS) to full autonomous driving, sensors such as radar and cameras are the critical input devices that enable the vehicle to accurately sense the environment around it. Future vehicles may include several types of radar sensors for short, mid and long range to provide a full 360° surround view of the car. When combined with other sensors in the vehicle, this information provides the context needed for the vehicle to make decisions.

Traditionally many of the radar sensors use frequencies in the 24 Hz band, but due to spectrum regulations the use of some of these frequencies will be phased out by 2022 in Europe and the US. Regulating authorities have instead opened up the 77GHz band, and most research and new automotive radar applications therefore focus around the 76-77 and 77-81GHz band. The benefit of having a higher frequency with higher bandwidth radar sensors is that it improves range resolution by almost 20 times, and velocity resolution by 3 times, whilst reducing sensor size compared to the 24GHz sensors.

The growth and advancement in automotive radar has led to several challenges for the test and validation of radar sensors. The first set of challenges center on meeting the increasing technical requirements for testing modern automotive radar while maintaining or lowering the cost of production test. It is typical for a modern radar sensor to require 1 GHz of bandwidth at 76-77GHz, and few companies have the expertise to build test systems in this frequency band. Higher bandwidth sensors provide finer resolution, and radar manufacturers have already demonstrated sensors with higher bandwidths approaching 4 GHz with a 79 GHz center frequency.

While the technology in radar sensors continues to improve, in order to enable the broad adoption of radar sensor and meet the price and volume requirements, manufacturers need to reduce testing costs and time.

Early radar sensor manufacturers used large anechoic RF chambers and corner reflectors to functionally test and calibrate modules. These chambers were commonly several meters long and consumed large amounts of manufacturing floor space.  To reduce floor space, radar functional testing evolved to use analogue delay lines to emulate long-distance radar obstacles followed by a second test station to perform parametric measurements of the radar. 

Technology for radar functional testing has further evolved, as showcased by the NI Vehicle Radar Test System (VRTS). The VRTS is a hybrid simulator built around the NI PXIe-5840 Vector Signal Transceiver (VST), which integrates an instrument-quality RF Vector Signal Analyzer with a RF Vector Signal Generator via a high-performance, low-latency Xilinx FPGA. The FPGA can calculate radar reflection signatures in incredibly short periods, allowing the RF Vector Signal Generator to generate a corresponding radar signal back to the radar sensor. This approach can consolidate a radar module production test cell by combining the functional test (object simulation) and the parametric tests into a single tester. The combination reduces manufacturing floor footprint dramatically and eliminates the overhead of transferring radar modules between test stations, improving throughput and freeing up space for additional testers.

Fig. 1: Block diagram of the NI Vehicle Radar Test System (VRTS). Because the system integrates an instrument grade Vector Signal Analyzer and a programmable FPGA, the VRTS performs both functional test with obstacle simulation as well as RF parametric measurements.

Beyond the higher frequency and bandwidth requirements of automotive radar testing, the next challenge of testing future radar sensors is the validation of increasingly complex software built into sensors. A radar sensor with 1GHz or more of bandwidth produces massive amounts of raw data. To avoid overwhelming the communication buses and ECU of the vehicle, radar sensors include a processor to reduce this data into a summarized snapshot. Periodically, the radar transmits a parameterized object table with a summary of all the objects currently tracked by the sensor. Each object includes a range, velocity, radar cross-section (RCS), object ID and confidence (a measure of the radar’s confidence that an object exists). The radars software detects these objects and tracks their real-time movements. Algorithms look for inconsistencies such as an obstacle that is moving away from the sensor but has a Doppler signature that indicates the obstacle is approaching.

In the lab, engineers must validate these algorithms and the software that implements them. Lab testing with a compact radar test system allows software developers to quickly validate software changes immediately. Combined with motion systems, such as a small robotic arm to move the radar simulator antennas, the NI VRTS can generate standardized radar environments to characterize and validate radar sensor software, including simulating corner case scenarios that would be difficult or dangerous to emulate with drive testing. Lab testing with simulators is critical to maintaining the pace of innovation of automotive radar sensor design.

Within the context of the entire ADAS or autonomous driving system, engineers must also consider radar emulation for system validation test.  Increasingly, these systems rely on a combination of sensors, including cameras, LIDAR and radar. Validating the overall performance of an ADAS function like Forward Collision Warning and Automatic Emergency Braking increasingly utilizes sensor fusion, the combination of two or more sensors to improve the quality or increase the confidence of an obstacle detection.  For example, if the ADAS radar sensors detect an obstacle but the cameras indicate the path is clear, then the ECU may disregard the radar obstacle as a ghost or interference. 

When testing these functions at the system level, engineers need a test platform that can support a wide set of synchronized sensor simulations to emulate the entire sensed environment around the vehicle. Because the NI VRTS is built on PXI Express, the standard in modular, automated test equipment, engineers can support and emulate additional sensors by adding additional PXI modules such as NI FlexRIO to emulate digital camera inputs in sync with VRTS radar emulation.

Fig. 2: Validating the software for ADAS and autonomous driving requires synchronized emulation of multiple vehicle sensors so the vehicle ‘thinks’ it is driving in the real world.

Next, advanced modulation techniques will have an impact on the future of automotive radar testing. Frequency Modulated Continuous Wave (FMCW) radar has been the standard bearer for automotive radar. Radar designers are now looking to combine multiple antennas in a MIMO setup (multiple input – multiple output) to augment automotive radar capability to accurately detect obstacle elevation or even provide a raster image similar to a camera. Radar sensor researchers are demonstrating higher performance based on modulation schemes that are similar to those commonly used in cellular communication.   

This approach promises to improve radar resolution and field of view while enhancing the radar’s immunity to interference from other vehicles. In response, radar test systems must also grow in sophistication. Accurately emulating a person or vehicle at the resolution of these imaging radars may require demodulating individual radar channels, applying the obstacle effects of distance, Doppler and RCS for each transmit channel, and reflecting that obstacle back to the sensor – all at the roundtrip speed of light.  These new requirements will challenge radar test vendors and suppliers and will require a high bandwidth, low-latency system architecture with extreme signal processing capabilities.

About the author:

Matt Spexarth is Principal Solution Manager for Automotive Radar at National Instruments –

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles