At EUMW 2016, companies look to advance ADAS with higher-frequency solutions

BY RICHARD COMERFORD, Senior Technical Editor, Electronic Products

If European Microwave Week 2016 (EUMW 2016, held in London the first week of October) was any indication, in the next decade, the expression “driving the car” may become as archaic as “dialing the phone.” The industry is in the process of developing new sensor technology for automobiles that will eventually allow a car to take over all the navigation and control functions needed to get from one place to another. Referred to as Advanced Driver Assistance Systems (ADAS), these systems will initially be introduced to help motorists drive more safely. The functions that ADAS can provide, such as automatic braking, collision avoidance, and lane-keeping assistance, are expected to serve as the foundation for future self-driving vehicles.

The sensors that designers are now testing for such applications fall into four broad categories: optical (cameras), sonar, radar, and lidar. The last of these, lidar, is similar to radar, but instead of bouncing radio waves off of an object to detect it, lidar uses coherent light from a laser. Hence, “lidar” — “light radar”, or “light detection and ranging.”

There has recently been a lot of activity as companies jockey to become the next great automotive technology suppliers. For instance, at EUMW 2016, Fujitsu Laboratories Ltd. announced development of millimeter-wave radar technology aimed at helping realize autonomous driving amongst objects — cars, pedestrians, bicycles, and such — moving at different speeds; detection at relative speeds up to 200 km/h is possible (Fig. 1 ).  

Fig. 1: Fujitsu’s millimeter wave radar would allow detection of both slow and fast moving objects in a vehicle’s vicinity.  

Fujitsu expects millimeter-wave radar to serve as the “eyes” of ADAS, as it can make up for the weaknesses of optical cameras in adverse environments, such as at night, during rain, fog, and backlighting. More than just the conventional use of the narrow 77-GHz band for monitoring in front of and behind vehicles, in recent years, there has been increasing interest in peripheral monitoring radar that uses a broader band of 79 GHz.

The new technique avoids a major disadvantage of frequency modulated, continuous wave (FMCW) millimeter-wave automotive radar: as FMCW radar approaches objects moving at different speeds — such as a vehicle and a pedestrian — it tends to overlook one of them. Fujitsu’s fast-chirp modulation (FCM) overcomes the problem by using higher-speed modulation to enable detection with better distance resolution and a broader range of target-object speeds.

To implement such a system, Fujitsu’s researchers created a CMOS-based millimeter-wave signal generator capable of modulating its frequencies across a 76- to 81-GHz band (Fig. 2 ). Combining this circuit with a concurrently developed four-channel CMOS transmitter circuit to measure and control millimeter-wave beams with a phase precision within one degree, the user can scan their surroundings with a high level of precision, for example, with resolution at a 5-cm interval anywhere within a 10-m radius. The signal-generator circuit, which controls the frequency of the millimeter-wave signal, continuously reads in and counts the millimeter-wave signal pulses, applies a voltage to the frequency controller based on that count, and modulates the frequency.  

Fig. 2: Fujitsu’s millimeter-wave CMOS signal-generator circuit (a) and four-channel CMOS transmitter circuit (b) promise higher-precision ADAS radar.  

Circuits used in automotive radar are expected to perform normally at ambient temperatures as hot as 150°C, but with conventional CMOS signal generators, the internal signals slow down and the counts become inaccurate. Without being able to increase their modulation speed, the relative speed where detection is possible is limited to about 50 km/h.

The new technology makes use of existing millimeter-wave CMOS design technologies and focuses on the block in the signal generator that has the most effect on the counting operation. By adding a function to the block that compensates for delays caused by temperature changes, Fujitsu Laboratories was able to develop a new time-compensating pulse counter that operates accurately, even at temperatures of 150°C (Fig. 3 ). This circuit enables the world's fastest modulation frequency of 2 GHz for every 1 µs at the 80-GHz band and achieves the maximum relative speed detection (200 km/h) that is expected of radar.  

Fig. 3: Fujitsu adds circuitry to its radar circuitry to correct pulse-readable timing error due to rising temperature.

Fujitsu Laboratories says it is working on developing a radar chip that integrates a high-performance processor and other elements and further advancing high-end functionality, with the goal of making these technologies practical from 2020.

Also at EUMW 2016, National Instruments gave a technology demonstration of a new ADAS Test Solution for radar in the 76−81-GHz range based on NI’s mm-wave front-end technology and the recently released PXIe-5840 second-generation vector signal transceiver (VST). Stefano Concezzi, vice president of the global automotive initiative at NI, noted that, “With regulatory requirements still evolving, the flexibility of this solution allows engineers to quickly adapt their test systems to address the challenges of new radar scenarios.”

Because it combines the VST with banded, frequency-specific up-converters and down-converters designed to test the 76−81-GHz radar band with 1 GHz of real-time bandwidth, the system can function as a mm-wave vector signal generator and vector signal analyzer. Engineers can program the VST’s FPGA with LabVIEW to use the ADAS Test Solution for radar target emulation, a technique in which test equipment emulates the radar cross-section, range, radial velocity, and angle of arrival of a particular object. This is essential for testing a radar system’s software and hardware.

Right after EUMW 2016, Infineon Technologies announced its acquisition of the Dutch fabless semiconductor company Innoluce to develop chip components for high-performance lidar systems. The company believes lidar, radar, and camera will be the three key sensor technologies for semi-automated and fully automated cars, so with the acquisition of Innoluce, it plans to deliver expertise in all three complementary sensor systems, which provide the redundancy required for autonomous driving.

Infineon says that the first lidar systems introduced in premium cars within the next couple of years will be based on mechanical scanning mirrors and, thus, be bulky and rather expensive. To become a standard feature in all car classes, lidar systems need to be semiconductor-based, thus getting more compact, cost-effective, and robust.

Over the horizon

Since standard sensors such as radar, optical, ultrasonic, and lidar are all line-of-sight, they can only detect risks within view of the vehicle. The Australian firm, Cohda Wireless is working on a system that can detect hidden-from-view threats, so it can extend the horizon of awareness beyond what the driver can see, such as when two cars are approaching each other around blind corners, over the crests of hills, or when there are trucks between them.

Cohda recently introduced vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) — or V2X — equipment that it sees as an essential technology for the next generation of Advanced Driver Assistance Systems (ADAS) as well as autonomous vehicles. V2X is a wireless sensor system that allows vehicles to share their sensor data with other vehicles and highway equipment around them. Cohda’s V2X technology is a non-line-of-sight sensor with 360° awareness that uses accurate satellite positioning with embedded dead reckoning technology provided by u-blox. Based on V2X, Cohda’s Dedicated Short‑Range Communications (DSRC) system enables, for example, early warning of an imminent collision, oncoming traffic, presence of road workers, and unsafe speed based on vehicles in the vicinity.