Considerations for Future Automotive Radar in the Frequency Range Above 100 GHz Mike K¨ohler #
#1 ,
Frank Gumbmann
#,
Jan Sch¨ur
#,
Lorenz-Peter Schmidt
#,
Hans-Ludwig Bl¨ocher
∗
Chair for Microwave Engineering and High Frequency Technology (LHFT), University of Erlangen-Nuremberg, Cauerstr. 9, D-91058 Erlangen, Germany #1 Email:
[email protected] ∗
Daimler AG, Group Research & Advanced Engineering, D-89013 Ulm, Germany
Abstract— Within the last years automotive radar sensors became more and more important. Started with comfort systems like adaptive cruise control (ACC) and parking aid, the safety aspect increasingly came to the fore. In the near future not only upper class cars will be equipped with various radar applications like pre crash, collision warning and collision avoidance to increase traffic safety. Therefore many high performance radar sensors have to be integrated in the car. This paper describes the demands on future radar sensors. It is discussed how the step to higher operating frequencies could be beneficial particularly for urban situations with high traffic density. Furthermore the paper will provide and discuss design considerations for future mmW-Radar sensors in automotive safety applications.
I. I NTRODUCTION Traffic statistics show that inattention of the car driver is the most common cause for accidents. Thus, in addition to the passive safety systems (seat belt, airbag) and active safety systems (ESP, ABS) the driver shall be supported by intelligent radar systems monitoring the environment. One of the first forward looking radar sensors for adaptive cruise control (ACC) was implemented in Mercedes S-Class. This long range radar (LRR) working at 77 GHz scans a sector in front of the car with an angle of view of 30 degrees up to a distance of 150 meters. To get more information about the sideways traffic lanes a 24 GHz UWB Sensor (SRR-short range radar) was realised, monitoring the short range area within a maximum distance of 30 meters and an angle of view of 80 degrees. The frequency regulation for UWB SRR was redefined in an European 2-Phase-Plan [1]. Instead of the 24 GHz a new operating frequency at 79 GHz was allocated. Since 2005 the development of radar sensors working in the frequency range of 77 GHz - 81 GHz was forced in projects like KOKON. Until mid-2013 the changeover to 79 GHz should be completed. After that sunset date 24 GHz systems can still be used but further selling has to be stopped. For realising a radar system in this high frequency range proper technologies must be available. To get reliable components (mixers, amplifiers) the transit frequency and the steadily maximum oscillation frequency of the devices should be 5 to 10 times higher than the operating frequency. Various companies (Infineon, IBM, Jazz) and institutes (FBH, IFH,
978-3-9812-6681-8
(c) IMA
284
Fraunhofer IAF) are working on increasing the maximum frequency of semiconductors towards the THz range. Projects like Dot Five were set up to push the limit of the cost-efficient SiGe technology to 500 GHz and above. Research covering this topic was continued in KOKON and is still in progress in the BMBF-funded RoCC project (Radar on Chip for Cars). Another topic started in KOKON and continued in RoCC is the monitoring of the complete traffic environment. Therefore the car has to be equipped with surrounding sensors realising an angle of view of 360 degrees. Figure 1 depicts different applications and positions of some current integrated radar sensors. In most applications it is necessary, that the radar
Parking aid
Blind spot detection
Pre crash
Backup parking aid Rear crash collision warning
Stop & Go for ACC Collision warning
Collision mitigation Fig. 1.
Blind spot detection
Lane change assistent
Automotive radar applications
system has a sufficiently high resolution in the direction of propagation. It must be able to distinguish a person from other objects like containers or street lamps. Furthermore it has to be sensitive enough to detect a weakly reflecting obstacle in presence of a strong second reflector, for example a person standing in front of a wall. This would be beneficial for parking maneuvers in areas with many buildings and pedestrians. Respective activities were started in Japan to lower the accident risk at traffic situations in the so-called mega cities. These and many other applications demand high resolution radar systems in the short and mid range area. It is a wellknown fact that high resolutions are better achievable in a
German Microwave Conference 2010
higher frequency range. The step from 24 GHz to 79 GHz systems is still in progress and should be completed by 2013. So there is a reasonable question: ”What can we expect from further increasing the frequency, perhaps to 150 GHz or even higher?”
77 GHz
150 GHz
2.4°
1.2°
II. M OTIVATION Fig. 3.
A. Miniaturisation As can be seen in figure 1, there are various applications, that will be implemented in cars. Without doubt in the near future the number of safety functions and therefore the number of sensors will grow. Thus, current radar sensors with dimensions of about 10 cm x 8 cm x 6 cm could be too large to place them all in the carbody and it could be necessary to further miniaturize them. Due to the step to a higher operating frequency the place required by the antenna would become significantly smaller. A demonstrative comparison is given in figure 2. Because
150 GHz
77 GHz
2 mm Fig. 2.
Comparison of angular resolution
C. Range Resolution The range resolution in radar systems depends on the length of the transmit pulse or the bandwidth in FMCW-radars, respectively. In current 79 GHz sensors a bandwith of 4 GHz is attainable and a range resolution of 3.8 cm can be achieved. Doubling the operation frequency with the same relative system bandwidth results in an absolute bandwidth of 8 GHz. This leads to an even better range resolution of less than 1.9 cm. Thus, in longitudinal direction the distance measurement accuracy increases and multiple objects may be easier distinguished. If the range resolution of 3.8 cm is sufficient there are alternative possibilities to use the larger bandwidth. So in areas with high traffic density the complete bandwidth could be used for frequency hopping to mitigate interference between automotive radars of several cars. III. B OUNDARY C ONDITIONS
1 mm Comparison of patch size
a single patch element has dimensions in the range of λ/2, the four patches of the 77 GHz array require approximately the same space like an 16-patch array at 150 GHz. So with a 150 GHz radar system a better performance due to the increasing number of elements with the same antenna area could be achieved. Alternatively by spending the same number of antenna elements a similar performance compared to 77 GHz can be attained. In that case the 77 GHz antenna area would be four times the area of a 150 GHz antenna.
Thinking about changing the frequency range of an established radar system means consideration of many new aspects, e.g. the wave propagation under different conditions (influence of weather) and the availability of suitable technologies and materials. A. Characterisation of Materials Various standard automotive radome materials have been tested in different frequency bands from 75 GHz to 325 GHz. This will result in detailled information about the materials performances at higher frequencies, because a radome working well at 77 GHz does not mean necessarily a good suitability for 150 GHz. The non-destructive testing was made with an open
B. Angular Resolution In section I the call for high resolution radar systems has been discussed. The resolution is classified in angular and range resolution. It is obvious that the development of a radar sensor with high lateral or angular resolution calls for a sufficiently large aperture. As exemplarily illustrated in figure 3 a 77 GHz radar with an aperture width of 5 cm has a lateral resolution of 4.2 meters in a distance of 50 meters. This corresponds to an angular resolution of 2.4 degrees. Compared to a 150 GHz radar with the same aperture width and in the same distance a lateral resolution of 2.1 meters can be attained, which equals an angular resolution of 1.2 degrees. This results in a better single aim detection only by doubling the operating frequency.
285
Converter
Horn antennas
Absorber
DUT
Electronic moveable sample holder Fig. 4.
Measurement setup with converters
Method of measurement
Polypropylen - series of measurement
Average value
tan δ
1
2
3
4
5
Transmission Wband
2,3667
2,3760
2,3749
2,3680
2,3708
2,3713
