The invention relates to a radar system for the use for driver assistance systems in a motor vehicle. According to the invention, the radar system has arrangements and methods for the decoupling of transmitting and receiving signals and for the suppression of interference radiation.
Motor vehicles are being increasingly equipped with driver assistance systems that cover the surroundings by means of sensor systems and derive automatic vehicle reactions from the traffic situation detected in this manner and/or instruct (in particular warn) the driver, wherein a distinction between comfort functions and safety functions is made.
FSRA (Full Speed Range Adaptive Cruise Control) is the most important comfort function as far as present development is concerned. The vehicle adjusts the ego-velocity to the desired velocity predetermined by the driver if said adjustment is possible in the present traffic situation. Otherwise, the ego-velocity is automatically adjusted to the traffic situation.
Besides increased comfort, safety functions will be more and more important in future, wherein a reduction of the length of the brake path in emergency situations will probably play the most important role. The corresponding driver assistance functions range from prefilling the brake automatically for reducing brake latency via an improved Brake Assist System (BAS+) to autonomous emergency braking.
Nowadays, radar sensors are used in most driver assistance systems of the type described above. Said radar sensors reliably operate even in bad weather and are capable of measuring the distance between objects as well as of directly measuring the relative velocity of the objects by means of the Doppler effect.
These radar sensors are still rather expensive, and the detection quality thereof is not perfect, which is very critical particularly with respect to safety functions. Reasons thereof are, for example:
The object of the invention is to provide a radar system and a method for a motor vehicle that suppresses the effect of interference radiation.
Said object is basically achieved by means of a radar system according to claims 1 to 14.
The suppression of interference radiation particularly comprises the decoupling or isolation of transmitting and receiving signals, which results in a precise determination of the lateral position of objects and in the avoidance of sensitivity losses. The suppression of interference radiation is also comprised.
The advantages of the invention result from reduced demands, particularly on the high-frequency electronic components and on the components of analog and digital signal processing, which reduces the costs of the radar system. Further advantages of the invention result from an improved and robust detection quality.
In the following, the invention will be explained on the basis of exemplary embodiments of radar systems. The invention described in the embodiments and the indicated numerical examples refer to a 24-GHz radar. However, it is not intended to restrict the invention to the 24-GHz range but the invention is claimed for high-frequency radar systems and can be easily realized by a person skilled in the art with other frequencies as well, e.g., with 77 GHz.
The first exemplary embodiment is the radar system that is roughly illustrated in
The following exemplary embodiment is one with four receiving antennas but can be easily realized with any plurality of receiving antennas or with at least one receiving antenna.
All antennas (transmitting and receiving antennas) have the same beam shape with respect to elevation and azimuth. The four receiving antennas are located in the same plane, and each of them has the same lateral, i.e., horizontal spacing d.
The transmitting signals are obtained from the high-frequency oscillator 1.3 in the 24-GHz range. The frequency of the oscillator 1.3 can be changed by means of a driving voltage vSteuer. The driving voltage is generated in the driving means 1.9. The signals received by the antennas are mixed down to the low-frequency range in the real-valued mixers 1.5 also with the signal of the oscillator 1.3. Moreover, the phase of the oscillator signal may be rotated by 180° by means of the switchable inverter 1.4, or it may be left unchanged (the switchable inverter is driven by the driving means 1.9). After that, each of the receiving signals passes a band-pass filter 1.6 with the transfer function shown, an amplifier 1.7 and an A/D converter 1.8, followed by further processing thereof in a digital signal processing unit 1.10.
In order to be able to measure the distance from objects, the frequency of the high-frequency oscillator (and thus of the transmitting signals) is linearly changed very quickly (so-called frequency ramp), e.g., by 187.5 MHz within 16 μs (see
The receiving signal of an individual object is, after mixing and thus also at the ND converter, a sinusoidal oscillation for each frequency ramp and for each of the four receiving channels, which can be explained by means of
The receiving signals at the A/D converter are sampled, e.g., 512 times at an interval of, e.g., 25 ns each (i.e., with 40 MHz) during each frequency ramp in all four receiving channels (see
After that, a discrete Fourier transform (DFT) in the form of a fast Fourier transform (FFT) is formed over the, e.g., 512 sampled values of each frequency ramp and of each receiving channel, whereby objects in different distances resulting in different frequencies can be separated (see
Each of the discrete frequency control points j of the DFT corresponds to a distance r. Therefore, it can be also called “range gate” (by analogy with pulse radar). In the design mentioned above, the range gates have an interspace and thus a width of just one meter (results from r·78.125 kHz/m=1/(12.8 μs)). In the DFT, power peaks occur in the range gates in which objects are present. Since the sampled receiving signals are real-valued and the upper transition region of the analog band-pass filters 1.5 has a frequency bandwidth of, e.g., 8.764 MHz (corresponds to the range of 112 frequency control points), only 200 of the 512 discrete frequency control points can be processed in this numerical example. Note that filter transition regions cannot have any desired narrowness. The filters 1.5 damp small frequencies and thus the receiving signals from close objects in order to avoid the overdriving of the amplifiers 1.6 and the ND converters 1.7 (the intensity of the signals received at the antennas increases with decreasing object distance).
Complex spectral values e(j,k,m) occur over the, e.g., 1024 frequency ramps (k=0, 1, . . . , 1023) in each receiving channel m (m=0, 1, 2, 3) for each range gate j (i.e., each of the, e.g., 200 considered frequency control points). If there is exactly one object at the distance that corresponds to a range gate, the complex spectral value in this range gate j rotates over the, e.g., 1024 frequency ramps at the Doppler frequency since the distance (in the mm range or below that) and thus the phase position of the associated oscillation uniformly changes from frequency ramp to frequency ramp. The phase change of 45° per frequency ramp (see example in
The result of this second DFT for the relative velocities is a two-dimensional complex-valued spectrum for each receiving channel, wherein the individual cells can be called “range—relative-velocity gates” and power peaks caused by objects occur at the respective associated range—relative-velocity gate (see
Finally, the information from the four receiving channels (to the four receiving antennas) is merged. The wave radiated by the transmitting antenna and reflected from an individual object arrives at the four receiving antennas m, m=0, 1, 2, 3, with different phase positions φ(m) in dependence on the azimuth angle α since the distances between the object and the receiving antennas are slightly different. Because of the horizontal equidistance of the receiving antennas, the phase differences linearly increase/decrease over the four receiving antennas (see
The result of this third DFT for the azimuth angles is a three-dimensional complex-valued spectrum, wherein the individual cells can be called “range—relative-velocity—angle gates” and power peaks caused by objects occur at the respective associated range—relative-velocity—angle gate (see
In real radar systems, interference coupling or interference radiation occurs in the radar frequency range (e.g., 24 GHz) or in the range in which the low-frequency part of the electronic evaluation unit operates or is sensitive (e.g., approximately in the range of 50 Hz to 1 GHz). These interferences can be caused by other systems or by the radar system itself. Examples for such interferences are:
If no special measures are taken, all these interferences may result in the supposed detection of objects that do not exist in reality (ghost objects), which may result in wrong reactions of driver assistance functions. For example, if the 500-kHz cycle of a voltage regulator equally couples into all four receiving channels, the result will be a power peak in the three-dimensional spectrum (after the third DFT), said power peak resulting in the detection of an object at a distance of a bit more than 6 m, at an azimuth angle of 0° and at a relative velocity of 0 km/h. If the FSRA function (Full Speed Range Radar) is implemented with the radar system, this supposed detection consists in the erroneous and permanent detection of a vehicle driving ahead of the ego-vehicle at a very short distance from the ego-vehicle and at the same velocity as the ego-vehicle. The ego-vehicle slows down relative to the vehicle driving ahead in order to achieve a sufficiently big distance therefrom. However, since the distance from and the relative velocity of this ghost object never change (it slows down more or less at the same rate as the ego-vehicle), the ego-vehicle is almost slowed down to a standstill, which is of course unacceptable and may also become critical with respect to safety.
For avoiding the problem described above, the phase of the oscillator signal used for mixing is rotated at random from ramp to ramp by 180° by means of the switchable inverter 1.4, or it is left unchanged. The selected setting of the switchable inverter remains constant within each ramp. This results in the phases of the receiving signals equally varying after mixing, i.e., they are rotated by 180° or they are not. For frequency ramps where inversion has taken place, this must be corrected later (e.g., after the first DFT) by just multiplying the respective values by −1 (corresponds to a reversed rotation by 180°). After that, the useful signals resulting from reflections from objects are coherently reintegrated in the three discrete Fourier transforms. The result is the same three-dimensional spectrum as without random inversion with power peaks at the corresponding range—relative-velocity—angle gates.
The coupling into the low-frequency receiving channels (caused by, e.g., a 500-kHz cycle of a voltage regulator) is coherent prior to the correction of the phase variations over the ramps but becomes non-coherent after the correction by the ramp-to-ramp random multiplication by −1 or +1 so that said coupling does not result in a power peak any more because of the integration over the ramps taking place in the second DFT and the third DFT, but its power is distributed at random among all discrete frequency control points and thus represents white noise. Said noise occurs in the three-dimensional spectrum in all cells of the range gate for 6 m and, at a reduced level, in the one or two preceding and one or two subsequent range gates. There is no increased noise in the cells of other range gates since the coupling is coherent within each ramp and is therefore not transformed into noise by the first DFT, yet. In the exemplary design described above (a total of 1024 ramps), the noise caused by the coupling is approximately 10·log10(1024)≈30 dB below the power peaks that the coupling would generate without the phase variation. This is shown for an intensive 500-kHz coupling in
The same principle applies to the other interference coupling or interference radiation mentioned above. Because of the random inversion, said coupling or radiation only results in possibly increased noise in few range gates (if the noise generated by it is above system noise) but does not result in ghost objects.
So far, the ideal case (power is only radiated by transmitting antenna 1.1) has been discussed (see embodiment 1 according to
For 24-GHz narrow-band radar systems operating in the so-called ISM band, the amount of power outputted to the transmitting antennas by the oscillator at least equals the amount of power outputted to the mixers. Since the mixers typically have an isolation of at least 20 dB, the power radiated by the receiving antennas is negligible as against the actual transmitting power (radiated by the transmitting antennas). For 24-GHz wide-band radar systems (so-called UWB radar systems), frequency allocation is very restrictive. It only allows the radiation of a very low transmitting power, which results in the fact that the amount of power unintentionally radiated by the actual receiving antennas on account of insufficient isolation comes close to the amount of power radiated by the transmitting antennas. This may result, in the arrangement according to
The following exemplary embodiment can be easily realized for a radar system with a plurality of transmitting antennas and at least one receiving antenna and will be presented on the basis of an embodiment with one receiving antenna and four transmitting antennas.
Therefore, the radar system shown in
Furthermore, a disadvantage of the radar system according to
For avoiding both problems described above (declines occurring in the transmitting-antenna diagrams and incorrect azimuth angle formation), the phase of the oscillator signal used for mixing can be rotated at random from ramp to ramp by 180° by means of the switchable inverter 10.4, or it can be left unchanged. The selected setting of the inverter remains constant within each ramp. Over the ramps, this results in the power radiated by the receiving antenna becoming uncorrelated and thus non-coherent to the power radiated by the transmitting antennas. Again, the power radiated by the receiving antenna and reflected from objects results only in low noise in the corresponding range gates in the receiving signals. Said noise is approximately 10·log10(1024)≈30 dB below the power that would result without the phase variation (i.e., in the case of a coherent integration by means of the second DFT and the third DFT over 1024 ramps).
The following exemplary embodiment can be easily realized for a radar system with a plurality of transmitting antennas and at least one receiving antenna and will be presented on the basis of an embodiment with one receiving antenna and two transmitting antennas.
The simpler radar system shown in
The switchable inverter 11.4 alternately varies the phase of the signal of the first transmitting antenna from ramp to ramp by 0° and 180°, i.e., the signal is inverted in every second ramp and remains unchanged therebetween. The phase of the signal of the second transmitting antenna is not varied. The alternating phase of the signal of the first transmitting antenna results in the receiving signals from this transmitting antenna being modulated at half the ramp recurrence frequency (25 kHz) over the ramps so that their Doppler frequency is also shifted by 25 kHz after the second DFT. The receiving signals from the second transmitting antenna are not shifted in the Doppler. For an object whose relative velocity corresponds to, e.g., a Doppler frequency of 5 kHz, the second DFT results in a power peak at 5 kHz for the receiving signals from the second transmitting antenna and a power peak at 30 kHz for the receiving signals from the first transmitting antenna. Thus, the components from the first transmitting antenna and from the second transmitting antenna can be separated on the basis of their frequencies after the second DFT. The component of the first transmitting antenna can be shifted back by 25 kHz, whereafter the third DFT (having a length of 2, for example) for angle formation can be performed.
Instead of the determinate alternating phase variation, there could be a random phase variation. In that case, however, the second DFT would have to be determined two times—one time with a correction of the phase variation and one time without said correction. In the DFT calculated with phase correction, the receiving signals from the first transmitting antenna would result in power peaks, whereas the receiving signals from the second transmitting antenna would generate noise that is approximately 30 dB below the power peaks. In the DFT calculated without phase correction, conditions would be vice versa. Because of that, the two components could be separated.
The following exemplary embodiment can be easily realized for a radar system with at least one transmitting antenna and a plurality of receiving antennas and will be presented on the basis of an embodiment with one transmitting antenna and two receiving antennas.
Finally, the simple radar system shown in
Number | Date | Country | Kind |
---|---|---|---|
10 2009 016 478.2 | Apr 2009 | DE | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/DE10/00417 | 4/1/2010 | WO | 00 | 9/14/2011 |