FI110966B - A method for processing a Frequency Modulated Interrupt Continuous Wave Radar (FMICW) signal - Google Patents

Description

110966

Method for Processing Frequency Modulated Interrupt Continuous Wave Radar (FMICW) signal

The invention relates to a method according to the preamble of claim 1 for processing a Frequency Modulated Interrupted Continuous Wave (FMICW) signal.

The invention relates to radar systems using frequency modulated interrupt continuous wave signal modulation (FMICW). This modulation method is also sometimes referred to as the roof home FMCW or IFMCW

The invention is applicable to situations in which the FMICW radar signal is used to search for objects, either discrete or continuous, over a continuous range or to track multiple targets at different distances.

15 The FMICW radar is a pulsed frequency sweep (see, e.g., "Principles and Applications of the Millimeter-Wave Radar", by Currie and Brown, ed., Page 683). In such a system, the interrupted linear frequency sub-bandwidth is transmitted, typically at 50% transmission time.

This type of radar method typically comprises the following steps. A signal is generated whose frequency changes with the set bandwidth. The signal is spun on and off at a predetermined frequency (fg), thereby forming on-time and off-time slots. This gated signal is transmitted to the desired destination by means of a transmitter and antenna-25 n. The reflected or scattered signal is received during the off-time slots of the gated signal, and the desired characteristics of the received signal are detected and collected during the off-time slots of the gated signal.

«

When the FMICW is applied to a windbreaker, typical operating parameters could be 30 such that the area resolution is 150 meters and the maximum detection range is 3 km. In such a case, a 1 MHz linear frequency sweep would be transmitted and the sweep 2110966 would be interrupted by 20 ps pulse sequences. As a result, the 3 km range would be optimal in terms of pulse frequency; the transmitted 20 ps pulses would arrive at the receiver 20 ps later, with a time delay of 3 km back and forth. As a result, the scattered power of the entire 20 ps pulse would arrive at the receiver. However, for other heights the situation is less favorable. The time-frequency output for 20 ps pulse, 20 ps reception timing with 1 MHz linear frequency sweep is shown in Figure 1.

Since the return signal is mixed with the continuous frequency sweep of the transmitted signal in the FMICW receiver, the different difference frequencies correspond to different ranges. Figure 2 shows 10 time-frequency output after mixing the received signal and the transmitted signal. The mixing can take place in one step (homodyne) or in several steps via intermediate frequencies (heterodyne).

Figures 1 and 2 show three exemplary radar return beacons: one fitted at a pulse modulation frequency of 3,000 meters, one return echo closer than a matching distance, and one farther away from the radar than a pulse modulation matching frequency. These three targets can be considered as part of geophysical scattering targets or as multiple point targets on a search or trace radar.

Figure 1 shows the transmitted signal and return hooks at three altitudes: 1200, 3000 and 4200 meters. The return echo of 1200 meters is delayed by 8 ps for transmission and the return echo of 3000 meters is delayed by 20 ps for transmission. The return echo from 4200 meters is delayed by 28 ps.

In the FMICW ice system, the signal is only allowed to enter the receiver when there is no transmission. The baseband signal is generated by mixing the received signal with a non-transmission portion of the frequency sweep. As a result, the system has a baseband return echo from three heights: 1200 meters, 3000 meters and 4200 meters. The return gates at these heights are shown in Figure 2.

30 3 110966.

The frequency sweep shown in Figure 1 is continuous in phase throughout a 5 ms period. Thus, the 1.6 kHz (1200 m), 4.0 kHz (3000 m) and 5.6 kHz (4200 m) signals shown in Fig. 2 also show a continuous phase through the entire frequency sweep, since the 5 ms scan is much shorter than the de-5 correlation time of the scatterers in the volume under investigation.

Because the return echoes have a phase continuity, each portion of each frequency of Figure 2 forms a blue pattern that is temporally interrupted. The intermittent 1.6 kHz signal has time to form 8 intermittent sine wave periods during a 5 ms scan, a 4 kHz signal for 20 10 cycles, and a 5.6 kHz signal for 28 cycles. These sine waves are summed in the receiver to produce a voltage that contains different intermittent frequency components associated with different heights.

Taking digital samples while the receiver is on, generates a 125-point timeout (if one sample is taken during one RX interval). The different frequencies in this time series are separated by a Fourier transform. For example, the 8th, 20th, and 28th samples in this Fourier Transform correspond to heights of 1,200, 3,000, and 4,200 meters. If the receiver is, as in the preferred case, capable of separating the so-called phase and cross-phase components at the baseband frequency, the corresponding Fourier transform is complex and the frequencies obtained in the given example are negative.

The object of the present invention is to overcome the drawbacks of the above-described technique and to provide an entirely new type of method for processing frequency modulated interrupted wave (FMICW) radar signal.

25

The object of the invention is achieved by detecting and / or processing only a desired portion of the off time jet during each oif time slot of the gated signal to obtain information only at a desired distance and to minimize signal noise.

The method according to the invention is particularly characterized by what is stated in the characterizing part of claim 1.

4, 110966

The invention offers considerable advantages over the prior art.

The invention improves the detection range of an FMICW radar outside a pulse cycle compatible range. Thus, it improves the usability of FMICW technology for detecting continuous profiles of volume rotors such as weather radar, wind profiler and other geophysical radar. It also improves the usability of FMICW in search radar and reduces the need for pulse rate scanning, since a greater range can be detected with a single pulse frequency implementation.

10

The invention will now be described in detail with reference to exemplary embodiments and the accompanying figures, in which: Figure 1 shows a frequency-time curve of the principle of FMICW modulation, Figure 2 shows another frequency-time curve of the principle of FMICW modulation, Fig. 4 is a multi-curve representation of the signals obtained by the invention, Fig. 5 is another multi-curve representation of the signals obtained by the invention.

This publication describes a novel method related to frequency modulated truncated continuous wave modulation (FMICW). The method improves radar return echo from distances that are not matched to the pulse frequency of the ice system and thus improves the ability of the 'FMICW' systems to detect observations over continuous ranges. FMICW ice systems are traditionally used in search in the form of pulse rate scanning.

As shown in Figure 2, the FMICW system signal at different heights consists of intermittent frequency components. The interruption is determined by the pulse parameters and the height in question. If one sample is taken per pulse, integrated over the reception period, the sample contains noise and signals from different heights. Now let's look at one of the reception batteries, the first one a little more closely.

5 Now, one sample per pulse, after integration over a reception period, each frequency component of the signal includes a signal portion and a noise portion. The signal portion consists of the integrated voltage over the duration of each curve shown in FIG. However, the noise portion of this frequency compartment is equivalent to integration over a period of 20 µls. Although the signal portion remains coherent over multiple pulses, the noise portion 10 does not, and the noise thus has a detrimental effect on the signal in such a way that it interferes with the phase and voltage of the actual desired signal.

Let us now consider one continuous 5 ms sweep and the corresponding output from the FMICW receiver. Let there now be an object, either discrete or volume scatterer, at a distance of 1200 m, which corresponds to a 1.6 kHz signal in Figures 2 and 3. If we have a receiver with a bandwidth of 1 MHz, a matched bandwidth value for the transmitted waveform, complete representation of the signal by 1 MHz quadrature sampling at baseband. We do not consider the effect of the finite steepness of the transition bands on the receiver's concatenated filters, or the folding there. In practice, taking these phenomena into account, a sampling bandwidth greater than 1 MHz could be chosen. However, it does not fundamentally change the situation - in this case, · · the same part of the time would be digitally integrated as in these examples. At that time, however, there would be more samples to be integrated per case.

In Figure 4, the top result line shows the time-domain signal (simulated) at 1200 meters, plus noise. The right-hand output shows the result of using a 5000 'high-speed Fourier transform to extract the various frequency components - and at the same time the range - from the signal. By this method, the signal-to-noise ratio of the target is 15.9 dB, measured over a bandwidth greater than 200 Hz. It occurs at 1.6 KHz baseband 30 at the 8th frequency compartment, which corresponds to the 8th elevation compartment, 1,200 meters. The signal consists of 20 zeros, which corresponds to the transmission time during which the receiver is muted, followed by 20 samples of signal and noise, followed by 20 zeros, etc. (see Figure 5). The time unit in the left column of Fig. 4 is 1 ps, and the Frequency unit in the right column is 1 MHz. In this specification, it is assumed that the mute is achieved in the digital components, which allows the transmission time signal to be completely reset. It should be noted that a bandwidth of 1 MHz is greater than a pulse repetition frequency of 25 kHz, which causes replicas of the signal within ± 25 kHz of center frequencies.

In Figure 4, the middle output row illustrates the effects of using over 20 ps cycle, full reception, integration, and 125-point fast Fou-10 rier conversion to reveal components of different signal areas. The length of the Fourier transform was reduced without degradation in the signal-to-noise ratio, which is now 15.7 dB. The pulse effect is no longer visible due to the reduced Nyquist frequency at the base frequency. The integration is accomplished by summing the values from each receiving period, resulting in higher voltage amplitudes. It is noteworthy that here the 15 signals to be examined, at 1200 meters, correspond to a rotating pointer which rotates less than 5 degrees during one receiving period. Thus, it is possible to sum up all samples from the receiving period without significantly reducing the signal voltage due to the addition of rotating indicators.

The top two rows of Figures 4 and 5 can be considered as being included in the prior art. The last line of Figure 4 shows the effects of the integration according to the present invention. In this case, a discrete Fourier transform, which can also be implemented using only one of the output components of the fast Fourier transform, is used to produce a frequency component corresponding to a 1200 meter range. The difference is that instead of using a full receive time integrated signal, we now use an integrated signal that adapts the integration time to the bidirectional delay of transmission. ' · And the echo reception of said subject. In this case, only the 8 first samples are integrated during each reception period to produce the last row left-hand time series. The integration of eight samples was chosen for this Fourier Transform (DFT) because it aligns the start of the reception period with the signal scattered from the target. This can be seen in Figure 3: in 7,110,966, only the first 8 microseconds of the receiving period contain a scattered signal at a distance of 1200 meters. The amount of noise in this time series is smaller than the noise in previous time series because the last 12 samples have been omitted from integration, and the noise in these samples does not contribute to the output of the 8-point integrator.

The signal itself remains unchanged since all information from this area is used as shown in Figure 3. In this example, the signal-to-noise ratio due to the integration of the partial reception time is 19.5 dB, 3.8 dB higher than the signal-to-noise ratio reception period integration.

In Figure 5, the data in the second row is obtained by summing all 20 points of the signal in the highest output per reception period. The last line of data is obtained by summing 8 points per reception period, resulting in less noise. A similar procedure is used for all heights separately.

The essence of the invention described herein is that a plurality of digital integrators are constructed, preferably by continuously summing and using different integration times for each region. For example, an area of 150 meters could thus be detected using single point integration, i.e., a first sample of each receive period in combination with a DFT that extracts the first non-zero frequency from a 125-point time series comprising said first samples. The 300-meter range would be expressed by using a 2-point integration, i.e., the first two samples summed, of each receiving period, and exporting the resulting 125-point time series to a DFT that computes another non-zero frequency component from the 125-point time series. This procedure is continued up to pulse-matched altitude, in this case 3000 meters, for which the full 20 25-point integration result is used and for which the DFT determines the 20 non-zero frequency component.

t

For altitudes above 3000 meters, only the remainder of the reception period is used to integrate to obtain all signal information and to minimize the amount of noise corresponding to the frequency component. For example, for the 4200 meter signal shown in Figure 2-3, 8 110966 are used for the last 12 samples of each 1 MHz sampled 20 (is durable) reception period so that the DFT displays a 28th frequency component.

One preferred embodiment of the present invention would be to use a DSP or FPGA-5 rigid system to compute area-specific signals using a cumulative sum register to determine the waveform transmitted at a sufficiently high frequency and multiplying the contents of said register with complex weights corresponding to DFT multiplication for that pulse and region. Multiplication would result in the input of another complex summary register for each area, and the contents of these registers would be read at the end of the fitness sweep. These values are the complex voltages corresponding to each height for that sweep. The DFT registers are read and reset at the end of each sweep, and the integrator registers are reset at the beginning of each receive period.

It is within the scope of the invention to use DFT instead of a digital summation of all samples during a single pulse reception period for a suitable subset of the original 1 MHz samples. This is useful when objects are detected in areas that are within the sweep repetition period. In the exemplary case presented in this document, for example, objects hundreds of kilometers away would be suppressed if a sum integrator was used. This is because the complex voltages of the rotating pointer are summed over the time at which the pointer rotates substantially. Thus, for objects that are very far from the radar, a preferred method is to use matched subsets from the original samples as a long input to the DFT. In these cases, the summing of the digital samples would result in some loss of signal amplitude due essentially to grounding of the rotating pointer sum-25.

*; In the example corresponding to Figure 4, the amount of detection enhancement, 3.8 dB, is due to the fact that the target to be detected remains somewhat coherent for a 5 ms scan period, while the noise is completely incoherent, i.e. has no phase correlation between samples. The noise power is thus additive during the reception period, and the preferred integration method receives 8/20 of the noise power instead of the full integration of 20/20 9 110966. The factor 8/20 corresponds to 4.0 decibels, which is somewhat consistent with the simulated result.

To better illustrate the partial integration during pulse reception, FIG. 5 is an enlarged view of the prints of FIG. 4.

A typical device according to the invention comprises the following elements: (i) a wide bandwidth sampling system (coupled to an analog-to-digital converter), (ii) digital logic generating more than one sum of samples per FMICW pulse reception period, (iii) a counting system using said more than one time series as input to the DFT or one or more frequency components of the FFT to produce the same number of ranges per corresponding time series.

Typically, the length of each of the samples and the frequency component detected in the DFT or FFT are matched to the same radar detection range.

In this application, "broadband" means a sampling bandwidth sufficient to determine the radar transmission bandwidth.

25

One alternative integration system according to the invention may comprise the following elements: i) a broadband sampling system, 30a 10 110966 ii) a digital logic that produces samples of more than one subset per FMICW pulse reception period, iii) subsets of time slots as input to the DFT or FFT to produce one or more frequency components for the same number of ranges per respective time series.

Typically, the length of each continuous subset of samples per reception period and the frequency component detected in the DFT or FFT are matched to the same radar detection distance.

Although a typical gate utilization period is 50%, according to the invention, the utility period can range from 10 to 90%.

15

Claims (6)

A method for processing a Frequency Modulated Interrupt Continuous Wave Radar (FMICW) signal, the method comprising the steps of: - generating a signal whose frequency changes within a set bandwidth (B); and off time slots, 10 - transmitting a gated signal to a desired destination by means of a transmitter and an antenna, - receiving a reflected or scattered signal of an off-gate signal. and detecting and collecting the desired characteristics of the received signal, characterized by 20. detecting and / or processing the desired portion of the off-time slots to obtain gated signal information at a desired distance and to minimize signal noise associated with the desired distance. Method according to claim 1, characterized in that only the beginning of the off-25 time slot is processed. A method according to claim 1, characterized in that only the end of the off-time slot is processed. Method according to claim 1, characterized in that the transmission benefit period due to gating is about 50%. i2 110966 Method according to any one of the preceding claims, characterized in that the gate frequency is varied within a range between 1 and 50% of the bandwidth (B). Method according to any one of the preceding claims, characterized in that the method is used for geophysical remote sensing. 13,10966