CIRCUITS AND METHODS FOR DETERMINING CHIRP SIGNAL LINEARITY AND PHASE NOISE OF AN FMCW RADAR

DE602016090167T2Active Publication Date: 2024-11-06TEXAS INSTRUMENTS INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
DE602016090167
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Priority Date
2015-12-18
Filing Date
2016-12-19
Publication Date
2024-11-06
Estimated Expiration
2036-12-19

AI Technical Summary

Technical Problem

Current methods for measuring frequency linearity and dynamic phase noise in FMCW radars are hindered by the high cost and limited bandwidth of existing test equipment, making it difficult to accurately assess the purity of chirp signals, especially for wider bandwidths.

Method used

The use of generic signal analyzers and signal processing algorithms to down-convert and digitize chirp signals, allowing for the calculation of frequency linearity and dynamic phase noise through polynomial fits and power spectral density analysis, even when the chirp signal bandwidth exceeds the equipment's capabilities, by segmenting the signal into manageable portions.

Benefits of technology

Enables accurate and cost-effective measurement of frequency linearity and dynamic phase noise, overcoming the limitations of traditional test equipment and allowing for the assessment of chirp signal purity across a broader range of frequencies.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader
Need to check novelty before this filing date? Find Prior Art

Description

BACKGROUND

[0001] FMCW (frequency modulated continuous wave) radars are systems that transmit a continuous wave, which modulates a periodic function such as a sine wave or saw-tooth waveform. The main constituent of a FMCW radar is the chirp signal or waveform. A chirp signal is a signal whose frequency increases or decreases with time. The chirp signal has to be extremely linear in order to avoid phase distortion effects seen at the receiver of the radar.

[0002] The radars have a loop filter in their local oscillator (LO) subsystems that is optimized for continuous wave phase noise, but may not be optimal for a dynamic phase noise. Dynamic phase noise is the term used to refer to the phase noise that is associated with a chirp signal that is being transmitted over a certain bandwidth. The measurements of frequency linearity and dynamic phase noise are important in FMCW radars because they determine the purity of the waveform of the chirp signal transmitted by the radars.

[0003] The current solution involves the use of test equipment to measure the frequency linearity and phase noise. However, most test equipment is expensive and time consuming to use. The more affordable test equipment typically does not have the bandwidth to measure the frequency linearity and / or they do not have the capability to operate on a wider chirp signal bandwidth to measure dynamic phase noise. US 2009 / 224963 discloses a FMCW radar apparatus. DE 10 2013 207 464 relates to a measurement device and method for determining a deviation of a broadband measurement signal from a reference signal. XP055603626 Steffen Heuel: "Automated Measurements of 77 GHz FMCW Radar Signals; Application Note" relates to an application note concerning automated measurements of 77 GHz FMCW radar signals. XP017447665 Geoffrey M Garner, Huawei Technologies Co et al: "Further Analysis of RMS Frequency Offset for a CPRI Client Transported over OTN; C 972", ITU-T Draft; study period 2009-2012, International Telecommunication Union, Geneva; CH, vol. 13 / 15 and US 2008 / 122427 represent further state of the art.SUMMARY

[0004] The invention is defined by the features of the appended claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] FIG. 1 is a graph showing an example of a chirp signal as a function of time. FIG. 2 is a flowchart describing a not claimed embodiment of measuring chirp signal linearity of a transmitted chirp signal. FIG. 3 is a flowchart describing a not claimed embodiment of measuring dynamic phase noise of a transmitted chirp signal. FIG. 4 is a flowchart describing the procedure to compute the metrics of a chirp signal in situations where the chirp signal bandwidth cannot be accommodated by the analyzing equipment. FIG. 5 is a block diagram of test equipment that measures frequency linearity and dynamic phase noise of a radar. FIG. 6 is a block diagram of an embodiment of more detailed test equipment that measures the frequency linearity and dynamic phase noise of a chirp signal. FIG. 7 is flowchart describing a method for determining frequency linearity of a chirp signal. FIG. 8 is a flowchart describing a method for determining dynamic phase noise of a chirp signal. DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

[0006] Circuits and methods that validate frequency linearity and dynamic phase noise of frequency modulated continuous wave (FMCW) radars are disclosed herein. Frequency linearity is the linearity of a chirp signal transmitted by a radar. Dynamic phase noise is the term used to refer to the phase noise that is measured when a chirp signal is being transmitted over a certain bandwidth. The circuits and methods are described below based on an example chirp signal having a waveform with the following parameters: transmitted chirp signal bandwidth: 0.025GHz chirp start frequency: 2.375GHz chirp stop frequency: 2.4GHz slope rate: 1000MHz / us

[0007] FIG. 1 is a graph showing an example of a chirp signal 100 as a function of time that has the above described parameters. The chirp signal 100 repeats continuously, starts at a start frequency of 2.375GHz, and ends at a stop frequency of 2.4GHz. The frequency rise between the start frequency and the stop frequency corresponds to the portion of the chirp signal 100 where frequency linearity is measured. More specifically, a slope rate that is constant between the start frequency and the stop frequency along a line 102 corresponds to a chirp signal that has high frequency linearity. If the line 102 is curved or has discontinuities, the chirp signal has low frequency linearity. Thus, the straightness of the line 102 plotted between the start frequency and the stop frequency is related to the frequency linearity.

[0008] Additional reference is made to FIG. 2, which is a flowchart 200 describing a not claimed embodiment of determining chirp signal linearity of a transmitted chirp signal. In step 202, the transmitted chirp signal is captured. In some examples the chirp signal is captured using a generic signal analyzer. For example, commonly available spectrum analyzers may be used to capture the transmitted chirp signals. The spectrum analyzers further function as signal analyzers even with high bandwidth chirp signals because they down-convert the chirp signal using their own internal or locally produced local oscillator (LO). The chirp signal is downconverted to its equivalent base-band intermediate frequency (IF) samples in step 204 and is digitized to yield a digitized IF signal, which is referred to as the IF signal in step 206.

[0009] The IF signal is quadrature in nature (typically digital I and Q samples). The bandwidth of the spectrum analyzer may or may not accommodate the bandwidth of the transmitted chirp signal, so the analysis bandwidth of the spectrum analyzer may be set to its maximum value in order to analyze the greatest portion of the chirp signal as possible. In some examples, the analysis bandwidth of the spectrum analyzer is approximately 40MHz.

[0010] The amplitude of the digitized quadrature samples that are captured by the spectrum analyzer are chosen in a way that it does not clip or saturate the dynamic range of the digitizer and so that the digitized quadrature samples are not buried under the noise floor of the digitizer. The amplitude selected should be optimal and should be in the linear range of operation of the digitizer. In some examples the IF signal is processed to compute the phase of the IF signal. The derivative of the phase of the IF signal with respect to time is calculated to yield the instantaneous frequencies of the chirp signal. In step 208, the instantaneous frequencies of the chirp signal are determined, which are sometimes referred to as the computed frequency samples.

[0011] The computed frequency samples can have frequency offsets caused by the frequency difference between the transmitter's LO frequency, which is typically the chirp signal's LO, and the receiving equipment's LO. The difference in the starting frequency of the chirp signal, as captured by the equipment, and the actual start frequency used for the transmission of the chirp signal 100 can also cause a frequency offset. The frequency offset may be removed at step 210, by conventional techniques.

[0012] Ideal frequency samples of the computed frequency samples are determined from the computed frequency samples. In some examples this is determined by applying a linear fit in the form of a first order polynomial fit, which may include simple terms such as slope and intercept, to the computed frequency samples. In some examples, the polynomial for the polynomial fit is f(t) = at + b, where 'f' is the ideal frequency samples, 't' is the time base; 'a' is slope, and 'b' is the intercept. All the variables are the polynomial coefficients that give the best fit for the computed frequency samples.

[0013] Predetermined frequency samples are frequency samples based on a mathematical model of the chirp signal. The difference between the ideal and predetermined frequency samples are determined in step 212 to yield the frequency error samples. Frequency linearity is calculated in step 214 from the frequency error samples. In some examples, statistical operations such as mean standard deviation on the frequency error samples yield a measure of the frequency linearity.

[0014] Having described process for measuring frequency linearity, processes for measuring dynamic phase noise will now be described. The dynamic phase noise is continually analyzed over the bandwidth of the chirp signal and is specified at one particular offset frequency. In some examples, the dynamic phase noise is referred as being offset 100KHz, 1MHz, and 10MHz from the carrier signal. In the case of FMCW radar, the carrier signal is not a single CW frequency, but constantly changes. For example, it may be the ramp signal shown in FIG. 1 that is specified over a particular bandwidth. Accordingly, the dynamic phase noise is specified as the phase noise of the ramp at an offset frequency, such as 10KHz, 1MHz or 10MHz.

[0015] FIG. 3 is a flowchart 300 describing a not claimed embodiment for determining the dynamic phase noise of a transmitted chirp signal. Phase of the chirp signal is determined at step 302. Phase samples of the chirp signal are determined at step 304. Phase samples can either be computed from the frequency samples or from the phase computed from the digitized I and Q samples. In the former case, the frequency error samples are integrated or summed up and divided by 2π to yield phase error samples. In the latter case, a linear polynomial fit (similar to linear fit on the computed frequency samples) is applied to compute phase samples. An example of the polynomial used is as follows: Φ(t) = at + b, where Φ is the phase samples of the chirp signal, 't' is the time base; 'a' is slope, and 'b' is the intercept. These variables are the polynomial coefficients that give best fit for the phase samples, which is sometimes referred to as the measured phase samples.

[0016] Predetermined phase samples are calculated in step 306. The predetermined phase samples are based on a mathematical model of an ideal chip signal. The difference between the phase samples and the predetermined phase samples is determined at step 308 to yield phase error samples. The power spectral density function of the phase error samples is calculated in step 310 to yield the dynamic phase noise of the transmitted chirp signal. The power spectral density of the phase error samples is useful to compute the noise power over a certain frequency offset.

[0017] FIG. 4 is a flowchart 400 describing a procedure for computing the metrics of a chirp signal in situations where the chirp signal bandwidth cannot be accommodated by the analyzing equipment. In summary, the entire bandwidth of the chirp signal is segmented into segments, whereby each segment is analyzed. In this example, the chirp signal has the following characteristics: transmitted chirp signal bandwidth: 70.3MHz chirp start frequency: 2.375GHz chirp stop frequency: 2.445GHz equipment bandwidth: 40MHz number of segments: 2 segmented bandwidth: 35MHz segment 1 start and stop frequencies: 2.375 - 2.41GHz segment 2 start and stop frequencies: 2.411 - 2.445GHz

[0018] As described above, the chirp signal bandwidth is greater than the bandwidth of the device or equipment analyzing the chirp signal. In step 402, the bandwidth of the chirp signal is segmented into a plurality of segments, whereby each segment has a bandwidth that is within the bandwidth of the analyzing equipment. In the example described herein, the bandwidth of the chirp signal is segmented into halves of equal bandwidth which are within the bandwidth of the analyzing equipment. The segmented halves start and stop frequencies are entered into the analyzing equipment to extract the digitized IQ samples for each segment. In step 404, each segment of the chirp signal is analyzed per FIGS. 2 and 3 to determine the metrics of each segment. With regard to the metrics described above, the first segment and the second segment are analyzed individually to determine their metrics. In step 406, the root-mean-square of the metrics of each segment are added together to yield the metrics of the entire chirp signal. The metrics include the frequency linearity and the phase noise.

[0019] FIG. 5 is a block diagram of test equipment 500 that measures frequency linearity and phase noise of a radar 502. In some examples, the test equipment 500 is automated and measures the frequency linearity and phase noise of the radar 502 during production of the radar 502. The radar 502 has a test mode that outputs a chirp signal. In the example of FIG. 5, the chirp signal has a start frequency of 2.375GHz and stop frequency of 2.41GHz. The test equipment 500 includes an IQ down-converter mixer 504 that is coupled to the output of the radar 502. The IQ down-converter mixer 504 is coupled to an external local oscillator 508 that provides a mixing signal for the IQ down-converter mixer 504. In the example of FIG. 5, the frequency of the local oscillator is 2.5GHz. The mixed signal output by the IQ down-converter mixer 504 is output to an intermediate frequency (IF) digitizer 510 that extracts the digital IQ signal from the mixed signal generated by the IQ down-converter mixer 504. The combination of the IQ down-converter 504 and the IF digitizer 510 serves a similar function as a conventional signal analyzer. A processor 514 analyzes the digitized IQ signals generated by the IF digitizer 510 and determines the frequency linearity and dynamic phase noise of the chirp signal per the methods illustrated in FIGS. 2 and 3.

[0020] The IQ down-converter 504 receives its input in the form of a chirp signal generated by the radar 502. The chirp signal in the examples provided herein is in the frequency range of 2.375GHz to 2.51GHz. The chirp signal is mixed with the external LO generated by the local oscillator 508 to yield a zero IF signal. The zero IF signal is passed through the IF digitizer 510 to produce digitized quadrature (IQ) samples. The signal processing techniques described herein can be applied on the zero IF signals by the processor 514 to evaluate frequency linearity and dynamic phase noise as described above.

[0021] FIG. 6 is a block diagram of an embodiment of more detailed test equipment 600 that measures the frequency linearity and dynamic phase noise of a chirp signal. In some examples, the measurements are achieved on a single integrated circuit chip or by a single circuit. A chirp signal having a defined configuration, such as slope, bandwidth, start frequency, and stop frequency, is transmitted from a radar 602. The frequencies of the chirp signal are scaled down by a divider 604. Scaling down the frequencies of the chirp signal accommodates simpler design of the devices described below. In the example of FIG. 6, the scaling is accomplished by a divider 604 that can divide the chirp signal down by a plurality of possible values, such as 2, 4, 8, 16 and 32.

[0022] The output of the divider 604 is coupled to an input of an auxiliary mixer 606 that is used for the down-conversion process. In the example of FIG. 6, the auxiliary mixer 606 is a complex mixer or a quadrature mixer. An external LO 610 generates a signal that has the same frequencies as the radar 602. In the embodiment of FIG. 6, the external LO 610 generates frequencies in the range of 76GHz to 81GHz. The output of the external LO 610 is input to a divider 612 that may be the same or substantially similar to the mixer 604. For example, the divider 612 may reduce the frequency of the external LO 610 the same as the divider 604 reduces the signal output by the radar 602. The output of the divider 612 is input to an auxiliary LO 616 that is used for down-conversion. The output of the auxiliary LO 616 is input to the mixer 606.

[0023] The output of the mixer 606 is coupled to a receiver 620 that includes an IF amplifier 622 and an ADC 624. In some examples, the IF amplifier 622 is a two-staged, bi-quad IF amplifier that amplifies the output of the auxiliary mixer 606. The output of the amplifier 622 is digitized by the ADC 624. The output of the ADC 624 is processed by a processor 630 per the methods described in the flowcharts 200 and 300 of FIGS. 2 and 3. In some examples, the processing of the processor 630 is performed on the same circuit or integrated circuit as the other portions of the radar 602, so the processing can be absorbed as a part of software / firmware written to characterize the radar 602. For example, the processing may be performed as a self test process in the radar.

[0024] The LO used for the radar 602 cannot be used to feed the auxiliary mixer 606 and the external LO 610 because the phase noise gets cancelled out during the down-conversion and will affect the dynamic phase noise measurement. This problem is overcome by feeding the external LO 610 from an external source. The use of other internal clocks may create noise that will adversely affect the above-described measurements.

[0025] Many laboratory testing devices do not have the capability to measure dynamic phase noise of a chirp signal as described above. The chirp signal frequency linearity measurement typically requires high-end test equipment. The methods and devices described herein overcome the problems associated with measuring the frequency linearity and dynamic phase noise together by using a combination of a generic signal analyzer and signal processing algorithms.

[0026] FIG. 7 is flowchart 700 describing a method for determining frequency linearity of a chirp signal. The method includes down-converting the chirp signal to quadrature signals at step 702. At step 704 the quadrature signals are digitized to yield an IF signal. At step 706 the instantaneous frequencies of the IF signal are determined. At step 708, the instantaneous frequencies of the IF signal are compared to predetermined instantaneous frequencies of the chirp signal to determine frequency linearity of the chirp signal.

[0027] FIG. 8 is a flowchart 800 describing a method for determining dynamic phase noise of a chirp signal. The method includes down-converting the chirp signal to quadrature signals at step 802. The method also includes digitizing the quadrature signals to yield an IF signal at step 804; measuring the dynamic phase of the IF signal at step 806; and comparing the measured dynamic phase of the IF signal to a predetermined dynamic phase of the chirp signal to yield dynamic phase noise of the chirp signal at step 808.

[0028] Modifications are possible in the described embodiments, and other embodiments are possible, within the scope of the claims.

Claims

1. A method of analyzing a chirp signal generated by a frequency modulated continuous wave FMCW radar, the method comprising: down-converting the chirp signal to quadrature signals using an local oscillator (204) which is not located within the radar; digitizing the quadrature signals to yield an IF signal (206); determining the instantaneous frequencies of the IF signal (208); comparing the instantaneous frequencies of the IF signal to predetermined instantaneous frequencies of the chirp signal to determine frequency linearity of the chirp signal (214); wherein the local oscillator is fed from an external source and is not the local oscillator of the FMCW radar; determining dynamic phase noise by calculating the derivative of the phase of the IF signal to yield the instantaneous frequency samples of the IF signal; determining the difference between the instantaneous frequency samples and predetermined frequency samples to yield frequency error samples; integrating the frequency error samples to yield the phase error samples; and calculating the power spectral density function of the phase error samples to yield the dynamic phase noise.

2. The method of claim 1, wherein determining the instantaneous frequencies of the IF signal comprises calculating the derivative of the IF signal.

3. The method of claim 1, further comprising: calculating the derivative of the phase of the IF signal to yield the instantaneous frequencies of the IF signal; fitting the instantaneous frequencies to a line; and determining the difference between the line of the instantaneous frequencies and a line constituting predetermined instantaneous frequencies to yield the frequency linearity.

4. The method of claim 1, wherein the down-converting includes down-converting the chirp signal to quadrature signals using a spectrum analyzer.

5. The method of claim 1, further comprising: analyzing a plurality of bandwidths constituting the chirp signal; and combining the results of analyzing the plurality of bandwidths to yield an analysis of the chirp signal.

6. A method of analyzing a chirp signal generated by a FMCW radar, the method comprising: down-converting the chirp signal to quadrature signals using an local oscillator (802) which is not located within the radar; digitizing the quadrature signals to yield an IF signal (804); measuring the dynamic phase of the IF signal (806); comparing the measured dynamic phase of the IF signal to a predetermined dynamic phase of the chirp signal to yield dynamic phase noise of the chirp signal (808); wherein the local oscillator is fed from an external source and is not the local oscillator of the FMCW radar; determining phase error samples comprising: calculating the derivative of the phase of the IF signal to yield the instantaneous frequencies of the chirp signal; determining the difference between the instantaneous frequencies and predetermined frequencies to yield frequency error; and integrating the frequency error to yield the phase error samples7. The method of claim 6, further comprising calculating the power spectral density of the phase error samples, wherein the power spectral density of the phase error samples yields dynamic phase noise of the chirp signal.

8. The method of claim 6, wherein the phase error samples are determined by fitting a linear polynomial to the measured dynamic phase samples and taking the difference between the linear polynomial and a linear polynomial of the predetermined dynamic phase.

9. The method of claim 6, further comprising determining the frequency linearity of the chirp signal.

10. The method of claim 9, wherein determining the frequency linearity comprises: calculating the derivative of the phase of the IF signal to yield the instantaneous frequency samples of the chirp signal; fitting the instantaneous frequency samples to a line; and determining the difference between the instantaneous frequency samples and a line constituting predetermined frequency samples to yield the frequency linearity.

11. The method of claim 6, wherein the down-converting is performed with a spectrum analyzer.

12. The method of claim 6, further comprising: analyzing a first bandwidth of the chirp signal; analyzing a second bandwidth of the chirp signal; and combining the results of analyzing the first bandwidth and analyzing the second bandwidth to yield an analysis of the chirp signal.

13. A testing device for FMCW radar, the testing device comprising: means for down-converting the chirp signal to quadrature signals using an local oscillator (204) which is not located within the radar; means for digitizing the quadrature signals to yield an IF signal (206); means for determining the instantaneous frequencies of the IF signal (208); means for comparing the instantaneous frequencies of the IF signal to predetermined instantaneous frequencies of the chirp signal to determine frequency linearity of the chirp signal (214); wherein the local oscillator is fed from an external source and is not the local oscillator of the FMCW radar; means for determining dynamic phase noise by calculating the derivative of the phase of the IF signal to yield the instantaneous frequency samples of the IF signal; means for determining the difference between the instantaneous frequency samples and predetermined frequency samples to yield frequency error samples; means for integrating the frequency error samples to yield the phase error samples; and means for calculating the power spectral density function of the phase error samples to yield the dynamic phase noise