Radar device related improvements
By adopting cyclic processing and block storage mode in the FMCW radar device, data is transmitted to remote memory in grid mode, solving the problem of local memory capacity limitation, improving data transmission and processing efficiency, and achieving faster range and speed data processing.
Patent Information
- Application Number
- CN201910327256.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2018-04-23
- Filing Date
- 2019-04-23
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2039-04-23
AI Technical Summary
When processing large amounts of data, the existing FMCW radar devices are limited by the capacity of high-speed local memory, and need to frequently transmit data to slower remote memory, resulting in limited overall processing speed and it is difficult to achieve good resolution within a wide range of distances. .
By cyclic processing data in the radar device, the data of each linear frequency modulation pulse is first captured and processed in the local memory, and then the data block is transferred to the remote memory in grid mode, and Fourier transform is performed before the data is transferred, using the block The storage and interleaved data modes reduce the demand for local memory and improve data transmission efficiency.
By reducing the demand for local memory, the data transmission speed and processing efficiency are improved, the problem of limited data transmission speed is avoided, and faster range and speed data processing is achieved.
Smart Images

Figure CN110389337B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to improvements in FMCW radar apparatus, particularly for mounting to a vehicle such as a car or truck. Background Art
[0002] It is known to use radar in distance measurement. In one application, radiation in the microwave region of the spectrum is emitted from a source towards a target. The target reflects some of this radiation back towards a detector, which is sensitive to radiation in the microwave region of the spectrum. A microcontroller can then be used to make a comparison between the emitted signal and the measured, detected signal, from which the range of the target from the source / detector can be estimated.
[0003] Radar devices mounted to vehicles can be used to detect the presence of objects around the vehicle, which can be used as part of a vehicle autonomous or semi-autonomous system (such as active cruise control).
[0004] It is known to provide a frequency shift keying scheme to determine range / distance from a signal. When the transmitted signal is modulated by a small step change in frequency, the distance can be measured from the phase difference seen in the returned signal. The accuracy of the measurement is related to the frequency step size. Similarly, relative speed can be measured by looking at the phase shift changes between closely spaced linear frequency modulation pulses (chirps).
[0005] There has been a great deal of development work in the field of radar systems, and there are many types of radar transmission technologies available. One technology that has proven suitable for automotive applications is called frequency modulated continuous wave (FMCW) radar. In an FMCW radar, the frequency of the signal transmitted by the antenna array is ramped up or down within a frequency range that defines a linear frequency modulation pulse. This can be relatively simply generated using a voltage controlled oscillator (VCO). The data collected from each linear frequency modulation pulse will include a set of possible 1024 data samples for each antenna in the array. The antenna array can be of any suitable size, such as a 16-antenna radar array. These data samples can then be processed using a discrete Fourier transform to determine a set of values for the range bins of the linear frequency modulation pulses. These can then be processed together with the range bin values for a series of linear frequency modulation pulses to produce a final set of velocity bin data for the radar device.
[0006] In a practical implementation, each chirp provides a collection of range measurements. The output of a DFT taken on a chirp is a collection of complex range bins whose frequency and phase indicate the precise range. Since the range changes from one chirp to the next (because of relative motion), the phase (and ultimately, the frequency) changes between chirps. By performing a Fourier transform (such as a Fast Fourier Transform (FFT)) across consecutive chirps in a given range bin, it is possible to extract the frequency (i.e., the rate of change of phase over time), which is related to the relative velocity.
[0007] A typical FMCW radar device includes a microcontroller, which typically uses a Fast Fourier Transform (FFT) or FFT sequence to perform the function of processing data from the antenna. The microcontroller includes a certain amount of fast access memory, which is usually integrated into the same chip as the processing unit (usually a programmable logic controller). This memory is referred to as local storage memory in this article. This local storage memory is physically close to the processor core of the programmable logic controller, so reading and writing to it is faster. It is used for short-term storage of data being processed during device operation.
[0008] The system also includes external memory, referred to herein as remote burst access storage memory, which is located outside the main processing element chip and physically farther away from the processing unit, so reading from it and writing to it is slower. This memory is usually provided as a dedicated memory chip and communicates with the processing unit across a memory bus. Remote burst access memory is also usually constrained so that data can be written or read efficiently in bursts of fixed line lengths. True random access (which is desirable for FFT operations) is very inefficient and reduces performance compared to burst access memory.
[0009] The local storage memory is designed for true random access but the remote memory is designed to allow burst sequences of memory, each burst consisting of a fixed length of data line to be transferred to the local storage memory according to a request made on the memory bus. A burst may consist of a data line of at least 2 bytes of data or more in length. A sequence of bursts may be written to any number of rows in memory, but it is generally fastest to write a sequence of bursts corresponding to a block of consecutive rows in memory.
[0010] A problem with prior art radar devices is that in order to achieve good resolution over a wide range of distances, very large amounts of data must be captured and processed. This often exceeds the amount of high speed local memory available to the processor, requiring the data to be transferred out to remote memory during signal processing and then transferred back for more processing. Because this transfer is relatively slow, the overall speed of the radar in generating range data will be limited. It is not economical to provide enough high speed local memory to avoid having to do this when writing. Summary of the invention
[0011] It is an object of the present invention to improve upon limitations present in prior art methods of processing radar data in which large amounts of data are processed.
[0012] According to a first aspect, the invention provides a method of operating a radar apparatus comprising an antenna array of N antennas and a processing unit receiving data captured from the antenna array and comprising an area of local storage memory and a larger area of remote burst access memory, the method comprising for a single cycle of the radar comprising two or more chirps of a transmitted radar frequency signal:
[0013] (a) capturing data of a first chirp in a cycle, which generates D data samples from each of the N antennas, and storing the data in a local storage memory,
[0014] (b) processing the captured data using a Fourier transform to generate sets of range interval data, one set of range interval data for each antenna, wherein each set includes R range intervals,
[0015] (c) transferring range interval data for chirps from a local storage memory to a remote burst access memory in a sequence of bursts, the data being arranged in a grid pattern in the remote burst access memory, the grid pattern comprising data-filled rows containing two or more unique range intervals of data for the chirps, the two or more unique range intervals being maintained in contiguous strips of the remote burst access memory, the data-filled rows being vertically separated from adjacent data-filled rows by a plurality of rows that do not contain any bursts of data associated with the chirps,
[0016] (d) repeating steps (a) and (b) for each subsequent chirp data in the loop,
[0017] (e) repeating step (c) after the range interval data for each subsequent chirp has been stored in the local storage memory, so as to transfer the range interval data for each subsequent chirp in a sequence of bursts from the local storage memory to the remote burst access memory where the data is arranged in the same grid pattern as used in step (c), but vertically offset by one or more rows to fit within the rows in the column to which data for any previously processed chirp in the cycle has not been written,
[0018] (f) transferring, in one or more bursts, at least one continuous vertical block of data from the remote block to the local storage memory, the at least one continuous vertical block of data having a length equal to the number of different range intervals,
[0019] (g) processing the data transferred to the local storage memory using a Fourier transform to generate a set of velocity data comprising a set of velocity intervals, wherein each velocity interval is generated based on the value of a corresponding range interval for each antenna of the array, and
[0020] Steps (f) and (g) are repeated until all data in the remote burst access memory data block has been processed by the second FFT.
[0021] Applicants have recognised that by transferring data from a fast local storage memory to a remote burst access memory in a repeating pattern of a single data line for each linear frequency chirp prior to performing a full Fourier transform on all linear frequency chirps in a cycle, it is possible to use significantly less local storage memory than in prior art arrangements where an FFT is performed on all linear frequency chirps prior to transferring the data.
[0022] By storing the data in blocks in the remote memory (each block corresponds to one chirp), fast transfer back to the local memory is possible. Block storage is achieved in a very efficient manner by processing one chirp at a time in the local storage memory and writing out to alternate locations in the remote memory to form a grid and then repeating to interleave the data.
[0023] The method in step (f) can transfer all the data corresponding to one chirp as a single burst. This is possible because the data is all present in one continuous vertical block in the remote memory, which is much faster than bursting the data out in many single bursts.
[0024] In steps (c) and (e), the method may store data in the burst access memory in a chirp order while vertically scanning down the memory block.
[0025] In steps (c) and (e), the method may transmit data in bursts of at least 2 bytes and at most N bytes in length, where N is an integer value less than the total number of antennas. The number of bytes that can be transmitted at one time depends on the relationship between the available burst length and the length of each byte of data.
[0026] In steps (c) and (e), the method may include transmitting the data as a sequence of bursts to form a pattern of data in the remote burst access memory, wherein each burst contains data for at least a corresponding pair of range intervals in the set of range intervals, wherein the data for one range interval is written to a left portion of a column in the grid and the data for the other range interval is written to a right portion of the column.
[0027] The sequence used in steps (c) and (e) may transmit a first single burst corresponding to range intervals 0 and 1, followed by subsequent single bursts corresponding to range intervals 2 and 3, and so on until all range intervals are transmitted to the remote memory to form a pattern of stored data.
[0028] Thus, reading the rows horizontally, for each entry in the column of data, there will be values for two range intervals, one after the other.
[0029] To maximize the speed of data transfer, step (b) may include writing the range interval values generated by the Fourier transform to a fast local memory, wherein each pair of range intervals in the range interval set is saved as a single data line, with the value for the first data interval in the left half and the value for the second data interval in the right half.
[0030] If more than two range intervals fit in one burst, then the speed can be increased in step (c) by transmitting three or more range intervals at a time in each burst.
[0031] The method can be applied to cycles including more than two chirps simply by repeating steps (a) and (b) as many times as there are chirps, and then repeating step (c) for each chirp with an appropriate vertical offset of the grid. For example, it can be applied to a total of 512 chirps in each cycle.
[0032] Steps (c) and (e) may each include transferring the data into a grid whereby all data is held in a single column having a width equal to the burst length. This column may store the range interval values for all chirps in the cycle in a single column. Thus, the width of the column may be two or more bytes.
[0033] In the presence of R range data intervals, after the data for the cyclic linear frequency modulation pulse is transmitted, the data can be arranged in a pattern in the remote memory so that all even range interval data is located on the left hand side of the burst-sized data line in the remote data block, and all odd range interval data is stored on the right hand side of the burst-sized data line in the remote data block.
[0034] To achieve this, in step (c), the data may be stored in the remote memory in a pattern whereby large gaps equal to C-1 rows are left between subsequent individual burst lines, thereby ensuring that all subsequent repetitions of step (f) will be interleaved into these gaps without overwriting any of the previously transmitted data from step (c) or step (f).
[0035] In a preferred arrangement, after all the data is transferred to the remote memory, it may form a block of data with no empty rows and wherein, scanning down the columns of the block, all data values for the N antennas for a given range interval are stored sequentially.
[0036] A block of data can be stored in a single column. In some cases, this can allow the data to be read back in a longer single burst, which is possible with some remote burst access memory types, to achieve even higher efficiency.
[0037] Steps (f) and (g) may be performed for each loop only after all data interval values for all antennas and all chirps have been transferred to the remote memory.
[0038] In a preferred arrangement, the method may be applied to a radar apparatus having N = 16 antennas in the antenna array, which may be identified as A0 ... A15.
[0039] For each antenna, there may be D=1024 real-valued data samples in each chirp, where each data sample is numbered D0...D1023, each corresponding to a different frequency in the chirp. Alternatively, the method may generate D / 2 complex-valued data samples.
[0040] The first FFT performed in step (b) can produce R complex value range bins for each chirp, where R is defined as D / 2. For each set of range bin data, this will result in 256 data lines in the long-range data block, with odd ranges on the right side of each data line and even ranges on the left side.
[0041] The first Fourier transform step may include a fast Fourier transform of the type described in the textbook "An algorithm for the machine calculation of complex Fourier series" Math. Comput. 19:297-301, written by Cooley, James W. and John W. Tukey in 1965.
[0042] Similarly, the second Fourier transform step may comprise a Fast Fourier Transform of the type described in this textbook.
[0043] For each cycle, the method may be repeated for a set of 512 chirps.
[0044] The method may of course be repeated for a series of cycles during operation of the radar device.
[0045] According to a second aspect, the present invention provides a frequency modulated continuous wave (FMCW) radar apparatus comprising: an antenna array of N antennas, a processing unit that receives data captured from the antenna array and includes an area of fast local storage memory and a larger area of remote burst access memory, and a program set or a logic instruction set stored in the memory area, which program set or logic instruction set is executed by the processing unit in use, wherein the method of the first aspect of the present invention is performed.
[0046] The antenna array may include a set of N antennas that receive the reflected analog signal, which is fed into an analog to digital converter that converts the analog signal into a digital signal.
[0047] The device may also include a transmitter array of N antennas that transmits a radar frequency signal from the device that is within a range that can be detected by the N antennas.
[0048] The device's processing unit can generate drive signals for the transmitter array, or can configure a remote radar processing chip to generate the required radar signals.
[0049] According to a third aspect, the present invention provides a set of program instructions maintained in a memory of a processing unit of a radar apparatus, which, when executed by the processing unit, causes the radar apparatus to perform the method of the first aspect of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] An embodiment of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:
[0051] Figure 1 is a block schematic diagram of a radar apparatus according to the present invention;
[0052] Figure 2 is a representation of the data arrangement in the local storage memory for one chirp N=n, showing how each data line stores a range data interval value for a range data interval pair;
[0053] Figure 3 shows the time sequence in which a data pattern is transferred to a remote memory one burst at a time to build a remote data block;
[0054] Figure 4 The final data block stored in the remote memory is shown;
[0055] Figure 5 shows the source of the first vertical bar of data being transferred from remote memory back to local storage memory for second stage FFT processing;
[0056] Figure 6 A second vertical bar showing the source of data being transferred from remote memory back to local storage memory for second stage FFT processing; and
[0057] Figure 7 and Figure 8 Shows Figure 5 and Figure 6 The destination results (in local storage memory); and
[0058] Fig. 9 Shows Figure 1 A complete set of method steps implemented by a device during one cycle of processing radar data. DETAILED DESCRIPTION
[0059] Figure 1 An exemplary FMCW radar device according to the present invention is shown. The device includes a transmitter antenna array 10 of N antennas and a receiver antenna array 20 of N antennas. In this example there are 16 antennas, A0 to A15. The device also includes a processing unit that sends high frequency analog signals in the form of linear frequency modulation pulses to the transmit antenna array and receives digitized data captured from the receiver antenna array. Between the processor of the processing unit and the transmit antenna is a VCO 30, and between the receive antenna and the processor is an RX mixer. The functions of the VCO and the mixer are clear to those skilled in the art.
[0060] The processing unit may include appropriate digital to analog converter circuitry and analog to digital converter circuitry, as shown, but more typically the output of the radar antenna array 20 is already in a digitized format. The A / D circuitry converts the received data from the mixer into a stream of data values. The processing unit includes a processor chip and a region of fast on-chip local storage memory 50. The processing chip 40 is encoded with a programmable logic instruction set. The device also includes a region of remote burst access memory 60. This is a memory that is slower to access than local memory, but is optimized for burst access (e.g., data of multiple bytes in length).
[0061] The skilled person will appreciate that Figure 1 The arrangement of the functional blocks shown in is not intended to limit the scope of the present invention. The VCO and the mixer may be implemented as part of the processing unit, and the A / D converter and the D / A converter may be implemented outside the processing unit.
[0062] Figure 8 Shown by Figure 1 The steps of implementing the embodiment of the radar device shown in FIG.
[0063] In operation, the device transmits a chirp of frequency from the transmit antenna array, and any echo signal reflected onto the receiver antenna array generates a corresponding set of data samples. In the described embodiment, each chirp generates 1024 data points for each of the 16 antennas, so a total of 1024*16 data points are generated for each chirp. The radar device transmits multiple chirps for a single complete processing cycle, and in this embodiment, a complete cycle consists of 512 sequentially transmitted chirps (C0 to C511). The antennas can be uniquely identified by a digital sequence from A0 to A15, and can be spatially arranged in this order in the antenna array.
[0064] Each data point for an antenna will correspond to a single voltage sample captured by the ADC, and will have a value indicative of the amplitude of the echo signal detected by the antenna at the associated instant.
[0065] The combination of the number of samples in each chirp, the high number of chirps per cycle, and the number of antennas means there is a lot of data to process - around 16MB per radar cycle.
[0066] For each chirp, an FFT of the data is performed to generate a set of range interval results for each antenna. Since there are 16 antennas in this example, 16 range interval data sets will be generated for each chirp. Each range interval set in this example consists of B individual range intervals, where B equals 512, so that each range interval can be uniquely labeled R0 to R511. Of course, this is not intended to be limiting, as in practice the number of antennas in the array or the number of data samples per chirp are not intended to be limiting.
[0067] As will be explained, the apparatus performs an FFT on one chirp before writing out some of the resulting range interval data to remote memory, but need not wait for all chirps to be processed.Each FFT creates a complete set of range intervals R0-511.
[0068] In practice, as explained in the following paragraphs in this described embodiment, data is written out each time a pair of range intervals in a set of range intervals for a chirp is calculated for each antenna. This corresponds to a total of only 32 data intervals needing to be calculated by the FFT between each data transfer, corresponding to two sets of 16 range interval values for each range interval, one value for each antenna.
[0069] In the first FFT processing stage, data for the first linear frequency modulation pulse begins to be processed within the processing unit before any data is written out to the remote burst access memory. For each of the 16 antennas, the FFT generates all range interval values R0-R511 for the first linear frequency modulation pulse. Taking advantage of the random access nature of the fast local storage memory, the data is written to the local memory in a pattern that is conducive to quickly transferring the data to the remote burst access memory. Specifically, the data is stored in the fast local storage memory, where the values for each antenna in range interval 0 are first, then the values for each antenna in range interval 1, and so on. Each set of 32 values (2 range intervals, 16 antennas) is located in a continuous data line. In the context of this description, the term "data line" can also be interpreted as a "data burst", that is, the smallest efficient data block that the remote burst access memory uses to work. This is in the context of the present description. Figure 2 , where n is any integer value between 0 and the total number of chirps in the cycle.
[0070] Once the chirps are processed and stored in the local storage memory, each line is transferred to the remote burst access memory as a single data burst. This is possible because the length of the data line is the same as the minimum length of a single burst. Transferring means that the data in the local storage memory is written out to the remote burst access memory; it can also remain in the local storage memory until it is subsequently overwritten because it no longer needs to be processed.
[0071] After the first data line is transmitted, each data line is transmitted in sequence. This is transmitted to the remote memory. Importantly, the second row is stored in the burst access memory at a position offset by C-1 rows relative to the first row of data, that is, if the first data line is written to row 0, then it is written to row 512. Subsequent bursts are also similarly offset.
[0072] This process is repeated for all B / 2 data lines needed to store the data that makes up the range interval data for the complete 512 sets of the first chirp.
[0073] A sequence of 512 individual data transfers for all data intervals generated from the first chirp is formed in the remote memory Figure 3 The data pattern shown in .
[0074] The next chirp is then processed in exactly the same manner and transferred to the remote memory one data line at a time. However, the starting row position for writing data to the remote memory in the transfer is offset by one row to prevent the new data from overwriting the first row of data transferred. The resulting data stored in the remote memory is as follows: Figure 3 as shown in .
[0075] This process is repeated for each of the remaining 510 linear frequency modulation pulses, with the offset increasing by one row each time. Finally, a data block will be formed in the remote burst access memory as shown. It can be seen that the horizontal dimension of the data lines in the remote data block is antenna 0 to antenna 15, twice, once for the even range intervals and once for the odd range intervals. The vertical dimension has the data for range intervals 0 and 1 stored in the first and subsequent rows, as well as the data for range intervals 2 and 3 offset by 512 rows from row 1, the data for range intervals 4 and 5 offset by 2×512 rows from row 1, and so on. At the end of processing of the last linear frequency modulation pulse in the cycle, there are no gaps in the data block in the remote burst access memory. All of the data associated with each pair of range intervals is located together in the remote burst access memory. This is done in the remote burst access memory. Figure 3 besides Figure 4 Shown in.
[0076] Storing data in the remote burst access memory in the above described mode allows all data for the 2 range intervals to be easily burst back to the fast local storage memory in burst blocks or a single long burst into the memory address block in the local storage memory, thereby allowing the second stage FFT set to be performed efficiently.
[0077] Data can be transferred back from remote storage in the following manner.
[0078] Initially, the entire block of data corresponding to range intervals 0 and 1, which is 512 lines long, is burst (512 bursts) into the local storage memory. This data is processed by the second stage FFT, which works vertically on the data for range interval 0 on the left side of each data line. This data is then processed by the FFT working vertically on range interval 1 on the right side of the data line. This is done in Figure 5 Shown in.
[0079] Further processing is then performed on the so-called "velocity processed data". This can occur in parallel with the reading in and FFT processing of the next pair of intervals, but does not have to.
[0080] After this is completed, another block of data corresponding to range intervals 2 and 3 is transmitted in exactly the same manner as the first block. This data is processed by the second stage FFT, which works vertically on the range interval 2 data on the left side of each data line. This data is then processed by the FFT working vertically on the range interval 3 data on the right side of the data line. This is done in Figure 6 Shown in.
[0081] This process is repeated for the blocks corresponding to range intervals 4 and 5, and so on, until the final range intervals 510 and 511 are transmitted and processed to form velocity data.
[0082] The final result can then be processed a little more and possibly also written out back to external storage. Again, this can be done in a long burst for the result pair.
Claims
1. A method of operating a radar apparatus comprising an antenna array of N antennas and a processing unit receiving data captured from the antenna array and comprising an area of local storage memory and a larger area of remote burst access memory, the method comprising performing the following operations for a single cycle of the radar comprising two or more chirp pulses of a transmitted radar frequency signal: (a) capturing data of a first chirp in said cycle, and storing said data in a local storage memory, said capturing generating D data samples from each of N antennas, (b) processing the captured data using a Fourier transform to generate sets of range interval data, one set of range interval data for each antenna, wherein each set includes R range intervals, (c) transferring range interval data for the chirps from the local storage memory to a remote burst access memory in a sequence of bursts, wherein the data is arranged in a grid pattern, the grid pattern comprising data-filled rows containing two or more unique range intervals of data for the chirps, the two or more unique range intervals being maintained in contiguous stripes in the remote burst access memory, the data-filled rows being vertically separated from adjacent data-filled rows by a plurality of rows that do not contain any bursts of data associated with the chirps, (d) repeating steps (a) and (b) for each subsequent chirp data in the cycle, (e) repeating step (c) after the range interval data for each subsequent chirp has been stored in the local storage memory, to transfer the range interval data for each subsequent chirp from the local storage memory to the remote burst access memory in a sequence of bursts, wherein the data is arranged in the same grid pattern as used in step (c), but is vertically offset by one or more rows to fit within the rows in the column to which the data for any previously processed chirp in the cycle has not been written, (f) transferring, in one or more bursts, at least one continuous vertical block of data from a remote block from the remote burst access memory to the local storage memory, the at least one continuous vertical block of data having a length equal to the number of different range intervals, (g) processing the data transferred to the local storage memory using a Fourier transform to generate a set of velocity data comprising a set of velocity intervals, wherein each velocity interval is generated based on the value of a corresponding range interval for each antenna of the array, and Steps (f) and (g) are repeated until all data in the remote burst access memory data block has been processed by the second FFT.
2. The method of claim 1, comprising: In steps (c) and (e), data is stored in the burst access memory in chirp order as the memory block is scanned vertically down, whereby the grid pattern is vertically shifted by one row for each subsequent chirp.
3. The method of claim 1, comprising: In steps (c) and (e), data is transmitted in bursts of at least 2 bytes and at most N bytes in length, where N is an integer value less than the total number of antennas.
4. The method of claim 3, comprising: In steps (c) and (e), data is transmitted as a sequence of bursts to form the grid pattern of data in the remote burst access memory, wherein each burst contains data for at least a corresponding pair of range intervals of a set of range intervals, wherein the data for one range interval is written to a left portion of the column in the grid and the data for the other range interval is written to a right portion of the column.
5. The method of claim 1, wherein: Steps (c) and (e) each include transferring the data into a grid whereby all data is held in a single column having a width equal to the burst length.
6. The method of claim 5, wherein: In the presence of R range data intervals, the method transmits the data arranged in a pattern to the remote burst access memory after the data for the cyclic linear frequency modulation pulse is transmitted, so that all even range interval data are located on the left hand side of the data line of the burst size in the remote data block, and all odd range interval data are stored on the right hand side of the data line of the burst size in the remote data block.
7. The method of claim 1, wherein: The data transferred to the remote burst access memory is stored in a pattern whereby the number of empty rows left between subsequent individual burst lines is equal to C-1 rows, thereby ensuring that all subsequent repetitions of step (f) will be interleaved into the gaps without overwriting any of the previously transferred data from step (c) or step (f).
8. The method of claim 1, wherein: After all the data is transferred into the remote memory, it forms a data block with no empty rows and where, scanning down the columns of the block, all the data values for the N antennas for a given range interval are stored sequentially.
9. A FMCW radar device, comprising: An antenna array of N antennas, a processing unit that receives data captured from the antenna array and includes an area of fast local storage memory and a larger area of remote burst access memory, and a program set or logic instruction set stored in the memory area, which program set or logic instruction set is executed by the processing unit in use, wherein the method as claimed in any one of claims 1 to 8 is performed.
10. The FMCW radar device according to claim 9, wherein: The antenna array comprises a set of N antennas which receive reflected analog signals which are fed into an analog to digital converter which converts the analog signals into digital signals.
11. A computer program product stored in a memory of a microcontroller of a radar device, the computer program product having a program instruction set stored thereon, which, when executed by the microcontroller, causes the device to perform the method according to any one of claims 1 to 8.
Citation Information
Patent Citations
Signal-based data compression
CN105277940A
Radar systems and methods thereof
CN107589406A
Dynamic programming of chirps in FMCW radar system
CN107710011A
Transceiver device has chirp signals generated within transmitter by accessing chirp sequences stored in memory
DE10215798A1
Radar data compression system and method
EP2980996A1
