Ultra-wideband low-delay signal processing method based on channelization technology
By adopting a multiphase filtering structure in signal processing, the filter operation is moved to data decimation, the filter design is simplified, the problems of large resource waste and delay in traditional methods are solved, and efficient ultra-wideband signal channelization processing is achieved.
Patent Information
- Application Number
- CN202510618537.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-14
- Publication Date
- 2025-08-22
AI Technical Summary
When processing ultra-wideband low-latency signals, the traditional low-pass filter channelization method has problems such as wasted computing resources, high power consumption, large processing delay and high complexity, especially in scenarios where resource limitations or high real-time requirements are difficult to apply.
Using a multiphase filtering structure, the filtering operation is placed after data decimation, and through multiphase decomposition, multiphase branch filter and discrete Fourier inverse transformation, the filter design is simplified, and the computing volume and hardware resource consumption are reduced.
It significantly improves the efficiency of ultra-wideband signal channelization processing, reduces hardware resource requirements and power consumption, simplifies filter bank design, and is suitable for a variety of signal types and channel division methods.
Smart Images
Figure CN120528398A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of ultra-wideband signal processing, and in particular to an ultra-wideband low-latency signal processing method based on channelization technology. Background Art
[0002] In modern communications, radar, and electronic countermeasures, real-time, efficient processing of ultra-wideband, low-latency signals is a key technical requirement. Typically, analog signals from antennas are sampled by high-speed analog-to-digital converters (ADCs) to generate digital signals. Due to the extremely wide signal bandwidth, directly performing complex signal processing (such as demodulation and parameter estimation) on high-rate digital signals poses significant challenges to processor performance and power consumption. Currently, while traditional low-pass filter-based channelization methods can achieve signal band division, they suffer from significant shortcomings in efficiency and resource consumption. For one thing, the anti-aliasing low-pass filter is placed before the data decimation operation, meaning that a large amount of data that would ultimately be discarded by the decimator still participates in the complex operations of the high-order filter. This not only results in a significant waste of computing resources, especially when the decimation factor is large, but also significantly increases the burden and power consumption of hardware resources such as field-programmable gate arrays (FPGAs). On the other hand, in order to ensure sufficient channel isolation and performance, the order of the prototype low-pass filter usually needs to be designed to be very high, which not only makes the implementation of a single filter difficult and introduces a large processing delay, but also requires the implementation of such a high-order filter in each sub-channel, resulting in very high complexity, resource consumption and processing delay of the overall system, which seriously restricts its application in scenarios with limited resources or high real-time requirements. Summary of the Invention
[0003] The purpose of the present invention is to provide an ultra-wideband low-latency signal processing method based on channelization technology. The method adopts a multi-phase filtering structure and places filtering after decimation, which significantly reduces the amount of calculation and hardware resource consumption, and greatly improves the efficiency of ultra-wideband signal channelization processing; at the same time, it simplifies the filter group design, has a regular structure, is easy to implement and has universal applicability.
[0004] The present invention is achieved through the following technical solutions:
[0005] An ultra-wideband low-latency signal processing method based on channelization technology, the method comprising the following steps:
[0006] Acquire an input digital signal, perform polyphase decomposition on the input digital signal, and obtain K polyphase components;
[0007] Selecting a prototype low-pass filter, and deriving K polyphase branch filters from coefficients of the prototype low-pass filter according to the total number of channels and a preset interpolation factor;
[0008] Performing filtering operations on K corresponding polyphase components through K polyphase branch filters, and performing K-point inverse discrete Fourier transform operations on the obtained K filtered component sets to obtain K transformation results;
[0009] The K transformation results are respectively multiplied by corresponding rotation factors to perform phase correction, and K sub-channel signals after phase correction are output.
[0010] Optionally, the polyphase decomposition of the input digital signal is performed, and the specific calculation formula is:
[0011] x p (m) = x(mD-p)
[0012] Among them, x p (m) is a multiphase component, representing the signal value of the p-th multiphase branch at the time index m after decimation, p is the multiphase branch index, m is the time index after decimation, and D is the preset decimation factor.
[0013] Optionally, the total number of channels, the decimation factor, and the preset interpolation factor are subject to the following constraints:
[0014] K=F·D
[0015] Wherein, K is the total number of channels, and F is the preset interpolation factor.
[0016] Optionally, the relationship between the total number of channels and the decimation factor is expressed as:
[0017] Optionally, the deriving of K polyphase branch filters from coefficients of the prototype low-pass filter according to the total number of channels and a preset interpolation factor further includes:
[0018] Extract the coefficients of the prototype low-pass filter to obtain the intermediate coefficients h p (i) = h(iK + p), where h p (i) is the i-th coefficient of the p-th polyphase branch filter, h() is the coefficient sequence of the prototype filter, and i is the coefficient index after decimation;
[0019] Perform F times zero value interpolation on the intermediate coefficients to obtain the coefficients of the polyphase branch filter, so that the coefficients of the polyphase branch filter are taken when m is an integer multiple of F. otherwise, it is zero.
[0020] Optionally, the filtering operation is performed on the K corresponding multi-phase components respectively through K multi-phase branch filters, and the specific calculation formula is:
[0021] s p (m) = x p (m)*g p(m)
[0022] Among them, g p (m) is a polyphase branch filter, s p (m) is the pth multiphase component x p (m) passes through the corresponding polyphase branch filter g p (m) The output sequence after filtering.
[0023] Optionally, the rotation factor is calculated as follows:
[0024]
[0025] Where j is the imaginary unit, ω k is the center angular frequency of the kth subchannel, and k is the subchannel index.
[0026] Optionally, the center angular frequency of the k-th subchannel is set to:
[0027] Optionally, the output of the K sub-channel signals after phase correction is specifically calculated as follows:
[0028]
[0029] Among them, y k (m) is the signal value of the kth subchannel at the time index m after extraction, IDFT k [s p (m)] is a pair set {s0(m),s1(m),...,s K-1 (m)} is the k-th output component obtained after performing K-point IDFT operation.
[0030] Optionally, the decimation factor D and the total number of channels K satisfy When the relationship is such that for the subchannel with an even index k (k = 0, 2, ..., K-2), after the IDFT operation, the obtained even subchannel data is multiplied by the correction factor (-1) n .
[0031] The technical solution of the present invention has at least the following advantages and beneficial effects:
[0032] On the one hand, the present invention cleverly moves the filtering operation to after polyphase decomposition and data extraction, which not only greatly reduces the number of multiplication and addition operations required and the amount of data processed, but also significantly reduces the demand for high-speed processing hardware and power consumption, thereby significantly improving the overall operational efficiency of the system. On the other hand, the present invention avoids the direct design and implementation of multiple independent high-order filters by deriving all polyphase branch filters from a prototype low-pass filter through coefficient extraction and zero-value interpolation. This not only simplifies and regularizes the design and implementation of the filter bank, facilitating engineering applications, but also has good universality, and its effectiveness is independent of whether the input signal is real or complex, and the specific channel division method. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 A schematic diagram of the process of processing ultra-wideband low-latency signals based on channelization technology provided by the present invention;
[0034] Figure 2 A schematic diagram of the principle of a traditional low-pass filter-based channelization model provided by the present invention;
[0035] Figure 3 A schematic diagram of the principle of a digital channelization model based on a multi-phase structure provided by the present invention;
[0036] Figure 4 This is a schematic diagram of the principle of the 128-channel complex even-type partitioned channelization structure provided by the present invention. DETAILED DESCRIPTION
[0037] To make the objectives, technical solutions, and advantages of the present invention more apparent, the following will provide a clear and complete description of the technical solutions of the present invention in conjunction with the accompanying drawings. It should be understood that the description is only a portion of the present invention, not all of it. The components of the present invention generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations.
[0038] like Figure 1 As shown, the present invention provides one embodiment: an ultra-wideband low-latency signal processing method based on channelization technology, the method comprising the steps of:
[0039] Acquire an input digital signal, perform polyphase decomposition on the input digital signal, and obtain K polyphase components;
[0040] Selecting a prototype low-pass filter, and deriving K polyphase branch filters from coefficients of the prototype low-pass filter according to the total number of channels and a preset interpolation factor;
[0041] Performing filtering operations on K corresponding polyphase components through K polyphase branch filters, and performing K-point inverse discrete Fourier transform operations on the obtained K filtered component sets to obtain K transformation results;
[0042] The K transformation results are respectively multiplied by corresponding rotation factors to perform phase correction, and K sub-channel signals after phase correction are output.
[0043] This embodiment is specifically based on Figure 2 The low-pass filter-based channelization model shown in FIG is improved, as shown in FIG. Figure 2 As shown in the figure, the signal input to the channel is down-converted and low-pass filtered to obtain the baseband signal. The baseband signal is then extracted to obtain the final output. Different channels use different frequency conversion local oscillators to down-convert signals of different frequency bands to baseband. Among them, h(n) is a low-pass filter with a filter order of N and a cutoff frequency of ω0. Channelization is based on the bandwidth of the low-pass filter. This low-pass filter is called the prototype low-pass filter. The output y of channel k is k (m) is:
[0044]
[0045] Although Figure 2 The original receiver digital channelization structure based on digital filter banks solves the defects of traditional digital down-conversion solutions. Although it can also achieve the goal of digital channelization, it has many disadvantages and limitations. On the one hand, the anti-aliasing filter is placed before the decimator, and the discarded data is also involved in the filter operation. The discarded data is far more than the retained data, which requires a lot of FPGA resources and is not conducive to economic efficiency. On the other hand, Figure 2 The filter order in [1] is generally large, making implementation difficult and resulting in significant processing delays. Implementing such a high-order filter for each subchannel is extremely inefficient. Therefore, it is crucial to improve the low-pass filter-based channelization structure and explore more efficient channelization processing implementations.
[0046] Therefore, in order to Figure 2 The structure in the Figure 2 The conclusion derived from the mathematical expression corresponding to the complete digital signal processing process in the structure is universal. This conclusion is universal because it is independent of whether the signal is real or complex, and is also independent of other limiting conditions such as the division method of digital channelization.
[0047] In this embodiment, the derivation process is as follows:
[0048] The frequency band is divided into K channels, and the decimation factor of each channel is D, where K = F·D, and F is a positive integer. The order of the low-pass filter h(n) is N, and L = N / K, where L is the value obtained by dividing the order N of the prototype low-pass filter by the total number of channels K. Let q = iK+p, p = 0, 1, ..., K-1, and i = 0, 1, ..., N-1, then the output of the kth subchannel becomes:
[0049]
[0050] Let x p (m) = x(mD-p), h p (i) = h(iK + p), we can get:
[0051]
[0052] definition:
[0053]
[0054] make g p (m) is h p (i) The result of F-fold interpolation is:
[0055]
[0056] Substitute y k (m) expression is:
[0057]
[0058] According to the derivation, a digital channelization model based on a multi-phase structure can be obtained. Figure 3 As shown. Figure 3 In the example, the new sequence x is obtained by extracting the sequence x(m) by D times. p (m) passes through filter g p (m) filtering, where the filter g p (m) is a polyphase branch filter obtained by extracting K times the coefficient of the prototype low-pass filter and then interpolating F times the zero value. The new sequence obtained after the polyphase filter operation is subjected to K-point IDFT operation and then combined with the rotation factor At this point, the ω obtained in different application scenarios can be multiplied according to the needs. k By substituting this, we can obtain an efficient digital channelization structure corresponding to this application scenario.
[0059] In summary, the digital channelization structure based on the polyphase structure is a special form of digital down-conversion derived from the traditional digital down-conversion transformation. Through the derivation and transformation of mathematical formulas, it changes the anti-aliasing filter from being located before the decimation in the traditional digital down-conversion to being located after the decimation in the polyphase filtering structure, greatly improving the system's operational efficiency.
[0060] In the specific application of this embodiment, the biggest difference between the low-pass filter actually designed and the low-pass filter in the ideal case is the existence of its transition band. For this reason, the signal spectrum will be overlapped after being down-sampled. Specifically, the signal in the transition band will be overlapped into the passband range of the adjacent sub-channel after being down-sampled. In order to avoid this problem, during the channelization process, it is necessary to make the decimation multiple D and the number of channels K satisfy the relationship D≤K / 2. At the same time, in order to reduce the data rate of channel processing as much as possible, the maximum value of D can be taken as D=K / 2=16. The 128-channel complex even-type division channelization structure can be obtained as follows Figure 4 shown.
[0061] The coefficients of the prototype low-pass filter are designed. In order to obtain the coefficients of the polyphase branch filter of 128 sub-channels, the coefficients of the prototype low-pass filter are first decimated by 128 times. It should be noted that since D = K / 2, in order to match the sub-channel data with the sub-channel branch filter coefficients, according to g p From the definition of (m), it can be seen that when K=FD, the coefficients of the polyphase branch filter, in addition to being extracted 128 times by the prototype low-pass filter, also need to be interpolated with 2 times zero value. The final result is the coefficient of the required polyphase branch filter. Each phase filter has 16 coefficients. The result of the polyphase filter operation is subjected to a 128-point IDFT operation, where the IDFT is implemented using the fast algorithm IFFT. In the FPGA, the xfftIP core can be directly called to complete the operation. Finally, after the data of the even sub-channel participates in the polyphase filter and the subsequent discrete Fourier inverse transform operation, it needs to be multiplied by (-1). n That is, it is realized, which greatly reduces the demand for processor DSP resources.
[0062] The above is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.
Claims
1. An ultra-wideband low-latency signal processing method based on channelization technology, characterized in that: The steps of the method include: Acquire an input digital signal, perform polyphase decomposition on the input digital signal, and obtain K polyphase components; Selecting a prototype low-pass filter, and deriving K polyphase branch filters from coefficients of the prototype low-pass filter according to the total number of channels and a preset interpolation factor; Performing filtering operations on K corresponding polyphase components through K polyphase branch filters, and performing K-point inverse discrete Fourier transform operations on the obtained K filtered component sets to obtain K transformation results; The K transformation results are respectively multiplied by corresponding rotation factors to perform phase correction, and K sub-channel signals after phase correction are output.
2. The ultra-wideband low-latency signal processing method based on channelization technology according to claim 1, characterized in that: The specific calculation formula for performing multi-phase decomposition on the input digital signal is: x p (m)=x(mD-p) Among them, x p (m) is a multiphase component, representing the signal value of the p-th multiphase branch at the time index m after decimation, p is the multiphase branch index, m is the time index after decimation, and D is the preset decimation factor.
3. The ultra-wideband low-latency signal processing method based on channelization technology according to claim 2, characterized in that: The total number of channels, decimation factor and preset interpolation factor are subject to the following constraints: K=F·D Wherein, K is the total number of channels, and F is the preset interpolation factor.
4. The method for processing ultra-wideband low-latency signals based on channelization technology according to claim 3, wherein: The relationship between the total number of channels and the decimation factor is expressed as:
5. The ultra-wideband low-latency signal processing method based on channelization technology according to claim 4, characterized in that: The method of deriving K polyphase branch filters from coefficients of the prototype low-pass filter according to the total number of channels and a preset interpolation factor further includes: Extract the coefficients of the prototype low-pass filter to obtain the intermediate coefficients h p (i) = h(iK + p), where h p (i) is the i-th coefficient of the p-th polyphase branch filter, h() is the coefficient sequence of the prototype filter, and i is the coefficient index after decimation; Perform F times zero value interpolation on the intermediate coefficients to obtain the coefficients of the polyphase branch filter, so that the coefficients of the polyphase branch filter are taken when m is an integer multiple of F. otherwise, it is zero.
6. The method for processing ultra-wideband low-latency signals based on channelization technology according to claim 5, wherein: The K corresponding multi-phase components are filtered by the K multi-phase branch filters respectively, and the specific calculation formula is: s p (m)=x p (m)*g p (m) Among them, g p (m) is a polyphase branch filter, s p (m) is the pth multiphase component x p (m) passes through the corresponding polyphase branch filter g p (m) The output sequence after filtering.
7. The method for processing ultra-wideband low-latency signals based on channelization technology according to claim 6, wherein: The specific calculation formula of the rotation factor is: Where j is the imaginary unit, ω k is the center angular frequency of the kth subchannel, and k is the subchannel index.
8. The method for processing ultra-wideband low-latency signals based on channelization technology according to claim 7, wherein: The central angular frequency of the kth subchannel is set to:
9. The ultra-wideband low-latency signal processing method based on channelization technology according to claim 7, characterized in that: The specific calculation formula for outputting the K sub-channel signals after phase correction is: Among them, y k (m) is the signal value of the kth subchannel at the time index m after extraction, IDFT k [s p (m)] is a pair set {s0(m),s1(m),...,s K-1 (m)} is the k-th output component obtained after performing K-point IDFT operation.
10. The ultra-wideband low-latency signal processing method based on channelization technology according to claim 9, characterized in that: The decimation factor D and the total number of channels K satisfy When the relationship is such that for the subchannel with an even index k (k = 0, 2, ..., K-2), after the IDFT operation, the obtained even subchannel data is multiplied by the correction factor (-1) n .