Signal data processing method and device, equipment, storage medium and computer program product

Abstract

The invention relates to the technical field of digital signal processing, and discloses a signal data processing method and device, equipment, a storage medium and a computer program product, and the method comprises the steps: obtaining signal data in a target channel, dividing the signal data according to a target quantization bit, and obtaining to-be-quantized subsets; setting a corresponding uniform quantization mapping formula for each subset to be quantized, and performing uniform quantization on the subset to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; and carrying out inverse quantization on the discrete value set to obtain processed signal data. According to the method, the signals are divided into the target number of groups according to the distribution characteristics of the signals, the uniform quantization mapping formula matched with the characteristics of each sub-group is designed for each sub-group, and the whole data mapping is non-uniform quantization, so that the flexibility and adaptability of quantization are improved, the calculation complexity and hardware resource requirements are effectively reduced, and the data mapping efficiency is improved. And the equalization performance and the system transmission quality are obviously improved.

Description

Technical Field

[0001] The present application relates to the technical field of digital signal processing, and in particular to a signal data processing method, device, equipment, storage medium and computer program product. Background Art

[0002] In view of the dynamics and complexity of signals in IM / DD (Intensity modulation and direct detection) systems, the automatic grouping non-uniform quantization method can group signals according to their distribution characteristics and assign quantization mapping functions to different groups. However, existing methods usually rely on manual settings or prior knowledge, with a low degree of automation, and it is difficult to meet the needs of real-time and efficient processing. Summary of the invention

[0003] The main purpose of the present application is to provide a signal data processing method, device, equipment, storage medium and computer program product, aiming to solve the technical problem that the existing non-uniform quantization method usually relies on manual settings or prior knowledge, has a low degree of automation, and is difficult to meet the needs of real-time and efficient processing.

[0004] To achieve the above object, the present application proposes a signal data processing method, the signal data processing method comprising: Acquire signal data in a target channel, and divide the signal data according to a target quantization bit number to obtain a subset to be quantized; Setting a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and uniformly quantizing the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; Dequantizing the discrete value set to obtain processed signal data.

[0005] Optionally, the step of setting a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and uniformly quantizing the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set includes: Generate a step size set and an offset set based on the subset to be quantized; Determine a uniform quantization mapping formula corresponding to the subset to be quantized according to the step size set, the offset set and the target number of quantization bits; The subset to be quantized is uniformly quantized according to the uniform quantization mapping formula to obtain a quantized discrete value set.

[0006] Optionally, the step of generating a step size set and an offset set based on the subset to be quantized includes: Counting the data range of the signal data in each of the subsets to be quantized, wherein the data range includes the maximum value and the minimum value of the signal data; Calculate the step size of the subset to be quantized based on the target number of quantization bits and the data range to obtain a step size set; The offset of the subset to be quantized is calculated according to the target number of quantization bits and the minimum value of the signal data to obtain an offset set.

[0007] Optionally, the step of acquiring signal data in a target channel and dividing the signal data according to a target quantization bit number to obtain a subset to be quantized includes: Acquire signal data in a target channel, and determine a data set to be quantized and a target number of quantization bits based on the signal data; Sorting the signal data in the data set to be quantified by a preset sorting algorithm to obtain a target data set; The target data set is divided according to the target quantization bit number, and each subset to be quantized after the division is determined.

[0008] Optionally, the step of determining the target number of quantization bits and the data set to be quantized based on the received real-time signal includes: Perform statistics on the received real-time signals at a preset time period to determine the data set to be quantified; A target number of quantization bits is generated according to the characteristics of the signal data in the data set to be quantized.

[0009] Optionally, after the step of dequantizing the discrete value set to obtain processed signal data, the method further includes: Determining a quantization error index of the processed signal data; If the quantization error index is greater than the preset error range threshold, the target quantization bit number is adjusted, and the step of dividing the signal data according to the target quantization bit number to obtain the subset to be quantized is returned until the quantization error index is less than the preset error range threshold.

[0010] In addition, to achieve the above-mentioned purpose, the present application also proposes a signal data processing device, the signal data processing device comprising: A data partitioning module is used to obtain signal data in a target channel and partition the signal data according to a target quantization bit number to obtain a subset to be quantized; A data quantization module, used to set a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and to uniformly quantize the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; The data dequantization module is used to dequantize the discrete value set to obtain processed signal data.

[0011] In addition, to achieve the above-mentioned purpose, the present application also proposes a signal data processing device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the signal data processing method described above.

[0012] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the signal data processing method described above are implemented.

[0013] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the signal data processing method described above are implemented.

[0014] The present application discloses the method of obtaining signal data in a target channel, dividing the signal data according to a target quantization bit number to obtain a subset to be quantized; setting a corresponding uniform quantization mapping formula for each subset to be quantized, and uniformly quantizing the subset to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; and dequantizing the discrete value set to obtain processed signal data. According to the distribution characteristics of the signal, the signal is divided into a target number of groups, and a uniform quantization mapping formula matching its characteristics is designed for each subgroup. The entire data mapping is non-uniform quantization, which improves the flexibility and adaptability of quantization, effectively reduces the computational complexity and hardware resource requirements, and significantly improves the equalization performance and system transmission quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0017] Figure 1 This is a flow chart of the first embodiment of the signal data processing method of the present application; Figure 2 A diagram of the test environment for the signal data processing method of this application; Figure 3This is a flow chart of the second embodiment of the signal data processing method of the present application; Figure 4 This is a non-uniform quantization flow chart of automatic grouping of the signal data processing method of this application; Figure 5 This is a flow chart of a third embodiment of the signal data processing method of the present application; Figure 6 This is a schematic diagram of the module structure of the signal data processing device according to an embodiment of the present application; Figure 7 Schematic diagram of the device structure of the hardware operating environment involved in the signal data processing method in the embodiment of the present application.

[0018] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0020] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0021] The main solution of the embodiment of the present application is: obtain signal data in the target channel, and divide the signal data according to the target quantization bit number to obtain subsets to be quantized; set a corresponding uniform quantization mapping formula for each subset to be quantized, and uniformly quantize the subset to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; dequantize the discrete value set to obtain processed signal data.

[0022] Short-distance optical interconnection is an indispensable key technology in modern data centers, high-performance computing, and high-speed access networks. With the rapid development of cloud computing, big data, and artificial intelligence, data traffic has exploded, and the demand for high-bandwidth, low-latency, and low-power communications has continued to increase. Short-distance optical interconnection uses optical communication technology to achieve high-speed data transmission. Compared with electronic interconnection, it has higher bandwidth density, lower transmission loss, and lower power consumption, which can significantly improve system performance. Especially in the internal interconnection of data centers, optical communication provides an ideal solution for large-scale parallel data exchange between switches, servers, and storage devices. With the rapid development of information technology, optical communication has become the core technology of modern communication systems. IM / DD systems have become the mainstream solution for short-distance optical communication due to their simple structure, low cost, and low power consumption. However, when high-bandwidth signals are transmitted in IM / DD systems, they often face problems such as channel nonlinear distortion, group delay, bandwidth limitation, and noise, which poses a great challenge to system performance. To address these problems, digital signal processing technology has been widely used, especially equalization algorithms, which have outstanding performance in suppressing nonlinear distortion. Among them, Volterra nonlinear equalizer, as a classic method, can effectively improve system performance. However, its computational complexity increases exponentially with the increase of nonlinear order and signal bandwidth, which makes hardware implementation face severe challenges.

[0023] In order to reduce the complexity of digital signal processing algorithms, quantization technology has become a key means. By reducing the data bit width, the amount of calculation and hardware power consumption can be significantly reduced. However, traditional uniform quantization methods are difficult to quantize according to the characteristics of data distribution. In view of the dynamics and complexity of signals in IM / DD systems, the automatic grouping non-uniform quantization method can group signals according to their distribution characteristics and assign quantization mapping functions to different groups, showing great potential in balancing quantization accuracy and complexity. However, existing methods usually rely on manual settings or prior knowledge, with a low degree of automation, and are difficult to meet the needs of real-time and efficient processing.

[0024] Therefore, the present application provides a signal data processing method for non-uniform quantization with automatic grouping, which realizes effective compensation for nonlinear distortion by combining signal characteristic analysis and dynamic optimization design. By designing an automatic grouping method, the size distribution characteristics of the input signal are analyzed, and the signal is divided into groups of the target number of bits according to the size of the data to be quantized. Subsequently, a uniform quantization mapping formula matching its characteristics is designed for each subgroup, and the entire data mapping is non-uniform quantization. After completing the signal grouping and quantization, a lightweight nonlinear equalizer is designed to quantize the tap coefficients in the equalizer to a lower bit. The implementation of the equalizer is combined with an automatic grouping strategy to further reduce the computational complexity and hardware power consumption by reducing the total number of bits occupied by the tap coefficients. It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a computer, or an electronic device capable of realizing the above functions. The following takes a nonlinear equalizer as an example to illustrate this embodiment and the following embodiments.

[0025] Based on this, the present application embodiment provides a signal data processing method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the signal data processing method of the present application.

[0026] In this embodiment, the signal data processing method includes: Step S10, acquiring signal data in a target channel, and dividing the signal data according to a target quantization bit number to obtain a subset to be quantized.

[0027] It should be noted that the target channel refers to a specific channel that requires signal processing in a specific communication or data transmission scenario. It can be a frequency band channel in wireless communication or a transmission link in wired communication (such as optical fiber, cable). Signal data is the digital data corresponding to the electrical signal, optical signal or other form of signal that contains specific information transmitted in the target channel. These data can represent various types of information such as sound, images, text, etc., and will be affected by factors such as noise and attenuation during the transmission process. Therefore, quantization and other processing are required to optimize data storage and transmission.

[0028] It can be understood that the target quantization bit number determines the number of discrete values ​​that the quantized signal can represent. For example, 8-bit quantization can represent 2 8 = 256 different discrete values. The higher the number of bits, the higher the quantization accuracy, but the greater the amount of calculation and storage space required; the lower the number of bits, the lower the quantization accuracy, but the calculation and storage costs are also reduced accordingly. The subsets to be quantized are multiple subset data sets obtained by dividing the signal data in the target channel according to certain rules (such as numerical value, data distribution, etc.) according to the target number of quantization bits. The data in each subset has similar characteristics or distribution patterns. These subsets will be quantized separately in the future to better adapt to the characteristics of the data and improve the effect and efficiency of quantization.

[0029] Specifically, the division rule can be determined based on the target number of quantization bits and the characteristics of the signal data (such as data range, distribution law, etc.). If the target number of quantization bits is n, the value range of the signal data can be divided into 2 nintervals; if the data is unevenly distributed, cluster analysis and other methods can be used to divide the data according to the density and distribution characteristics, so that the data in each interval has similar statistical characteristics. Then, according to the determined division rules, the signal data is allocated to different subsets. If divided by data range, the signal data needs to be sorted from small to large, and the data is placed in the corresponding intervals in turn to form the subset to be quantified.

[0030] Step S20, setting a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and performing uniform quantization on the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set.

[0031] It should be noted that the uniform quantization mapping formula is used to map the signal data in the subset to be quantized to a finite number of discrete quantization values, and through specific quantization parameters, ensure that the signal data in the subset is quantized according to a uniform rule.

[0032] It can be understood that the parameters in the uniform quantization mapping formula corresponding to different subsets to be quantized are different. For each subset to be quantized, it is necessary to determine the key parameters in the uniform quantization mapping formula, such as the quantization step size and offset, and construct a specific uniform quantization mapping formula based on the determined quantization step size and offset.

[0033] It should be understood that after uniform quantization, the originally continuous signal data in the subset to be quantized is converted into a set of discrete values, and the set of these discrete values ​​is the discrete value set after quantization. The number of elements in the set is determined by the number of quantization bits, and each element is the result of the quantization mapping formula acting on the original data.

[0034] In one example, in a uniform quantized Volterra equalizer, when the tap coefficients are uniformly quantized, the floating point tap coefficients R are first passed into the uniform quantization mapping formula to obtain the quantized low-bit integers, where Represents the quantized low-bit integer, ranging from Represents the number of bits to be converted. The uniform quantization mapping formula is as follows:

[0035] Among them, S is the step size, Z is the offset, and the calculation formula is as follows:

[0036]

[0037] in, Represents the maximum value of the previously calculated quantized low-bit integers, The function represents the nearest integer operation, which is the key step in converting floating-point numbers into integers. and Represents the maximum and minimum values ​​of floating-point tap coefficients The initial value of the setting is 0. The meaning of the clamp function is as follows:

[0038] Among them, w, a, c are demonstration parameters.

[0039] Finally, the norm quantization operation can be performed as follows:

[0040] Among them, the tap coefficients after dequantization Due to rounding, Not an exact match.

[0041] Step S30, dequantizing the discrete value set to obtain processed signal data.

[0042] It can be understood that, contrary to the quantization process, dequantization is the process of restoring the quantized discrete values ​​to be closer to the original signal data. Through a specific calculation method, the discrete values ​​are converted back to numerical values ​​with a certain accuracy to restore the original characteristics of the signal or as close to the original state as possible. The processed signal data is the data obtained after the dequantization operation. These data retain the information of the original signal to a certain extent, but due to the inevitable errors in the quantization and dequantization process, there may be some differences between the processed signal data and the original signal data.

[0043] It should be understood that the dequantization operation needs to be based on the parameters used in the previous quantization, such as the quantization step size and offset. These parameters determine how the original data is mapped to discrete values ​​during the quantization process, and are used to restore the discrete values ​​during the dequantization process.

[0044] In one example, when non-uniform quantization is used, different quantization intervals have different step sizes and offsets. and are the step size and offset of the i-th subset to be quantized, which need to be calculated in different sub-data. In the case of non-uniform quantization, the formula for the inverse quantization operation is:

[0045] in, is the final quantized value of each data point in the i-th subset to be quantized. is the inverse quantized data point. Each subset uses independent quantization rules to form non-uniform quantization characteristics, making the overall data distribution more consistent with the signal characteristics. By allocating different step sizes and offsets, regional optimization of quantization accuracy is achieved, while significantly reducing the computational complexity of data processing.

[0046] Furthermore, in order to continuously optimize the quantization results in practical applications, ensure the accuracy and reliability of the final quantization data, and reduce the problem of excessive errors caused by improper initial quantization parameter settings, after step S30, the following is further included: Determining a quantization error index of the processed signal data; If the quantization error index is greater than the preset error range threshold, the target quantization bit number is adjusted, and the step of dividing the signal data according to the target quantization bit number to obtain the subset to be quantized is returned until the quantization error index is less than the preset error range threshold.

[0047] It should be noted that the quantization error index is a parameter used to quantitatively describe the size of the quantization error between the processed signal data and the original signal data, such as mean square error, peak signal-to-noise ratio, etc. The preset error range threshold is a quantization error limit value pre-set according to specific application scenarios and requirements. When the quantization error index is less than the threshold, it is considered that the signal data after quantization processing is within an acceptable error range and can meet the actual application requirements; conversely, when the quantization error index is greater than the threshold, it indicates that the quantization effect is not ideal and the quantization process needs to be adjusted.

[0048] Specifically, the processed signal data is compared with the original signal data to calculate the quantization error index value of the current processed signal data. When the quantization error index is greater than the preset error range threshold, the target quantization bit number needs to be adjusted. If the error is too large, the target quantization bit number will be appropriately increased to improve the quantization accuracy and reduce the quantization error; if the quantization error index is close to the preset error range threshold, the target quantization bit number can also be fine-tuned to observe the change in the quantization error.

[0049] In one example, reference Figure 2 , Figure 2This is a test environment diagram for the signal data processing method of this application. The figure includes the digital signal processing (DSP) process and related components of the transmitter (Tx) and the receiver (Rx). The physical equipment in the test environment includes: AWG (Arbitrary Waveform Generator): an arbitrary waveform generator with a sampling rate of 224 GSa / s, used to generate signals of a specific waveform, used to generate original signals at the transmitter; DML (Directly Modulated Laser): directly modulated laser, converting electrical signals into optical signals for transmission; 10-km SSMF (10 km standard single-mode optical fiber): transmits optical signals, but will be affected by factors such as dispersion; VOA (Variable Optical Attenuator): a variable optical attenuator used to adjust the intensity of optical signals; 60-GHz PD (60GHz photodetector): converts optical signals back to electrical signals; RTO (Real-Time Oscilloscope): a real-time oscilloscope with a sampling rate of 256 GSa / s, used for observation and analysis of signals at the receiving end.

[0050] The transmitter (Tx DSP) process includes: 1. Bit Generation: bit generation, generating original binary data; 2. PAM-4 / 6 Modulation: pulse amplitude modulation, select PAM-4 (4 levels) or PAM-6 (6 levels) modulation mode, and map the bit data to different amplitude levels; 3. Up-sampling: up-sampling, increasing the sampling rate of the signal to prepare for subsequent processing; 4. RRC Filter: root raised cosine filter, used to filter the signal, reduce inter-symbol interference, and make the signal spectrum meet the transmission requirements; 5. Resampling to 224 GSa / s: resampling to 224 GSa / s to match the sampling rate of the AWG.

[0051] The receiving end (Rx DSP) process includes: 1. Resampling & Matched filtering: resampling and matched filtering, resampling the received signal, and enhancing the useful signal through the matched filter, suppressing noise and interference; 2. Frame synchronization: frame synchronization, used to determine the starting position of the signal frame in order to correctly receive and decode the data; 3. Quantized Volterra Equalization: quantized Volterra equalization, equalizing the signal and compensating for channel distortion; 4. Down-sampling and GD-NW: down-sampling and group delay equalization, reducing the sampling rate, and performing group delay equalization to further optimize the signal; 5. PAM-4 / 6 DeMod and BER count: PAM-4 / 6 demodulation, and performing bit error rate (BER) statistics after demodulation, demodulating the received signal back to the original bit data, and evaluating the performance of the communication system.

[0052] In experimental tests, this application successfully carried out 50-70 GBd 5 / 10 km pulse amplitude modulation (PAM)-4 and 50-60 GBd 10 km PAM-6 C-band IM / DD transmission, and compared this application with the uniform quantization method. The experimental results show that this application reduces the tap coefficients of the equalizer to a lower bit amount to reduce the total storage space without much loss in communication quality. In the case of PAM4@5km transmission, the effective quantization bit number is as low as 7 bits, which can save up to 75.6% of the storage space, which is 22.8% higher than the optimal storage of traditional uniform quantization. In the case of PAM6@10km transmission, the storage efficiency of this application is improved by 24.0% compared with traditional uniform quantization.

[0053] In this embodiment, the signal data in the target channel is obtained, and the signal data is divided according to the target quantization bit number to obtain the subsets to be quantized; a corresponding uniform quantization mapping formula is set for each of the subsets to be quantized, and the subsets to be quantized are uniformly quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; the discrete value set is inversely quantized to obtain the processed signal data. According to the distribution characteristics of the signal, the signal is divided into the target number of groups, and a uniform quantization mapping formula matching its characteristics is designed for each subgroup. The entire data mapping is non-uniform quantization, which improves the flexibility and adaptability of quantization, effectively reduces the computational complexity and hardware resource requirements, and significantly improves the equalization performance and system transmission quality.

[0054] Reference Figure 3 , Figure 3This is a flow chart of the second embodiment of the signal data processing method of the present application. Based on the above-mentioned first embodiment, the second embodiment of the signal data processing method of the present application is proposed.

[0055] In the second embodiment, the step S10 includes: Step S101: acquiring signal data in a target channel, and determining a data set to be quantized and a target quantization bit number based on the signal data.

[0056] It should be noted that the data set to be quantified is a part of the data set selected from the acquired signal data. It can be screened from the original signal data according to specific needs or conditions, or it can select signal data in a specific frequency band as the data set to be quantified based on the frequency range of the signal, which is used to quantify and subsequently process the signal characteristics of the frequency band.

[0057] Furthermore, in order to divide the data in an orderly manner according to the target quantization bit number and ensure that the division of each subset to be quantized is more reasonable, the step S101 may include: The received real-time signal is counted with a preset time period as a cycle to determine the data set to be quantized; and the target quantization bit number is generated according to the characteristics of the signal data in the data set to be quantized.

[0058] It should be noted that the preset time period refers to a fixed time interval that is preset in advance and is used to periodically process and analyze real-time signals. The preset time period can be determined according to specific application scenarios and requirements.

[0059] It is understandable that when the received real-time signals are counted with a preset time period as a cycle, the real-time signals are first obtained through the corresponding receiving devices (such as sensors, antennas, etc.). These signals may be analog signals, which need to be converted into digital signals through preprocessing steps such as analog-to-digital conversion (ADC). The received real-time digital signals are counted with a preset time period as a cycle. The statistical content may include various characteristic parameters such as the mean, variance, maximum value, minimum value, frequency distribution, etc. of the signal. For example, in each preset time period, the mean value, standard deviation and other statistics of the signal are calculated, and these statistics can reflect the overall characteristics of the signal in the time period.

[0060] It should be understood that the selection of the data set to be quantified based on statistical results can be when the signal fluctuation range is small, all the signal data in the time period is used as the data set to be quantified; it can also be when the signal has abnormal values ​​or special change patterns in certain time periods, and some data is selected as the data set to be quantified according to needs and characteristics. When generating the target quantization bit number, it is necessary to conduct a detailed analysis of the signal data in the data set to be quantified, including the dynamic range of the data (the difference between the maximum and minimum values), the distribution of the data (such as whether it is evenly distributed, normally distributed, etc.), the frequency components of the data, etc.

[0061] Step S102 , sorting the signal data in the data set to be quantized by using a preset sorting algorithm to obtain a target data set.

[0062] It should be noted that the preset sorting algorithm refers to a specific algorithm that is determined in advance for sorting data, such as quick sort, merge sort, and selection sort. When sorting, you can also select the desired sorting algorithm according to the characteristics of the data set. The target data set is the data set obtained after the preset sorting algorithm sorts the quantization data set.

[0063] Step S103 , dividing the target data set according to the target quantization bit number, and determining each subset to be quantized after the division.

[0064] It should be understood that the target quantization bit number corresponds to the number of subsets to be quantized after division. When grouping, the last subset may contain fewer elements than the previous subsets due to uneven data distribution.

[0065] It is understandable that after determining the division, each subset to be quantized can also use a deep learning model (such as an autoencoder or a variational autoencoder) to automatically learn the signal characteristics to optimize the grouping rules, and use an adaptive quantization algorithm to adjust the quantization level allocation in real time.

[0066] In one example, reference Figure 4 , Figure 4 This is a flowchart of non-uniform quantization of automatic grouping of the signal data processing method of the present application, in which { x} represents the original data sequence, which contains a series of data points, such as Etc. First, sort the data values ​​x that need to be quantized, arranging the data from small to large to ensure that the data can be more effectively allocated to different subsets in subsequent grouping. Then, divide the sorted data evenly into multiple subsets, the number of subsets corresponds to the target quantization bit b, and obtain the divided subsets to be quantized:

[0067] in, is the subset to be quantized. Perform group quantization on each subset and map it to a set of discrete values ​​within the quantization range: , Finally, de-quantization operations are performed one by one:

[0068] in, is the final quantized value of each data point in the i-th subset to be quantized. is the dequantized data point. and are the step size and offset of the i-th subset to be quantized respectively.

[0069] In this embodiment, the signal data in the target channel is obtained, and the data set to be quantized and the target quantization bit number are determined based on the signal data; the signal data in the data set to be quantized is sorted by a preset sorting algorithm to obtain the target data set; the target data set is divided according to the target quantization bit number, and each subset to be quantized after the division is determined. According to the different characteristics of the signal data in different time periods, a target quantization bit number that can adapt to the dynamic changes of the signal is generated, avoiding the performance degradation caused by the fixed quantization bit number, and improving the quantization accuracy and processing efficiency of the signal data.

[0070] Reference Figure 5 , Figure 5 This is a flow chart of the third embodiment of the signal data processing method of the present application. Based on the above second embodiment, the third embodiment of the signal data processing method of the present application is proposed.

[0071] In the third embodiment, the step S20 includes: Step S201: Generate a step size set and an offset set based on the subset to be quantized.

[0072] It can be understood that by using a step set and an offset set to perform quantization mapping on each subset, non-uniform quantization mapping based on signal statistical characteristics can be achieved. When each subset is uniformly quantized to a fixed number of quantization points, a step set and offset of a specific number of bits can be generated.

[0073] Furthermore, in order to optimize according to their own characteristics in subsets with different data ranges, quantization deviation caused by unified quantization is avoided, which helps to improve quantization effect and resource utilization efficiency. The step S201 may include: The method comprises the following steps: counting the data range of the signal data in each of the subsets to be quantized, wherein the data range includes the maximum value and the minimum value of the signal data; calculating the step size of the subset to be quantized based on the target number of quantization bits and the data range to obtain a step size set; and calculating the offset of the subset to be quantized according to the target number of quantization bits and the minimum value of the signal data to obtain an offset set.

[0074] It can be understood that, for each subset to be quantized, the step size can be calculated according to its data range and the target number of quantization bits, and the offset can be calculated based on the target number of quantization bits and the minimum value of the subset signal data.

[0075] In one example, for each data point, a set of steps of a certain number of bits is generated and offset The calculation formula is as follows:

[0076]

[0077] in, represents the subset to be quantized, and b is the target number of quantization bits.

[0078] Step S202: determining a uniform quantization mapping formula corresponding to the subset to be quantized according to the step size set, the offset set and the target number of quantization bits.

[0079] It can be understood that by dynamically allocating different quantization mapping formulas to each subgroup according to the grouping result, a lower bit quantization effect can be achieved.

[0080] In one example, for each data point, the final quantization value mapping formula is:

[0081] Among them, s i and z i The calculations are performed in different sub-data sets, and each subset adopts independent quantization rules, thus forming non-uniform quantization characteristics, making the overall data distribution more consistent with the signal characteristics.

[0082] Step S203 , uniformly quantizing the subset to be quantized according to the uniform quantization mapping formula to obtain a quantized discrete value set.

[0083] In this embodiment, a step set and an offset set are generated based on the subset to be quantized; a uniform quantization mapping formula corresponding to the subset to be quantized is determined according to the step set, the offset set and the target quantization bit number; the subset to be quantized is uniformly quantized according to the uniform quantization mapping formula to obtain a quantized discrete value set. By setting independent quantization rules for each subset, a non-uniform quantization characteristic is formed, so that the overall data distribution is more consistent with the signal characteristics.

[0084] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the signal data processing method of the present application. More simple transformations based on this technical concept are all within the protection scope of the present application.

[0085] The present application also provides a signal data processing device, please refer to Figure 6 , the signal data processing device comprises: The data partitioning module 10 is used to obtain signal data in the target channel and partition the signal data according to the target quantization bit number to obtain a subset to be quantized; A data quantization module 20, configured to set a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and to uniformly quantize the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; The data dequantization module 30 is used to dequantize the discrete value set to obtain processed signal data.

[0086] The signal data processing device provided by the present application adopts the signal data processing method in the above embodiment, which can solve the technical problem that the existing non-uniform quantization method usually relies on manual settings or prior knowledge, has a low degree of automation, and is difficult to meet the needs of real-time and efficient processing. Compared with the prior art, the beneficial effects of the signal data processing device provided by the present application are the same as the beneficial effects of the signal data processing method provided by the above embodiment, and the other technical features in the signal data processing device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0087] The present application provides a signal data processing device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the signal data processing method in the above-mentioned embodiment 1.

[0088] Reference below Figure 7, which shows a schematic diagram of the structure of a signal data processing device suitable for implementing an embodiment of the present application. The signal data processing device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The signal data processing device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0089] like Figure 7 As shown, the signal data processing device may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the signal data processing device are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the signal data processing device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a signal data processing device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0090] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0091] The signal data processing device provided by the present application adopts the signal data processing method in the above embodiment, which can solve the technical problem that the existing non-uniform quantization method usually relies on manual settings or prior knowledge, has a low degree of automation, and is difficult to meet the needs of real-time and efficient processing. Compared with the prior art, the beneficial effects of the signal data processing device provided by the present application are the same as the beneficial effects of the signal data processing method provided by the above embodiment, and the other technical features in the signal data processing device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0092] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0093] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0094] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the signal data processing method in the above-mentioned embodiment.

[0095] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0096] The computer-readable storage medium may be included in the signal data processing device; or may exist independently without being installed in the signal data processing device.

[0097] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the signal data processing device, the signal data processing device executes the signal data processing method described above.

[0098] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0099] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0100] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0101] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned signal data processing method, and can solve the technical problem that the existing non-uniform quantization method usually relies on manual settings or prior knowledge, has a low degree of automation, and is difficult to meet the needs of real-time and efficient processing. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the signal data processing method provided by the above-mentioned embodiment, and will not be repeated here.

[0102] The present application also provides a computer program product, including a computer program, which implements the steps of the signal data processing method as described above when executed by a processor.

[0103] The computer program product provided by the present application can solve the technical problem that the existing non-uniform quantization method usually relies on manual settings or prior knowledge, has a low degree of automation, and is difficult to meet the needs of real-time and efficient processing. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the signal data processing method provided by the above embodiment, and will not be repeated here.

[0104] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A signal data processing method, characterized in that: The signal data processing method comprises: Acquire signal data in a target channel, and divide the signal data according to a target quantization bit number to obtain a subset to be quantized; Setting a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and uniformly quantizing the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; Dequantizing the discrete value set to obtain processed signal data.

2. The signal data processing method according to claim 1, characterized in that: The step of setting a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and uniformly quantizing the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set includes: Generate a step size set and an offset set based on the subset to be quantized; Determine a uniform quantization mapping formula corresponding to the subset to be quantized according to the step size set, the offset set and the target number of quantization bits; The subset to be quantized is uniformly quantized according to the uniform quantization mapping formula to obtain a quantized discrete value set.

3. The signal data processing method according to claim 2, characterized in that: The step of generating a step size set and an offset set based on the subset to be quantized comprises: Counting the data range of the signal data in each of the subsets to be quantized, wherein the data range includes the maximum value and the minimum value of the signal data; Calculate the step size of the subset to be quantized based on the target number of quantization bits and the data range to obtain a step size set; The offset of the subset to be quantized is calculated according to the target number of quantization bits and the minimum value of the signal data to obtain an offset set.

4. The signal data processing method according to claim 1, characterized in that: The step of acquiring signal data in the target channel and dividing the signal data according to the target quantization bit number to obtain a subset to be quantized includes: Acquire signal data in a target channel, and determine a data set to be quantized and a target number of quantization bits based on the signal data; Sorting the signal data in the data set to be quantified by a preset sorting algorithm to obtain a target data set; The target data set is divided according to the target quantization bit number, and each subset to be quantized after the division is determined.

5. The signal data processing method according to claim 4, characterized in that: The step of determining the target quantization bit number and the data set to be quantized based on the received real-time signal includes: Perform statistics on the received real-time signals at a preset time period to determine the data set to be quantified; A target number of quantization bits is generated according to the characteristics of the signal data in the data set to be quantized.

6. The signal data processing method according to claim 1, characterized in that: After the step of dequantizing the discrete value set to obtain processed signal data, the method further includes: Determining a quantization error index of the processed signal data; If the quantization error index is greater than the preset error range threshold, the target quantization bit number is adjusted, and the step of dividing the signal data according to the target quantization bit number to obtain the subset to be quantized is returned until the quantization error index is less than the preset error range threshold.

7. A signal data processing device, characterized in that: The device comprises: A data partitioning module is used to obtain signal data in a target channel and partition the signal data according to a target quantization bit number to obtain a subset to be quantized; A data quantization module, used to set a corresponding uniform quantization mapping formula for each of the subsets to be quantized, and to uniformly quantize the subsets to be quantized based on the uniform quantization mapping formula to obtain a quantized discrete value set; The data dequantization module is used to dequantize the discrete value set to obtain processed signal data.

8. A signal data processing device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the signal data processing method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the signal data processing method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the signal data processing method according to any one of claims 1 to 6 are implemented.

Images