CN-117420517-B - TIADC mismatch error digital calibration method and system applied to radar data acquisition system

CN117420517BCN 117420517 BCN117420517 BCN 117420517BCN-117420517-B

Abstract

The invention provides a TIADC mismatch error digital calibration method and system applied to a radar data acquisition system, which belong to the technical field of signal sampling and processing, construct a residual sequence containing all mismatch errors, since the frequency of the mismatch error is limited and computable and the information in the mismatch is fixed, complex signals containing the mismatch information can be decomposed into sub-signals with different information granularity using a signal adaptive decomposition method. The invention solves the error problem generated by bias mismatch, gain mismatch and time mismatch in TIADC, further improves the performance index of the data acquisition system and enhances the accuracy of data acquisition.

Inventors

LIU DATONG
YU RUNZE
PAN DAWEI
PENG XIYUAN

Assignees

哈尔滨工业大学

Dates

Publication Date: 20260505
Application Date: 20230925

Claims (7)

1. A digital calibration method for TIADC mismatch errors applied to a radar data acquisition system is characterized by comprising the following steps of: The method specifically comprises the following steps: Step 1, selecting a pair of input and output data of a radar as a tag model to train in an off-line state, calculating ideal sampling output of the radar signal according to a mathematical formula of the input analog signal, or directly connecting a high-end digital oscilloscope to a signal generator to acquire a desired signal, and constructing a residual sequence containing all mismatch errors by combining actual output data of TIADC; In step 1, the problem of TIADC mismatch is solved from a time series perspective, and a residual sequence containing all mismatch is constructed: (1) In the middle of , In order to construct a residual signal, The data is output for the TIADC system, Is the desired output signal and does not contain any mismatch error; step2, carrying out standardization processing on the residual sequences containing all mismatch errors in the step 1, so that the characteristic scales of the data are similar; Step 3, decomposing the standardized residual signals constructed in the step 2 into sub-signals with different information granularities by using a signal self-adaptive decomposition method, and training by using an LSTM model to obtain the time sequence data correlation in the signals; in step 3, the signal adaptive decomposition method is VMD decomposition of normalized residual sequence Decomposing to obtain multiple IMF components as VMD ~ N is the number of VMD decompositions; Training using LSTM model, wherein the mapping function of LSTM regression network is as follows Thus: (3) In the formula, The estimated values obtained for the model training are, Is after reconstruction Each output value Corresponding to a group of corresponding The trained LSTM model is subsequently applied to the calibration stage; Step 4, acquiring real-time radar data in an on-line state, initializing to obtain residual data for estimation, performing standardization processing, performing self-adaptive decomposition on the residual data, and constructing an LSTM model input; And 5, carrying out residual estimation, reconstruction and inverse standardization on the input data in the step 4 by using the model trained in the step 3, calculating to obtain calibrated TIADC output data by combining the acquired real-time data stream of the radar to be processed, and verifying the accuracy of digital calibration of the TIADC mismatch error.
2. The method of calibrating according to claim 1, wherein: in step2, the constructed residual data is subjected to normalization processing: (2) In the middle of Is the residual error data after standardization; as the original residual data is to be obtained, Is the maximum value of the raw data and, Is the minimum value of the original data, and the constructed standardized residual signal provides input for subsequent signal decomposition and model training work.
3. The calibration method according to claim 2, characterized in that: In step 4, the initial residual data is , The preprocessing operation is needed to meet the input condition of the calibration model, and the VMD decomposition is carried out on the residual data, so that the LSTM model is built to be input The estimate is generated by the following model: (4) In the formula, For the estimated value of the model output, Is the residual signal used for the test.
4. A calibration method according to claim 3, characterized in that: In step 5 Reconstructing to obtain a normalized residual sequence The residual error estimation result is obtained after the inverse normalization By measuring signals in real time And a new residual signal And calculating to obtain a calibrated TIADC signal: (5) Finally, digital calibration of the TIADC mismatch error is completed.
5. A TIADC mismatch error digital calibration system applied to a radar data acquisition system is characterized in that the system is used for executing the TIADC mismatch error digital calibration method applied to the radar data acquisition system according to any one of claims 1 to 4; The system comprises an offline training module and a calibration module; The off-line training module selects a pair of input and output data of the radar as a tag model to train in an off-line state, calculates ideal sampling output of the radar signal according to a mathematical formula of the input analog signal, or directly connects a high-end digital oscilloscope to a signal generator to acquire an expected signal, and combines actual output data of TIADC to construct a residual sequence containing all mismatch errors; The problem of mismatch of TIADC is solved from the time sequence angle, and a residual sequence containing all mismatch is constructed: (1) In the middle of , In order to construct a residual signal, The data is output for the TIADC system, Is the desired output signal and does not contain any mismatch error; Carrying out standardization processing on residual sequences containing all mismatch errors so that the characteristic scales of the data are similar; decomposing the constructed standardized residual signals into sub-signals with different information granularities by using a signal self-adaptive decomposition method, and training by using an LSTM model to obtain the time sequence data correlation in the signals; The signal self-adaptive decomposition method is VMD decomposition, namely normalized residual sequence Decomposing to obtain multiple IMF components as VMD ~ N is the number of VMD decompositions; Training using LSTM model, wherein the mapping function of LSTM regression network is as follows Thus: (3) In the formula, The estimated values obtained for the model training are, Is after reconstruction Each output value Corresponding to a group of corresponding The trained LSTM model is subsequently applied to the calibration stage; The calibration module obtains real-time radar data in an on-line state, initializes residual data for estimation, carries out standardization processing, carries out self-adaptive decomposition on the residual data, builds the input of an LSTM model, carries out residual estimation, reconstruction and anti-standardization on the input data of the LSTM model by using a model trained by the off-line training module, calculates and obtains calibrated TIADC output data by combining acquired real-time radar data flow to be processed, and verifies the accuracy of TIADC mismatch error digital calibration.
6. An electronic device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 4 when the computer program is executed.
7. A computer readable storage medium storing computer instructions which, when executed by a processor, implement the steps of the method of any one of claims 1 to 4.

Description

TIADC mismatch error digital calibration method and system applied to radar data acquisition system Technical Field The invention belongs to the technical field of radar signal sampling and processing, and particularly relates to a TIADC mismatch error digital calibration method and system applied to a radar data acquisition system. Background Analog-to-digital Converter (ADC) is a key device for converting an Analog signal into a digital signal, and a data acquisition system using the ADC as a core device is widely used in modern electronic systems such as radar systems, communication, test instruments, digitizers, and the like. In particular in the field of radar systems, the signals received by the radar are typically analog signals, which are continuous voltage waveforms. The ADC converts these analog signals into discrete samples in digital form. This allows the signal to be processed by digital systems, including storage, analysis, display, etc. The sampling rate of the ADC determines the response rate of the system to the signal. Higher sampling speeds may capture signal changes faster, thereby increasing the response time of the radar system. The rise speed of the ADC manufacturing process level is much slower than the rapid rise of the application requirements, and it is very difficult to manufacture an ADC meeting the application requirements by means of the rise of the process level, so that the performance of the data acquisition system using the ADC as a core device gradually becomes a performance bottleneck of a modern electronic system. Therefore, how to realize a high-speed and high-resolution data acquisition system by using the existing ADC chip has important significance. The time-interleaved Analog-to-digital Converter (TIADC) uses a plurality of identical ADCs to sample the same signal source sequentially, so that the overall sampling rate of the system can be effectively improved, and the performance limit of the single-channel ADC is broken through. However, due to process variations between ADC chips, it is not ensured that each ADC has the same physical characteristics, which directly affects the performance of TIADC. These undesirable physical characteristics can cause tia dc mismatch problems, resulting in non-uniform signal sampling and spurious components in the output signal. These mismatches mainly include bias mismatch, gain mismatch, time mismatch, etc. Therefore, exploring and calibrating the influencing factors restricting TIADC indexes is important to improving the performance of a data acquisition system and the capability of a radar system and enhancing the reliability and the efficiency of the radar system in various applications. The existing TIADC calibration method mainly researches different filter structures from the point of signal processing and completes calibration by combining signal characteristics. Some calibration methods are limited by the number of TIADC channels, some cannot break through the Nyquist band limitation, and some require a separate design of a calibration algorithm according to each mismatch, which results in a complex calibration process. Disclosure of Invention The invention provides a TIADC mismatch error digital calibration method and system applied to a radar data acquisition system, which are applied to TIADC in the radar data acquisition system, and solve the error problem generated by bias mismatch, gain mismatch and time mismatch in the TIADC, thereby improving the performance index of the data acquisition system, enhancing the accuracy of data acquisition and improving the reliability and the working efficiency of the radar. The invention is realized by the following technical scheme: A TIADC mismatch error digital calibration method applied to a radar data acquisition system specifically comprises the following steps: Step 1, selecting a pair of input and output data of a radar as a tag model to train in an off-line state, calculating ideal sampling output of the radar signal according to a mathematical formula of selecting an input analog signal, or directly connecting a high-end digital oscilloscope to a signal generator to acquire an expected signal, and constructing a residual sequence containing all mismatch errors by combining actual output data of TIADC; step2, carrying out standardization processing on the residual sequences containing all mismatch errors in the step 1, so that the characteristic scales of the data are similar; Step 3, decomposing the standardized residual signals constructed in the step 2 into sub-signals with different information granularities by using a signal self-adaptive decomposition method, and training by using an LSTM model to obtain the time sequence data correlation in the signals; Step4, initializing residual data for estimation in an on-line state, performing standardization processing, and then performing self-adaptive decomposition on the residual data to construct the input of an LSTM mo