CN108008660A - Orthogonal signalling high-speed, high precision processing method based on DSP and FPGA - Google Patents

Abstract

The invention discloses a high-speed and high-precision processing method for orthogonal signals based on DSP and FPGA, including two parts: an FPGA industrial-grade processor and a DSP. The FPGA industrial-grade processor and the DSP realize data communication through a dual-port RAM; the DSP is used as the data The processing core is responsible for the main mathematical transformation calculation; the FPGA is used as the coprocessor, responsible for all digital logic control, and realizes the acquisition of quadrature signals, mean value filtering, error compensation, real-time phase calculation, phase fine resolution and storage of related data and display; DSP calculates the error compensation parameters of the quadrature signal data collected by FPGA through matrix operation, and transmits the error compensation parameters to FPGA through dual-port RAM communication. In the whole process of signal collection, transmission, correction, fine resolution and data storage, the present invention cooperates with DSP and FPGA, which can not only meet the real-time requirements of the system, but also realize high-precision processing of orthogonal signals.

Description

High-speed and High-precision Processing Method of Orthogonal Signal Based on DSP and FPGA

Technical field:

The invention relates to the field of precision measurement, and mainly relates to a high-speed and high-precision processing method for orthogonal signals based on DSP and FPGA.

Background technique:

Grid sensors and laser interferometers are widely used in modern precision measuring instruments, CNC machine tools, lithography machines, ultra-finishing, quantum physics, biomolecules and other fields, and are typical measurement equipment for modern precision testing. Therefore, it is of great practical significance to improve the range and precision of grid sensor and laser interferometer precision distance measurement technology. In practical applications, due to the influence of external environmental noise, manufacturing errors of optical components, electronic noise introduced by electronic circuits, and other comprehensive factors, there are unequal AC amplitudes and DC level components in the actually measured two-way signals. As well as errors such as quadrature error, they will ultimately affect the accuracy of displacement measurement. In order to meet the increasingly sophisticated measurement requirements and applications, grid sensors and laser interferometers are also correspondingly developing in several directions such as high resolution, high precision, and high speed measurement.

The existing methods for improving the measurement accuracy of grid sensors and laser interferometers mainly rely on the traditional Heydemann ellipse correction method. The ellipse correction method is based on the method of real-time estimation of orthogonal interference signal parameters, uses mathematical models to compensate errors, and then improves the accuracy of the measurement system through fine resolution. The quadrature signal processing method based on the Heydemann ellipse correction method involves data acquisition, error parameter calculation, error compensation process, fine resolution process, storage and display, etc., and has very high speed and processing accuracy in each link of signal processing. Therefore, it is usually only possible to perform offline analysis and compensation of quadrature signals, and it is difficult to realize real-time compensation and high-precision processing of quadrature signals at the same time.

Invention content:

The purpose of this invention patent is to solve the problems of poor real-time performance, low integration and low precision in the high-precision processing process of quadrature signals, optimize the precision, resolution, real-time performance and other indicators of the measurement data output by the quadrature signal measurement system, and provide A high-speed and high-precision processing method of orthogonal signals based on DSP and FPGA with small size, flexible use, low power consumption, high performance and fast response speed; this method can overcome external environmental noise, manufacturing errors of optical components, electronic circuits, etc. The introduction of electronic noise and other comprehensive factors, based on the FPGA and DSP architecture, improves the signal processing speed and signal processing accuracy, expands the use range of optical instruments such as homodyne interferometers, and provides high-precision grating signal analysis and high-speed error compensation. Provide hardware solutions.

In order to achieve the above object, the technical scheme adopted by the patent of the present invention is:

A high-speed and high-precision processing method for orthogonal signals based on DSP and FPGA, characterized in that: it includes two parts: FPGA industrial-grade processor (2) and DSP (3), FPGA industrial-grade processor (2), DSP (3) Realize data communication by dual-port RAM (10); The concrete realization steps of described method are as follows:

First, the light intensity signal outputs an analog orthogonal voltage signal containing displacement information through the photoelectric conversion and signal conditioning circuit (1), and the FPGA industrial-grade processor (2) drives the AD sampling module (4) to perform high-speed analog-to-digital conversion on the signal. And in the FPGA industrial-grade processor (2), obtain further optimized orthogonal signal data through mean filtering (5), and write it into the dual-port RAM (10) by the FPGA industrial-grade processor (2) for storage and high-speed buffering;

Subsequently, the DSP (3) reads the quadrature signal data in the dual-port RAM (10) through the EMIF interface to expand the memory, and screens (11) the sampled data to eliminate abnormal points, and then calculates the compensation by matrix operation parameter (12), and the compensation parameter is written in the dual-port RAM (10) and SDRAM (13), and then read by the FPGA industrial-grade processor (2);

At this time, the FPGA industrial-grade processor (2) uses the calculated compensation parameters to perform real-time error compensation (6), real-time phase calculation (7) and fine resolution (8) for the quadrature signal;

Finally, the real-time updated phase value is written into the dual-port RAM (10) and the displacement is calculated and stored in real time by the DSP (3), and the displacement data is read by the FPGA industrial-grade processor (2) through the dual-port RAM (10) and then Save through the data storage SD card (14), and simultaneously send to the VGA display module (9) for display.

A kind of high-speed and high-precision processing method for orthogonal signals based on DSP and FPGA is characterized in that: the real-time data interaction between the FPGA industrial-grade processor (2) and DSP (3) is as follows: FPGA industrial-grade processing The instruction to the direction of DSP (3) from device (2) is by setting an interrupt signal in FPGA industrial-grade processor (2), when DSP (3) receives this signal, then starts to follow the rule from dual-port RAM (10) Perform mean filter data reading; when DSP (3) transmits data to FPGA industrial-grade processor (2), follow the same execution rule, DSP (3) provides an IO signal, informs FPGA industrial-grade processor (2) from Read data in dual-port RAM (10).

A kind of high-speed high-precision processing method of quadrature signal based on DSP and FPGA is characterized in that: carry out compensation parameter calculation (12) in DSP (3), concrete steps are: step by step compensation quadrature signal by error source Errors, including quadrature error α, amplitude fluctuation, and DC drift error in the quadrature signal, a total of 7 errors.

Described a kind of high-speed high-precision processing method for orthogonal signals based on DSP and FPGA is characterized in that: phase calculation (7) and fine resolution (8) are carried out in FPGA (2), and concrete steps are as follows: phase calculation ( 7) According to the CORDIC algorithm, the phase value corresponding to each sampling point is calculated based on the arctangent trigonometric function operation, and the subdivision number is obtained through interval conversion, and the movement direction of the measurement data is judged by the size of the subdivision number, and the tracking of the displacement is realized, and the output of each Effective sampling points correspond to data containing real-time displacement information, and are converted into displacement information in DSP (3).

FPGA industrial-grade processor (2) is a co-processor responsible for all digital logic control; DSP (3) is the core function module of data processing, which adopts embedded hardware multiplier, improved Harvard structure, pipeline structure and optimized integrated circuit design, The throughput of the signal processor is increased, and the execution speed of DSP instructions is greatly improved, thereby realizing efficient calculation of error compensation parameters.

When the FPGA industrial-grade processor (2) drives high-speed sampling: when the measured data increases in the positive direction, the real-time phase value increases in the same cycle, and the phase difference between two adjacent points is always positive and close to 0, moving forward to the new cycle time When the real-time phase value changes from π to π, the phase difference between two adjacent points at this moment is negative and close to ; when the measured data decreases in reverse, the real-time phase value decreases in the same cycle, and the phase difference between two adjacent points is always negative and close to 0, the real-time phase value changes from π to 0 in the reverse direction to the new cycle moment, and the phase difference between two adjacent points at this moment is positive and close to 2π. Therefore, the integral cycle number judges the signal moving direction according to the phase difference between two adjacent points, and the integral cycle number is added and subtracted when entering a new cycle.

Compared with prior art, the advantage of the present invention is:

Based on the patent "FPGA-Based Real-Time Processing Method of Orthogonal Signals" (Patent No.: 201710234903X), the present invention has made improvements in hardware architecture and algorithms, and is a pure hardware implementation of high-precision real-time processing of orthogonal signals.

The present invention is structurally based on FPGA and DSP architecture, with DSP and FPGA as the core hardware architecture, with DSP as the data processing core, mainly responsible for the main mathematical transformation calculation; with the FPGA industrial-grade processor as the coprocessor, responsible for all digital logic Control, give full play to the high-speed parallel characteristics of FPGA and the fast calculation ability of DSP. FPGA and DSP work together to complete the communication between the two through a piece of dual-port RAM, and divide its interior into two equal parts according to its storage space, and use ping-pong technology to complete high-speed and real-time data buffering of orthogonal signals. In the whole process of signal acquisition, transmission, correction, fine resolution and data storage, the method of the present invention uses DSP and FPGA to cooperate in division of labor, which can not only meet the real-time requirements of the system, but also realize high-precision processing of orthogonal signals. The invention has a clear structure, fast and efficient signal processing, and can realize online compensation for the quadrature error, amplitude fluctuation, and DC drift error of the grating sensor and single-frequency laser interferometer signal, and is useful for the analysis and error detection of high-precision grating signals. Real-time compensation provides engineering solutions.

Description of drawings:

Figure 1 is a system structure diagram of a high-speed and high-precision processing method for orthogonal signals based on DSP and FPGA.

Figure 2 is a block diagram of a real-time quadrature signal processing system.

Fig. 3 is the working flowchart of FPGA industrial grade processor.

Fig. 4 is the working flowchart of DSP.

Detailed ways:

Referring to accompanying drawing 1, a kind of orthogonal signal high-speed high-precision processing method based on DSP and FPGA, comprises FPGA industrial grade processor 2 and DSP3 two parts, FPGA industrial grade processor 2, DSP3 realize data communication by dual-port RAM10; The specific implementation steps of the method are as follows:

First, the light intensity signal outputs an analog orthogonal voltage signal containing displacement information through the photoelectric conversion and signal conditioning circuit 1, and the FPGA industrial-grade processor 2 drives the AD sampling module 4 to perform high-speed analog-to-digital conversion on the signal, and processes it in the FPGA industrial-grade In the device 2, the orthogonal signal data obtained through mean filtering 5 is further optimized, and is written into the dual-port RAM 10 by the FPGA industrial-grade processor 2 for storage and high-speed buffering;

Subsequently, the DSP (3) reads the orthogonal signal data in the dual-port RAM 10 through the EMIF interface to expand the memory, and screens 11 the sampled data to remove abnormal points, and then calculates the compensation parameters 12 through matrix operations, and Compensation parameters are written into dual-port RAM10 and SDRAM13, and then read by FPGA industrial-grade processor 2;

At this time, the FPGA industrial-grade processor 2 uses the calculated compensation parameters to perform real-time error compensation 6, real-time phase calculation 7 and fine resolution 8 for the quadrature signal;

Finally, the real-time updated phase value is written into the dual-port RAM 10 and the displacement is calculated and stored in real time by the DSP3. The displacement data is then read by the FPGA industrial-grade processor 2 through the dual-port RAM 10, and stored through the data storage SD card 14, and sent simultaneously. To the VGA display module 9 for display.

The real-time data interaction between FPGA industrial-grade processor 2 and DSP3 is as follows: FPGA industrial-grade processor 2 sends instructions to DSP3 by setting an interrupt signal in FPGA industrial-grade processor 2, and when DSP3 receives the signal, it starts Read mean value filtering data from dual-port RAM10 according to the rules; when DSP3 transmits data to FPGA industrial-grade processor 2, follow the same execution rules, and DSP3 provides an IO signal to notify FPGA industrial-grade processor 2 from dual-port RAM10 read data in.

Referring to accompanying drawing 2, the signal processing method system can be divided into: a signal acquisition module, a logic control module, a data storage module, a real-time display module, a high-speed transmission module, and a compensation parameter calculation module. According to the different responsibilities of FPGA industrial-grade processor 2 and DSP3, FPGA industrial-grade processor 2 is mainly responsible for the acquisition of orthogonal signals, display and storage of displacement signals, high-speed buffering of data, and logic control of the entire circuit; DSP3 is mainly responsible for normal Cross signal error compensation parameter calculation.

See Figures 3 and 4, the workflow of the entire signal processing system is as follows: the system is powered on, and after the FPGA circuit completes the initialization action, high-speed acquisition and mean value filtering of orthogonal signals are started; the DSP completes the self-test, and waits for an external interrupt signal from the FPGA direction 1. The FPGA opens the data channel to the DSP, and starts to buffer the orthogonal signal sampling data. The data is stored in the dual-port RAM. When the storage capacity reaches a certain amount, it starts to send an interrupt signal 1 to the DSP in real time; the DSP receives the signal 1 After that, read the data from the dual-port RAM immediately and perform the matrix operation of the compensation parameters based on the Heydemann ellipse fitting algorithm. After the calculation is completed, the compensation parameters are stored in the dual-port RAM, and an interrupt signal 2 is provided to the FPGA; the FPGA receives the interrupt signal Immediately after 2, the compensation parameters are used to compensate the sampling signal, and the compensated correction signal enters the arctangent operation based on the CORDIC algorithm and the fine resolution module based on the interval conversion, and the real-time updated phase information is stored in the dual-port RAM and sent to the DSP provides interrupt signal 3; DSP receives interrupt signal 3, internally converts phase information into real-time displacement and stores it in dual-port RAM, and provides interrupt signal 4 to FPGA; FPGA reads displacement from dual-port RAM after receiving signal 4 data, and store the displacement data in the SD card in the form of files. At this time, if the outside world does not send an end signal, the data transmission and signal processing will continue; if the outside world sends an end signal, the FPGA will notify the DSP to end the encoding.

The quadrature error α, amplitude fluctuation, and DC drift error in the quadrature signal have a total of 7 errors. The signal is calculated by the error compensation parameters realized by DSP, and the algorithm is as follows:

The actual mathematical model of the quadrature signal is:

Then there are:

According to the relationship of trigonometric functions, the elliptic equation of the quadrature signal can be obtained:

In order to facilitate curve fitting, the above formula can be rewritten as:

AX ² +BY ² +CXY+DX+EY＝1

Perform the following matrix operations in DSP:

By calculating the fifth-order matrix, the values of coefficients A, B, C, D, and E can be obtained, and then the value of α can be obtained by simple calculation, and written into FPGA to compensate the orthogonal signal:

The embodiments of the present invention described above are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the claims of the present invention.

Claims (4)

1. a high-speed and high-precision processing method for orthogonal signals based on DSP and FPGA, is characterized in that: comprise FPGA industrial grade processor (2) and DSP (3) two parts, FPGA industrial grade processor (2), DSP ( 3) Realize data communication by dual-port RAM (10); The concrete realization steps of described method are as follows: First, the light intensity signal outputs an analog orthogonal voltage signal containing displacement information through the photoelectric conversion and signal conditioning circuit (1), and the FPGA industrial-grade processor (2) drives the AD sampling module (4) to perform high-speed analog-to-digital conversion on the signal. And in the FPGA industrial-grade processor (2), obtain further optimized orthogonal signal data through mean filtering (5), and write it into the dual-port RAM (10) by the FPGA industrial-grade processor (2) for storage and high-speed buffering; Subsequently, the DSP (3) reads the quadrature signal data in the dual-port RAM (10) through the EMIF interface to expand the memory, and screens (11) the sampled data to eliminate abnormal points, and then calculates the compensation by matrix operation parameter (12), and the compensation parameter is written in the dual-port RAM (10) and SDRAM (13), and then read by the FPGA industrial-grade processor (2); At this time, the FPGA industrial-grade processor (2) uses the calculated compensation parameters to perform real-time error compensation (6), real-time phase calculation (7) and fine resolution (8) for the quadrature signal; Finally, the real-time updated phase value is written into the dual-port RAM (10) and the displacement is calculated and stored in real time by the DSP (3), and the displacement data is read by the FPGA industrial-grade processor (2) through the dual-port RAM (10) and then Save through the data storage SD card (14), and simultaneously send to the VGA display module (9) for display. 2. a kind of high-speed high-precision processing method for orthogonal signals based on DSP and FPGA according to claim 1, is characterized in that: the real-time data interaction between the FPGA industrial grade processor (2) and DSP (3) As follows: the instruction from the FPGA industrial grade processor (2) to the DSP (3) direction is by setting an interrupt signal in the FPGA industrial grade processor (2), and when the DSP (3) receives the signal, it starts to read from the dual-port RAM (10) according to the rules to read the average value filter data; DSP (3) to FPGA industrial grade processor (2) when data transmission, follow the same execution rules, DSP (3) provides an IO signal to notify the FPGA industrial grade The processor (2) reads data from the dual-port RAM (10). 3. a kind of high-speed high-precision processing method for orthogonal signals based on DSP and FPGA according to claim 1, is characterized in that: carry out compensation parameter calculation (12) in DSP (3), concrete steps are: press error source Step-by-step compensation of quadrature signal errors, including quadrature error α, amplitude fluctuation, and DC drift error in the quadrature signal, a total of 7 errors. 4. a kind of high-speed high-precision processing method for orthogonal signals based on DSP and FPGA according to claim 1 is characterized in that: phase calculation (7) and fine resolution are carried out (8) in FPGA (2), specifically The steps are as follows: Phase calculation (7) According to the CORDIC algorithm, the phase value corresponding to each sampling point is calculated based on the arctangent trigonometric function operation, and the subdivision number is obtained through interval conversion, and the movement direction of the measurement data is judged and realized by the size of the subdivision number. For displacement tracking, output data containing real-time displacement information corresponding to each effective sampling point, and convert it into displacement information in DSP (3).