CN106330209B - It is a kind of based on FPGA front-end digital makeup set - Google Patents

Abstract

A front-end digitization device based on FPGA, including a digital front-end and an analysis back-end, the digital front-end is connected to the analysis back-end; the digital front-end is used to complete the collection, preprocessing, storage and transmission of space electromagnetic signals, and convert the original space electromagnetic signals and The feature value data is encapsulated into a data frame and transmitted to the analysis backend; the analysis backend is used to receive and analyze the data frame sent by the digital front end, perform data post-processing on the data frame, and provide interface management at the same time. The invention has strong electromagnetic compatibility, and can improve the measurement capability of space electromagnetic signal characteristic parameters of important sensitive equipment.

Description

A front-end digital device based on FPGA

technical field

The invention relates to the technical field of wireless communication, in particular to an FPGA-based front-end digital device.

Background technique

In the military field, "complex electromagnetic environment" is usually used to describe objectively existing complex electromagnetic phenomena, and the electromagnetic compatibility of information equipment is used to characterize the anti-interference ability and robustness of equipment; in the industrial field, with the modern communication and power electronics The rapid development of technology and the large-scale application of high-power nonlinear equipment produce power harmonics that pollute the power grid and radiate electromagnetic waves to the surroundings. Therefore, the electromagnetic environment of industrial measurement and control sites is becoming increasingly complex. However, electronic system equipment such as satellite communication, mobile communication, and electronic countermeasures work in this increasingly complex electromagnetic environment for a long time, and have extremely high requirements on the reliability of electromagnetic compatibility of the equipment.

The front-end equipment shoulders the task of collecting and extracting the characteristic parameters of the wireless communication channel with high precision, high resolution, and strong real-time performance in the fast time-varying, large-bandwidth, and multi-dimensional complex electromagnetic environment. It is the basis of the internal relationship of relevant characteristic parameters such as frequency domain, energy domain and air domain, and it is also the premise of comprehensively and dialectically understanding the complexity of wireless channels. The traditional front-end equipment instrument system is composed of the front-end primary instrument, the back-end secondary instrument and the transmission line between them. The analog/digital converter (ADC: Analog to Digital Converter) is usually arranged in the back-end secondary instrument. Therefore, The useful signal on the transmission line is a highly sensitive analog signal, which is extremely vulnerable to electromagnetic interference and reduces the reliability of the equipment.

Contents of the invention

The technical problem to be solved by the present invention is to provide an FPGA-based front-end digital device with strong electromagnetic compatibility to overcome the shortcomings of the above-mentioned background technology.

The technical scheme that the present invention solves its technical problem adopts is, a kind of front-end digitization device based on FPGA, comprises digitization front-end and analysis back-end, and digitization front-end is connected with analysis back-end; Described digitization front-end is used for finishing the collection of space electromagnetic signal, Preprocessing, storage and transmission, packaging the original space electromagnetic signal and eigenvalue data into a data frame, and transmitting it to the analysis backend; the analysis backend is used to receive and analyze the data frame sent by the digital front end, and perform Data post-processing, while providing interface management.

Further, the digital front-end and the analysis back-end are connected through optical fiber cables, and the communication between the digital front-end and the analysis back-end adopts optical fiber communication.

Further, the digital front-end includes a receiving antenna array, a radio frequency switch matrix, a signal transmission module, an A/D acquisition module, an acquisition drive module, a digital intermediate frequency converter, a deep learning convolution calculation layer, and an optical fiber transmission protocol layer; the receiving The antenna array is connected with the radio frequency switch matrix; the radio frequency switch matrix is connected with the signal transmission module; the signal transmission module is connected with the A/D acquisition module; the A/D acquisition module is connected with the acquisition drive module; the acquisition The drive module is respectively connected with the digital intermediate frequency converter and the optical fiber transmission protocol layer; the digital intermediate frequency converter is connected with the deep learning convolution calculation layer; the deep learning convolution calculation layer is connected with the optical fiber transmission protocol layer;

The acquisition drive module, digital intermediate frequency converter, deep learning convolution calculation layer and optical fiber transmission protocol layer are all integrated in FPGA in the form of IP CORE;

The FPGA is embedded with a processor, and the processor performs task management on the IP CORE of the acquisition drive module, digital intermediate frequency converter, deep learning convolution calculation layer and optical fiber transmission protocol layer;

The receiving antenna array is used to complete the receiving of electromagnetic signals in the wireless communication channel, and the antenna elements of the receiving antenna array are arranged in an array form to support the receiving of multi-dimensional space electromagnetic signals;

The radio frequency switch matrix is used to complete automatic switching and flexible access of space electromagnetic signals;

The signal transmission module is used to complete the down-conversion of the space electromagnetic signal, and transmit the radio frequency signal to the intermediate frequency;

The A/D acquisition module is used to complete the high-precision analog-to-digital conversion of the intermediate frequency signal;

The acquisition driver module is used to complete the signal reception after the A/D acquisition module, and outputs the digital signal on-chip according to the timing characteristics of the Block RAM in the FPGA;

The digital intermediate frequency converter is used to complete the relevant data domain transformation and matrix operation of the on-chip digital signal time domain, frequency domain, energy domain and space domain;

The deep learning convolution calculation layer is used to establish a deep learning training model for the characteristic parameters of the wireless communication channel in a complex electromagnetic environment, to approximate the transfer function expression of the wireless communication channel, or to establish a mapping table of spatial parameters;

The optical fiber transmission protocol layer is used to encapsulate the on-chip digital signal after the acquisition drive module and the on-chip characteristic signal after the deep learning convolution calculation layer into a data frame signal, complete the behavior and logic control of the optical fiber communication interface, and complete the digital front-end The AXI bus address of the FPGA and the memory physical address mapping of the analysis backend.

Further, the processor is an 8-bit scalable PicoBlaze soft-core processor.

Further, the analysis backend includes an upper computer software system and an optical fiber transmission network; the upper computer software system is connected to the optical fiber transmission network, and the optical fiber transmission network is connected to the optical fiber transmission protocol layer of the digital front end;

The upper computer software system is used to provide data analysis software, data post-processing software and interface software. The data analysis software is used to receive and analyze the data frames sent by the digital front-end; the data post-processing software is used to perform data post-processing on the data frames; the interface Software for interface management;

The optical fiber transmission network is used to provide a two-way high-speed communication link between the digital front end and the upper computer software system.

Further, the content of the interface management includes observing and recording the analysis results of the spatial electromagnetic signal characteristic value of the digital front-end and the alarm situation of the identification of the interested target; observing the consumption status of RAM resources in the FPGA.

Further, the data flow direction of the FPGA-based front-end digital device is: the original space electromagnetic signal passes through the receiving antenna array, the radio frequency switch matrix, the signal transmission module and the A/D acquisition module in turn to obtain the original digital signal; The on-chip digital signal obtained after the module, the on-chip digital signal is divided into two channels, one signal directly leads to the optical fiber transmission protocol layer, and the other signal passes through the digital intermediate frequency converter and the deep learning convolution calculation layer in turn, and the deep learning convolution calculation layer The on-chip characteristic signal is obtained through layer calculation, and the obtained on-chip characteristic signal leads to the optical fiber transmission protocol layer; the optical fiber transmission protocol layer encapsulates the on-chip digital signal and the on-chip characteristic signal into a data frame signal, and the data frame signal passes through the optical fiber transmission network Transparent transmission to the host computer software system.

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

(1) Analog signals susceptible to electromagnetic interference are collected and processed locally at the digital front-end of the present invention, and then transmitted to the analysis back-end through optical fiber communication, completely cutting off the propagation path of electromagnetic disturbance, ensuring the electromagnetic compatibility of sensitive equipment, and eradicating traditional The problem of analog front-end or measuring instruments being susceptible to electromagnetic disturbances further improves the ability to measure the characteristic parameters of space electromagnetic signals for major sensitive equipment.

(2) Further introduce FPGA (Field-Programmable Gate Array, Field Programmable Gate Array) to form a new digital front-end with powerful data preprocessing function, so that massive space electromagnetic signals can be pre-processed and analyzed locally at the digital front-end, from Fundamentally solve the task load and processing lag of traditional back-end secondary instruments (or back-end processors).

(3) The digital front-end designed by the present invention not only takes advantage of the high-speed parallelism of the FPGA, but also uses the extremely streamlined and reconfigurable PicoBlaze soft-core processor to complete the task management of the IP CORE in the FPGA, reducing the development difficulty of the FPGA and shortening the The equipment development cycle is short, the circuit scale is small, and the hardware integration is high.

Description of drawings

Fig. 1 is the overall structural block diagram of the present invention.

Fig. 2 is a specific structural block diagram of an embodiment of the present invention.

Detailed ways

The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

With reference to Fig. 1, the FPGA-based front-end digital device of the present invention includes a digital front-end U1 and an analysis back-end U2; the digital front-end U1 and the analysis back-end U2 are connected by an optical fiber cable, and optical fiber communication is adopted between the digital front-end U1 and the analysis back-end U2 way of communication.

The digital front-end U1 is used to complete the collection, preprocessing, storage and transmission of the space electromagnetic signal, and encapsulate the original space electromagnetic signal and eigenvalue data into a data frame, which is transmitted to the analysis back-end U2.

The analysis backend U2 is used to receive and analyze the data frames sent by the digital front-end U1, perform data post-processing on the data frames, and provide interface management at the same time.

Wireless communication channel U3 is the research object that the FPGA-based front-end digitization device of the present invention points to. During use, digital front-end U1 is connected with wireless communication channel U3, and space electromagnetic signals are generated in wireless communication channel U3. The present invention mainly aims at wireless communication channel Acquisition, analysis and extraction of characteristic parameters of U3 in time domain, frequency domain, energy domain and air domain.

Referring to Figure 2, the digital front-end U1 includes a receiving antenna array U11, a radio frequency switch matrix U12, a signal transmission module U13, an A/D acquisition module U14, an acquisition driver module U15, a digital intermediate frequency converter U16, a deep learning convolution calculation layer U17 and Optical fiber transmission protocol layer U18; receiving antenna array U11 is connected with RF switch matrix U12; RF switch matrix U12 is connected with signal transmission module U13; signal transmission module U13 is connected with A/D acquisition module U14; A/D acquisition module U14 is connected with The acquisition drive module U15 is connected; the acquisition drive module U15 is respectively connected with the digital intermediate frequency converter U16 and the optical fiber transmission protocol layer U18; the digital intermediate frequency converter U16 is connected with the deep learning convolution calculation layer U17, and the deep learning convolution calculation layer U17 is connected with the optical fiber transmission The protocol layer U18 is connected; in use, the receiving antenna array U11 is connected with the wireless communication channel U3.

Acquisition driver module U15, digital intermediate frequency converter U16, deep learning convolution calculation layer U17, and optical fiber transmission protocol layer U18 are all integrated in FPGA in the form of IP CORE (Intellectual Property Core IP core, that is, intellectual property core). Communicate with AXI (Advanced eXtensible Interface, a bus protocol) bus.

The processor is embedded in the FPGA, and the processor performs task management on the acquisition driver module U15, the digital intermediate frequency converter U16, the deep learning convolution calculation layer U17, and the optical fiber transmission protocol layer U18 IP CORE.

The processor embedded in the FPGA can be a soft-core processor or a hard-core processor. The soft-core processor can be an 8-bit PicoBlaze soft-core processor or a 32-bit MicroBlaze soft-core processor. The soft-core processor can be tailored; The hard-core processor may be a 32-bit PowerPC hard-core processor. In this embodiment, an 8-bit scalable PicoBlaze soft-core processor is selected.

The receiving antenna array U11 is used to complete the reception of electromagnetic signals in the wireless communication channel U3, and the antenna elements of the receiving antenna array U11 are arranged in an array to support the reception of multi-dimensional space electromagnetic signals;

RF switch matrix U12, used to complete the automatic switching and flexible access of space electromagnetic signals;

The signal transmission module U13 is used to complete the down-conversion of the space electromagnetic signal and transmit the radio frequency signal to the intermediate frequency;

A/D acquisition module U14 is used to complete high-precision analog-to-digital conversion of intermediate frequency signals, and its resolution is generally not lower than 14 bits;

The acquisition drive module U15 is used to complete the signal reception after the A/D acquisition module U14, and output the on-chip digital signal S3 according to the timing characteristics of the BlockRAM (block random access memory) in the FPGA;

The digital intermediate frequency converter U16 is used to complete the relevant data domain transformation and matrix operation of the on-chip digital signal S3 such as time domain, frequency domain, energy domain and space domain;

The deep learning convolution calculation layer U17 is used to establish a deep learning training model for the characteristic parameters of the wireless communication channel U3 in a complex electromagnetic environment, to obtain the transfer function expression of the wireless communication channel U3 in an approximate manner, or to establish a mapping table of spatial parameters ;Deep learning convolution calculation layer mainly involves convolution and multiplication and addition operations;

The optical fiber transmission protocol layer U18 is used to encapsulate the on-chip digital signal S3 after the acquisition driver module U15 and the on-chip characteristic signal S4 after the deep learning convolution calculation layer U17 into a data frame signal S5 to complete the behavior and logic of the optical fiber communication interface Control and complete the mapping between the AXI (Advanced eXtensible Interface) bus address of the FPGA in the digital front-end U1 and the memory physical address of the analysis back-end U2.

The analysis backend U2 includes the upper computer software system U21 and the optical fiber transmission network U22; the upper computer software system U21 is connected to the optical fiber transmission network U22, and the optical fiber transmission network U22 is connected to the optical fiber transmission protocol layer U18 of the digital front end U1.

The upper computer software system U21 is used to provide data analysis software, data post-processing software and interface software. The data analysis software is used to receive and analyze the data frames sent by the digital front-end U1; the data post-processing software is used to perform data post-processing on the data frames ; The interface software is used for interface management; the content of interface management includes observing and recording the analysis results of the eigenvalues of the space electromagnetic signal of the digital front-end U1 and the alarm situation of the identification of interested targets; observing the consumption of RAM resources in the FPGA.

The optical fiber transmission network U22 is used to provide a two-way high-speed communication link between the digital front-end U1 and the upper computer software system U21. The communication rate of a single channel is up to 10 Gbps, and a single chip can be configured with up to 4 transceiver channels.

The FPGA of this embodiment selects the Virtex-7 series FPGA of Xilinx Company, and adopts the "Aurora" communication protocol, which can easily realize the optical fiber communication between the digital front-end U1 and the analysis back-end U2, and the two-way single-channel communication rate is as high as 10Gbps. Demand, conveniently configure up to 4 transceiver channels. The invention can be widely used in collecting and extracting characteristic parameters of wireless communication channels with high precision, high resolution and strong real-time performance under fast time-varying, large bandwidth and multi-dimensional complex electromagnetic environments.

The data flow direction of the FPGA-based front-end digital device of the present invention is: the original space electromagnetic signal S1 originating from the wireless communication channel U3 is an analog signal, which passes through the receiving antenna array U11, the radio frequency switch matrix U12, the signal transmission module U13 and the A /D Acquisition module U14 to obtain the original digital signal S2; the on-chip digital signal S3 obtained after the acquisition drive module U15, the on-chip digital signal S3 is divided into two routes, one route directly leads to the optical fiber transmission protocol layer U18, and the other route The signal passes through the digital intermediate frequency converter U16 and the deep learning convolution calculation layer U17 in turn, and the deep learning convolution calculation layer U17 calculates the on-chip characteristic signal S4, and the obtained on-chip characteristic signal S4 leads to the optical fiber transmission protocol layer U18; The transmission protocol layer U18 encapsulates the on-chip digital signal S3 and the on-chip characteristic signal S4 into a data frame signal S5, and the data frame signal S5 is transparently transmitted to the upper computer software system U21 through the optical fiber transmission network U22.

Use the FPGA-based front-end digital device of the present invention to carry out the method for wireless communication channel feature parameter extraction as follows:

(1) Install the digital front-end U1 in a mechanical structure with good heat dissipation and electromagnetic shielding; power on the digital front-end U1, and check whether the power supply is working normally. After normal power-on, the power indicator light is on, and the entire digital front-end equipment is working without overheating ,odorless;

(2) Download the optical fiber communication test subroutine of the digital front-end U1 in the digital front-end U1, connect the digital front-end U1 and the analysis back-end U2 with the matching optical fiber cable, and confirm that the base address register (Base Address Register) space between the two can normal reading and writing;

(3) Download the on-chip digital signal S3 of the digital front-end U1, collect the subroutine in the digital front-end U1, restart the digital front-end U1, run and analyze the upper computer software of the back-end U2, and confirm that the on-chip digital signal S3 can be collected normally;

(4) Download the main program of the digital front-end U1 in the digital front-end U1, restart the digital front-end U1, run and analyze the upper computer software of the back-end U2, and observe and record the characteristic value analysis of the space electromagnetic signal of the digital front-end U1 in the interface software program The result and the alarm situation of the object of interest identification, and the consumption of RAM resources in the FPGA can be observed at the same time.

(5) During operation, according to the actual situation of space electromagnetic signals in the time domain, frequency domain, energy domain and air domain characteristic parameters, combined with the electromagnetic environment conditions where the digital front-end U1 is located, adjust, improve, increase, and delete digitalization as appropriate Some functions of the digital intermediate frequency converter U16 and the deep learning convolution calculation layer U17 in the front end U1;

(6) After the optimization is completed, repeat the above steps (4) and (5) until the extraction accuracy index and speed requirements of the characteristic parameters of space electromagnetic signals in a specific complex electromagnetic environment are met.

The FPGA-based front-end digital device of the present invention has small circuit scale, high hardware integration, good signal integrity and electromagnetic compatibility, and can be used for high-precision measurement in complex electromagnetic environments.

Those skilled in the art can make various modifications and variations to the present invention, and if these modifications and variations are within the scope of the claims of the present invention and equivalent technologies, then these modifications and variations are also within the protection scope of the present invention.

The content not described in detail in the specification is the prior art known to those skilled in the art.

Claims (5)

1. a kind of front-end digital makeup based on FPGA is set, it is characterised in that：Including Digital front end and analysis rear end, digitlization Front end is connect with analysis rear end；The Digital front end is used to complete acquisition, pretreatment, storage and the biography of spatial electromagnetic signal It is defeated, and luv space electromagnetic signal and characteristic value data are packaged into data frame, it is transmitted to analysis rear end；It uses the analysis rear end In the data frame received and parsing Digital front end is sent, Data Post is carried out to data frame, while providing interface management；

The Digital front end is connected with analysis rear end by fiber optic cable, and optical fiber is used between Digital front end and analysis rear end Communication mode communicates；

The Digital front end includes receiving antenna array, RF switch matrix, signal transmitting module, A/D acquisition modules, acquisition Drive module, digital intermediate frequency converter, deep learning convolutional calculation layer and optical fiber transmission protocol layer；The receiving antenna array It is connect with RF switch matrix；The RF switch matrix is connect with signal transmitting module；The signal transmitting module is adopted with A/D Collect module connection；The A/D acquisition modules are connect with acquisition drive module；The acquisition drive module becomes with digital intermediate frequency respectively Parallel operation is connected with optical fiber transmission protocol layer；The digital intermediate frequency converter is connect with deep learning convolutional calculation layer；The depth Study convolutional calculation layer is connect with optical fiber transmission protocol layer；

The acquisition drive module, digital intermediate frequency converter, deep learning convolutional calculation layer and optical fiber transmission protocol layer with The form of IP CORE is integrated in FPGA；

The FPGA is embedded in processor, and the processor is to acquisition drive module, digital intermediate frequency converter, deep learning convolution The IP CORE of computation layer and optical fiber transmission protocol layer carry out task management；

The receiving antenna array is used to complete the reception of the electromagnetic signal in radio communication channel, the antenna of receiving antenna array Array element is arranged in the form of an array, supports the reception of the spatial electromagnetic signal of various dimensions；

The RF switch matrix is used to complete automatic switchover and the flexible access of spatial electromagnetic signal；

The signal transmitting module is used to complete the down coversion of spatial electromagnetic signal, by radiofrequency signal pick-up to intermediate frequency；

The A/D acquisition modules are used to complete the high precision analogue conversion of intermediate-freuqncy signal；

The acquisition drive module is used to complete the signal after A/D acquisition modules and receives, and according in FPGA Block RAM when Digital signal in sequence characteristic output chip；

The digital intermediate frequency converter is for completing to digital signal time domain, frequency domain, energy domain and the related data in spatial domain in piece Domain converts and matrix operation；

The deep learning convolutional calculation layer is used to establish the depth to radio communication channel characteristic parameter under complex electromagnetic environment Learning training model, seeks the transmission function expression formula of radio communication channel approximantly, or establishes the mapping table of spatial parameter；

The optical fiber transmission protocol layer is for digital signal and deep learning convolutional calculation layer in the piece after acquiring drive module Characteristic signal is packaged into data frame signal in piece afterwards, completes the behavior and logic control of fiber optic data communication interface, completes digitlization The memory physical address map of the AXI bus address of FPGA and analysis rear end in front end.

2. the front-end digital makeup based on FPGA is set as described in claim 1, it is characterised in that：The processor can for 8 The PicoBlaze soft-core processors of cutting.

3. the front-end digital makeup based on FPGA is set as described in claim 1, it is characterised in that：The analysis rear end includes upper Position machine software systems and optical fiber transmission network；Upper computer software system is connect with optical fiber transmission network, optical fiber transmission network and number The optical fiber transmission protocol layer of word front end connects；

Upper computer software system is for providing data analysis software, data post-processing software and interface software, data analysis software The data frame sent for receiving and parsing Digital front end；After data post-processing software is used to carry out data to data frame Reason；Interface software is used for interface management；

Optical fiber transmission network is used to provide the bidirectional high speed communication link between Digital front end and upper computer software system.

4. the front-end digital makeup based on FPGA is set as claimed in claim 3, it is characterised in that：The content of the interface management Including observing and recording the spatial electromagnetic signal characteristic value analysis result of Digital front end and the alarm of interesting target identification Situation；Observe the RAM resource consumption situations in FPGA.

5. the front-end digital makeup based on FPGA is set as claimed in claim 3, it is characterised in that：The front end based on FPGA The data flow direction of digitalizer is：Luv space electromagnetic signal passes through receiving antenna array, RF switch matrix, letter successively Number transmitting module and A/D acquisition modules, obtain raw digital signal；Using number letter in the piece obtained after acquisition drive module Number, digital signal is divided into two-way in piece, and signal is directly communicated to optical fiber transmission protocol layer all the way, and another way signal passes through number successively Medium-frequency changer and deep learning convolutional calculation layer, deep learning convolutional calculation layer are calculated characteristic signal in piece, obtain Characteristic signal leads to optical fiber transmission protocol layer in piece；Optical fiber transmission protocol layer encapsulates characteristic signal in digital signal in piece and piece At data frame signal, data frame signal is pass-through to upper computer software system by optical fiber transmission network.