CN112242942A - Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system - Google Patents

Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system Download PDF

Info

Publication number
CN112242942A
CN112242942A CN202010794926.8A CN202010794926A CN112242942A CN 112242942 A CN112242942 A CN 112242942A CN 202010794926 A CN202010794926 A CN 202010794926A CN 112242942 A CN112242942 A CN 112242942A
Authority
CN
China
Prior art keywords
data
transmission
distributed
unit
layer
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010794926.8A
Other languages
Chinese (zh)
Other versions
CN112242942B (en
Inventor
宫新保
张正
王彧
宋罡
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Jiaotong University
Original Assignee
Shanghai Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Jiaotong University filed Critical Shanghai Jiaotong University
Priority to CN202010794926.8A priority Critical patent/CN112242942B/en
Publication of CN112242942A publication Critical patent/CN112242942A/en
Application granted granted Critical
Publication of CN112242942B publication Critical patent/CN112242942B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/42—Loop networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/46—Interconnection of networks
    • H04L12/4637—Interconnected ring systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/24—Multipath
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/42—Loop networks
    • H04L2012/421—Interconnected ring systems

Abstract

An information transmission method of a double-layer topological structure of a multi-channel radiometer imaging system is realized by distributed modules in a mode of bottom layer full interconnection connection and top layer ring connection, wherein: each distributed module carries out signal preprocessing, data transmission and sending and complex correlation operation processing according to the acquired digital signal information of a plurality of channels and outputs complex correlation operation result information, one core distributed module in the double-layer topological framework collects the correlation operation results of all other distributed modules and outputs the correlation operation results to a subsequent module for image inversion, and meanwhile, the core distributed module receives external control signals and forwards the control signals to other distributed modules. The invention meets the requirement of low complexity, real-time performance and low power consumption of a multi-channel radiometer system and simultaneously has expandability, and an internal structure corresponding to the topological structure is designed to meet the requirement of real-time processing of a radiometer imaging system.

Description

Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system
Technical Field
The invention relates to a technology in the field of radiometer imaging, in particular to an information transmission method of a double-layer topological structure of a multi-channel radiometer imaging system for satellite-borne equipment.
Background
The interferometric passive microwave imaging technology obtains a visibility function by sampling a spatial frequency domain of a target, and then obtains a final bright temperature image of the target by performing inverse imaging algorithms such as inverse Fourier transform and the like. The sampling of the spatial frequency domain of the target is done by a binary interferometer of different baseline lengths. Binary interferometers essentially perform a complex correlation of the signals received by two antennas. The binary interferometers with different base line lengths can sample to obtain different points of the visibility function, the more kinds of the base line lengths are, the more the visibility function is sampled, and the more complete the coverage of a target space frequency domain is.
In recent years, various meteorological remote sensing applications have put increasing demands on the spatial resolution of radiometric imaging technology. The distance between the two antennas that are furthest apart, i.e., the longest baseline, determines the spatial resolution; the distance between the two antennas that are closest apart, i.e., the shortest baseline, determines the aliasing-free region. Theoretically, the longer the longest baseline, the higher the spatial resolution; the shorter the shortest baseline, the larger the extent of the aliasing-free region. Thus, there is a need for a multi-antenna element, i.e., multi-channel radiometric imaging system to meet the ever-increasing spatial resolution.
The imaging performance of radiometer imaging, which requires the calculation of cross-correlation and auto-correlation between individual channels, depends mainly on two factors: the first is that the more the number of channels, the higher the resolution of the image, the better the imaging quality; secondly, the more the accumulated times of correlation operation, the better the quality of the image, and the two factors are contradictory to the calculation amount of correlation operation: as the number of channels increases, the amount of correlation computation increases quadratically. For example, the number of channels is increased by 2 times, and the calculation amount of the correlation operation is increased by 4 times. This directly blocks the increase in the number of passages. In the prior art, the number of channels processed by the digital signal is smaller than the number of radio frequency channels of the front end, for example, only a fraction of the number of radio frequency channels, and the data of the front end is processed in a polling manner. In this way, a large amount of effective data is discarded in unit time, the same accumulated times of correlation operation can be obtained in a longer time, and the real-time performance of the system is also deteriorated.
Disclosure of Invention
The invention provides an information transmission method of a double-layer topological structure of a multi-channel radiometer imaging system, aiming at the problems that the number of channels imaged by the existing satellite-borne radiometer does not exceed 100, the real-time performance is low and the energy consumption is high, the number of channels imaged by the radiometer can reach 500-1500, the low complexity, the real-time performance and the low power consumption required by the multi-channel radiometer system are simultaneously provided with expandability, and an internal structure corresponding to the topological structure is designed to meet the requirement of real-time processing of the radiometer imaging system.
The invention is realized by the following technical scheme:
the invention relates to an information transmission method of a double-layer topological structure of a multi-channel radiometer imaging system, which is realized by distributed modules in a mode of bottom layer full interconnection connection and top layer ring connection, wherein: each distributed module carries out signal preprocessing, data transmission and sending and complex correlation operation processing according to the acquired digital signal information of a plurality of channels and outputs complex correlation operation result information, one core distributed module in the double-layer topological framework collects the correlation operation results of all other distributed modules and outputs the correlation operation results to a subsequent module for image inversion, and meanwhile, the core distributed module receives external control signals and forwards the control signals to other distributed modules.
The number of the distributed modules is L1 × L2, wherein: L1E N*Representing the number of distributed modules of the underlying full interconnect layer, L2 ∈ N*Number of distributed modules, N, representing a top ring structure*Representing a non-zero set of natural numbers.
The double-layer topological structure comprises the following components: the distributed modules, namely the vertex set and the link set between the vertices, are in one-to-one correspondence, specifically: h _ L1_ L2 ═ G-{ V, E }, with a set of vertices V ═ DUi,iI t ∈ N1, i ∈ N2}, where: n1 { [ N | N ∈ N and N<L1}, N2 ═ N | N ∈ N and N<L2, t represents the number of the distributed module in the top ring structure, i represents the number of the distributed module in the bottom full interconnect layer; set of links is E ═ Ef,Ec}, wherein: ef＝{<DUt1,i,DUt2,i>L t1, t2 belongs to N1, i belongs to N2, t1 is not equal to t2 is a link set of a bottom layer full interconnection layer,
Figure BDA0002625212480000021
Figure BDA0002625212480000022
is a link set of the top ring structure.
The number of the links of the double-layer topological structure is
Figure BDA0002625212480000023
Wherein: when the L1 is reduced to 1, the topological structure is degenerated into a ring, when the L2 is reduced to 1, the topological structure is degenerated into a classic full interconnection structure, and the L2 is recommended in a general practical system>2to characterize a ring-type structure.
The bottom layer full interconnection connection refers to that: all distributed modules in the bottom layer are connected by links, and the link set is Ef(ii) a And data collected by the distributed modules and subjected to time slot division are transmitted among all the distributed modules in the bottom layer in a small circle data transmission mode.
The top ring-shaped connection means that: all distributed modules in the top layer are connected end to end in a ring link mode, and the link set is Ec(ii) a And data received by each distributed module through small circle data transmission is transmitted among all distributed modules in the top layer in a large circle data transmission mode, so that the data of all distributed modules can meet each other pairwise.
The signal preprocessing refers to: the method comprises the steps of sampling input intermediate-frequency analog signals of a plurality of channels of a distributed module to obtain digital signals, then carrying out synchronous processing on the sampled digital signals to remove hardware delay possibly existing among different channels, and finally carrying out quadrature down-conversion on the synchronized digital signals to obtain in-phase signals and quadrature signals located at a fundamental frequency.
The image inversion refers to: and performing inverse Fourier transform on the result obtained by the distributed module through relevant operation to obtain the radiation intensity of the target, and displaying to obtain the brightness temperature image of the target.
The control signals are: the external part obtains the working state of each distributed module by using the control signal, thereby completing a multi-channel radiometer imaging system which can control and monitor the working state of each module.
The small circle data transmission is as follows:
s1 all vertex DUs of topological structuret,iThe corresponding original data frame P is acquired after the time Tt,i；
S2 reaction of Pt,iEqually dividing the data into L1 data frames according to the sequence of acquisition, and respectively recording the data frames as pt,i 0,pt,i 1,pt,i 2,...pt,i L1-1The upper corner mark can be one-to-one corresponding to the top point number of the bottom full interconnection layer, i.e. Pt,i＝{pt,i t′|t′∈N1}；
S3:DUt,iWill L this1L in a slotted data frame 11 other modules addressed to the corresponding underlying full interconnect layer number, leaving p alonet,i tThe transmission process is marked as
Figure BDA0002625212480000031
Simultaneous DUt,iReceiving from L 11 data frame with time slot number t sent by other modules, namely:
Figure BDA0002625212480000032
after a small circle of data transmission, all vertex DUst,iGet DU from self and same top ring structure number it′,iThe collection of acquired data frames of time slot number t, i.e.Pi t，Pi t＝{p0,i t,p1,i t,p2,i t,...pL1-1,i t}。
The large circle data transmission means that:
s4: data frame { P) obtained by small circle transmissioni tI ∈ N2} totals the ring structure's neighboring vertices with the same underlying full interconnect layer number t
Figure BDA0002625212480000037
Clockwise and counterclockwise alternating data migration in secondary ring, DUt,iIs DUt,inAnd DUt,ipWherein: for brevity, the subscripts in (i +1) mod L2, ip (i + L2-1) mod L2, i ∈ N2, and DU may be used for brevityt,i+1And DUt,i-1Replacing DUt,inAnd DUt,ipTo represent DUt,iDefining the clockwise data direction as DU for the next and previous nodes in the ringt,iFrom DUt,i-1Receive data and forward to DUt,i+1Transmitting data with DU in anti-clockwise directiont,iFrom DUt,i+1Receive data and forward to DUt,i-1Sending data, if the first transmission direction is not clockwise, the data transmission is as follows:
Figure BDA0002625212480000033
further, the number of transitions N in step S4transWhen the number of the data is odd, the direction of the first data migration is changed into anticlockwise when the large-circle data transmission of the next frame data is carried out, and at the moment, the positive and negative signs in the data transmission expression are just required to be totally reversed, so that the full-duplex bidirectional bandwidth of the ring-shaped structure link is fully utilized; number of migration NtransAnd when the number of the data is even, the problem of changing the direction of the next frame data large circle transmission in the data migration direction does not need to be considered.
Preferably, the small circle data transmission and the large circle data transmission are performed simultaneously, and the two data transmission modes are coupled through a data cache unit arranged in the distributed module.
Preferably, to ensure real-time transmission, when the sampling rate of signal preprocessing is fs(Hz) and a sampling bit width Wadc(bit), when the number of sampling channels integrated by each distributed module is m, the required bidirectional bandwidth of each link of the full interconnection layer is
Figure BDA0002625212480000034
The required bidirectional bandwidth per link of the ring structure is
Figure BDA0002625212480000035
Thereby corresponding to a required total transmission bandwidth of the system of
Figure BDA0002625212480000036
Because the bandwidth requirements of the two links are mutually decoupled, when a real-time system needs to expand the scale by increasing the number of L2, only the bandwidth requirement of the ring link needs to be considered, and the bandwidth requirement of the full interconnection link does not need to be considered, so that the scale expandability of the topological structure is embodied.
Preferably, the transmission power consumption, the transmission bandwidth and the link number of the small circle data transmission and the large circle data transmission are optimized by adjusting the values of L1 and L2, specifically: the power consumption of each transmission interface of the bottom full interconnection layer is Ptrans_L1＝P0_L1+Pd_L1·BWL1The power consumption of each transmission interface of the top ring structure is Ptrans_L2＝P0_L2+Pd_L2·BWL2Total transmission power consumption of
Figure BDA0002625212480000041
Wherein: the power consumption of the transmission interface consists of static power consumption and dynamic power consumption, Ptrans＝P0+Pd·BW，P0For static power consumption, corresponds to P0_L1Static power consumption, P, for each transmission interface of the bottom fully interconnected layer0_L2Static power consumption of each transmission interface of the top ring structure isIs kept constant in transmission, PdFor dynamic power consumption per bit, corresponding to Pd_L1Dynamic power consumption per bit, P, for each transmission interface of the underlying fully interconnected layerd_L2For each bit dynamic power consumption of each transmission interface of the top ring structure, the dynamic power consumption is in direct proportion to the bandwidth BW of the transmission of the interface, and the corresponding BWL1Transmission bandwidth, BW, for each transmission interface of the underlying fully interconnected layerL2For the transmission bandwidth of each transmission interface of the top ring structure, the coefficient of 2 exists in the total transmission power consumption because each link corresponds to two transmission interfaces.
Technical effects
The invention integrally solves the problem of how to complete the real-time processing of input data while reducing the complexity of the system under the condition that the number of input channels of the radiometer imaging system is fixed and unchanged.
Compared with the prior art, the invention can reduce the complexity of the system under the condition of unchanged input channel number by a double-layer topological structure; the transmission bandwidth and the transmission power consumption can be adjusted according to real-time transmission requirements through small circle data transmission and large circle data transmission, and the condition of transmission blockage or advance caused by different transmission delays of each link is effectively prevented through a request frame sending mechanism and different from a distributed module spontaneous data return.
Drawings
FIG. 1 is a schematic of a topology;
FIG. 2 is a schematic diagram of small-turn data transmission;
FIG. 3 is a schematic diagram of a first clockwise transmission and a counterclockwise transmission of a large circle of data;
FIG. 4 is a schematic structural diagram of a distributed module;
FIG. 5 is a distributed module DU0,0The back-transmission flow logic diagram of the related result of (1);
FIG. 6 is a distributed module non-DU0,0The back-transmission flow logic diagram of the related result of (1);
FIG. 7 is a schematic diagram of a data cache unit structure inside a distributed module;
FIG. 8 is a schematic diagram of the structure of the cross-correlator internal complex correlation unit array and the two-stage accumulator of the correlation processing module inside the distributed module;
FIG. 9 is a schematic diagram of an implementation structure of a two-stage accumulator of a cross-correlator of a correlation processing module inside a distributed module;
fig. 10 is a schematic diagram of the overall cross-correlator structure of the correlation processing module inside the distributed module.
Detailed Description
As shown in fig. 1, in this embodiment, taking the topology architecture of H _4_5 as an example, 20 distributed modules are connected by a double-layer topology architecture, and the number of sampling channels integrated by each distributed module is 32, so that radiometric imaging with 640 acquisition channels can be performed under the topology architecture.
The bottom layer full interconnection layer in the double-layer topological structure is in full interconnection connection with 4 distributed modules as a group, the distributed modules at the same position of the full interconnection layer form a top layer ring structure in a top layer ring connection mode, and the H _4_5 topological structure only needs 50 links, so that the complexity of the system is greatly reduced.
As shown in FIG. 2, { DU) in H _4_5 topologyt,0|t∈[0,3]Giving DU as an example of a full interconnect layer0,0The small circle transmission required to be carried out comprises the following steps:
A1. distributed module DU0,0Collecting original data frame P0,0
B1.DU0,0According to the time slot number t epsilon [0,3]Collected data P0,0Divided into 4 time slots p0,0 0,p0,0 1,p0,0 2,p0,0 3
C1.DU0,0Data p identical to its own time slot number t equal to 00,0 0Leaving the remaining 3 slots of data p0,0 1,p0,0 2,p0,0 3Respectively transmitted to another 3 distributed modules DU corresponding to time slot numbers1,0,DU2,0,DU3,0In (1).
D1. Simultaneous DU0,0Receiving the same time slot number t-0 from the same full interconnectionData p of the other 3 distributed modules of the layer1,0 0,p2,0 0,p3,0 0Finally form data P after small circle transmission0 0＝{p0,0 0,p1,0 0,p2,0 0,p3,0 0}。
As shown in FIG. 3, { DU) in H _4_5 topology0,i|i∈[0,4]The ring structure of DU is given as an example0,0The large circle transmission required to be carried out comprises the following steps:
A2. will DU0,0Data P received over small turns0 0Clockwise transmission to DU0,4And is transmitted counterclockwise to DU0,1。
B2. Simultaneous DU0,0Receiving a message from a DU0,4Small circle of data P transmitted in counter-clockwise4 0And from DU0,1Clockwise transmitted data P after small circle transmission1 0。
For the topology structure of H _4_5, the number of data transmission circles, that is, the number of clockwise and counterclockwise alternate data migration in the circle, is all the same
Figure BDA0002625212480000051
Next, the process is carried out. The above DU0,0The data migration of 1 time of clockwise data migration and 1 time of anticlockwise data migration are completed, and through 2 times of data migration, the data collected by different distributed modules can be guaranteed to complete pairwise surface collision so as to complete required related operation.
Obviously, the small circle data transmission and the large circle data transmission shown in fig. 2 and fig. 3 can be performed simultaneously, that is, while the small circle transmits the collected ith frame data, the large circle can transmit the ith-1 frame data which is transmitted by the small circle before, thereby completing the pipeline transmission.
After the small circle data transmission and the large circle data transmission as shown in fig. 2 and 3, the distributed module DU is configured0,0Can obtain P in turn0 0,P4 0,P1 0Can correspondingly obtain
Figure BDA0002625212480000052
The correlation result of (1).
In order to ensure the real-time property of transmission under the double-layer topological framework, when the sampling rate of the ADC of the system is fsHz, sampling bit width Wadcbit, when the number of sampling channels integrated by each distributed module is 32, and under the topological architecture of H _4_5, the required bidirectional bandwidth of each link of the full interconnection layer is BWL1＝16fsWadcThe required bidirectional bandwidth per link of the ring structure is BWL2＝64fsWadcThe total transmission bandwidth of the system required thereby is BWtotal＝1760fsWadc。
The transmission power consumption required by the system can be calculated according to the real-time transmission bandwidth, the power consumption without a transmission interface consists of static power consumption and dynamic power consumption, and Ptrans＝P0+PdBW, wherein: p0For the static power consumption of the interface, which remains unchanged during transmission, the static power consumption of different types of transmission interfaces is different, PdThe dynamic power consumption of each bit is proportional to the bandwidth of transmission, and the dynamic power consumption of different types of transmission structures is different. The total transmission power consumption required by the system is related to the bandwidth and the number of links of the full interconnection layer and the bandwidth and the number of links of the ring structure, and taking the topological structure of H _4_5 as an example, P can be obtainedtrans-total＝
Figure BDA0002625212480000061
Since the transmission bandwidth required by the links of the full interconnection layer is much smaller than that required by the links of the ring structure, the links of the full interconnection layer can be specified to use copper cables, the links of the ring structure use optical fibers, and have a typical value P0L1＝0.5W,PdL1＝10pJ/bit,P0L2＝2W,PdL140 pJ/bit. Suppose again fs＝1GHz,WadcThe total transmission power consumption P can be calculated as 1bittrans-totalH _4_5 | 222 watts.
After the topological structure is changed into H _2_10, the total transmission power consumption P is obtained through recalculationtrans-total|H_2_10＝352.4 watts, the total transmission power consumption is increased compared to H _4_ 5. However, the number of links of H _2_10 is only 30, which is less than 50 of the number of links required by H _4_ 5. The topological structure can be changed into a typical full interconnection structure, and the total transmission power consumption P is obtained through recalculationtrans-totalThe total interconnect power consumption is 202.16 watts, which is slightly less than the total transmission power consumption of H _4_5, but the number of links in the full interconnect structure is 190, which is much larger than the number of links in H _4_ 5. In contrast, the topology is changed to the ring structure H _1_20, and the total transmission power P is obtained by recalculationtrans-totalI H _1_20 is 592 watts, which is much larger than the total transmission power consumption of H _4_5, and the number of links of H _1_20 is 20, which is slightly smaller than the number of links of H _4_ 5.
It can be seen that there is a trade-off between transmission power consumption and system complexity when the number of distributed modules is determined. The transmission power and the system complexity can be adjusted by adjusting L1 and L2 in the system structure, and a system which best meets the design physical requirements can be found. Meanwhile, the transmission power consumption of the double-layer structure is superior to that of a ring structure which is directly used.
As shown in fig. 4, the distributed module includes: AD sampling unit, inter-board synchronization unit, quadrature down conversion unit, data transmission unit, data buffer unit and relevant arithmetic unit, wherein: the AD sampling unit samples an externally input intermediate frequency analog signal into a digital signal through an ADC chip and transmits the digital signal to the inter-board synchronization unit; the inter-board synchronization unit carries out synchronization processing on the digital signals and transmits the synchronized data to the orthogonal down-conversion unit; the orthogonal down-conversion unit carries out orthogonal filtering on the digital signal to obtain an in-phase signal and an orthogonal signal and transmits the in-phase signal and the orthogonal signal to the data transmission unit; data transmission units of adjacent distributed modules in a bottom layer full interconnection layer and a top layer ring structure are respectively connected with each other to realize small-circle data transmission and large-circle data transmission, a data cache unit is connected with the data transmission units and transmits in-phase orthogonal signal data obtained by an orthogonal down-conversion unit and data obtained by small-circle data transmission, a correlation operation unit is connected with the data transmission units and transmits a complex correlation operation result and a control signal, and the data transmission units ensure that the data can be orderly transmitted back through a request frame sending mechanism.
The distributed modules are further provided with interaction units connected with the data transmission units, serve as connection points of the whole double-layer topology framework and the data post-processing module, receive related operation results of all other distributed modules and send the results to the data post-processing module, and meanwhile receive control instructions sent from the outside and forward the control instructions to other DU modules.
The data transmission unit transmits the data to different modules according to different receiving data sources, and specifically includes:
receiving data of same time slot number p from bottom full interconnection layert′,i tIf t 'belongs to N1 and t' is not equal to t }, directly sending the data to a data cache unit for merging;
when receiving the signal from the orthogonal down-conversion unit, the signal is processed and divided into L1One-slot data frame { pt,i t′If t ' belongs to N1 and t ' is not equal to t }, sending the data to a distributed module corresponding to the time slot number t ', thereby completing data transmission of a small circle;
thirdly, when receiving the data P obtained by the small-circle data transmission transmitted from the data buffer uniti tThen the data is transmitted to the DUs respectivelyt,i+1And DUt,i-1；
Fourthly, when the data which is sent by the data transmission unit of the adjacent distributed module in the top ring structure and is transmitted in a small circle is received, the data is Pj tJ ≠ i, and the sender is DUt,i+1Then the data is transmitted to the DUt,i-1When the sender is DUt,i-1Then the data is transmitted to the DUt,i+1Therefore, full duplex work of a transmission interface of the distributed module is met, and large circle data transmission is completed;
when receiving a Request frame (Request) sent by the data post-processing module or forwarded by the data transmission unit of the adjacent distributed module, reading the operation result of the relevant processing module, and transmitting the operation result back to the distributed module DU connected with the data post-processing module0,0Completing the return of the related operation result;
when receiving the control instruction sent by the data post-processing module, executing or forwarding the control instruction to other distributed modules according to the instruction to finish control information instruction forwarding.
As shown in fig. 5 and fig. 6, the request frame uses a specific sending mechanism to ensure that the core distributed module reads the correlation operation results of all the distributed modules in sequence, and then outputs the correlation operation results to the data post-processing module to complete the return of the correlation operation results, specifically: the destination distributed module periodically sends request frames to ask for correlation operation results until data is received, when the module receiving the request frames is not ready for the correlation operation results needing to be transmitted, the sent request frames are directly discarded, otherwise: i) the distributed module adjacent to the core distributed module directly transmits the related processing operation result or ii) the distributed module not adjacent to the core distributed module transfers the related operation result and forwards the request frame through connecting the intermediate node.
Still taking the topology architecture of H _4_5 as an example, the distributed module DU0,0Sequentially receiving data from DU through Request frame1,0,DU2,0,DU3,0,DU0,1Transmitted data, and DU1,0,DU2,0,DU3,0,DU0,1As non-DUs0,0The distributed modules of (1) respectively receive the data from the DUs only through the Request frame1,1,DU2,1,DU3,1,DU0,2Data of (2), i.e. data of the next number of the corresponding ring structure, to be DU-0,0DU after reading1,0,DU2,0,DU3,0,DU0,1Upon the result of the transmission, DU1,0,DU2,0,DU3,0,DU0,1Also at the same time receive DU1,1,DU2,1,DU3,1,DU0,2Result of the transmission, hence DU0,0Can continue by reading the DU1,0,DU2,0,DU3,0,DU0,1To receive data from DU1,1,DU2,1,DU3,1,DU0,2By analogy, the result of (D) is to complete DU0,0And (5) utilizing a request frame sending mechanism to sequentially read the returned related operation results.
As shown in fig. 7, the data cache unit further includes: a data buffer (data buffer) and a two-stage FIFO, wherein: the data buffer (data buffer) receives and combines the data frames with the same time slot number of the same full interconnection layer after the small circle transmission sent by the data transmission unit into Pi tThen merging the merged data Pi tSending the data to a data transmission unit for subsequent large circle data transmission, and simultaneously combining the data Pi tSending the data to a correlation processing module for subsequent correlation operation; two-stage FIFO sequential storage data frame of large circle transmission adjacent distributed module sent by data transmission unit
Figure BDA0002625212480000081
Figure BDA0002625212480000082
Each frame of data is stored in FIFO1 upon receipt, and the data received from the previous frame is stored in sequential order in FIFO2 above FIFO1, noting that when FIFO2 just received P in the data bufferi tStoring data frames P from a data buffer on demandi tSo that the data stored in the FIFO1 and the FIFO2 can be input to the relevant processing modules in time sequence.
As shown in fig. 8, 9 and 10, the correlation operation unit further includes an autocorrelator and a cross correlator, wherein: the autocorrelator calculates the correlation operation of the data frames of the same full interconnection layer with the same time slot number after small circle transmission sent by the data buffer of the data cache unit, and the cross correlator calculates the correlation operation of the data frames of the adjacent distributed modules of large circle transmission sent by the two-stage FIFO of the data cache unit.
The autocorrelator inputs the data packets input into the autocorrelator into two input ports of a correlation operation matrix, and the correlation operation matrix consists of a plurality of complex multiplication accumulators, so that a plurality of correlation operation results can be output in parallel.
The cross correlator further comprises: input data buffer, complex correlation unit array, second grade accumulator unit, wherein: after the input data buffer area is positioned after the input data and before the complex correlation unit array, the input data buffer area is utilized to ensure that time points corresponding to the input data received by the complex correlation units positioned at different positions in the array are the same; the complex correlation unit array is an integral structure of the whole cross-correlator, and the complex correlation units are arranged according to a matrix form to complete correlation operation among a plurality of channels in parallel; the second-stage accumulator unit reads the output result of the multiple correlation units within a certain range through time division multiplexing, further accumulates the output result and stores the accumulated result in the RAM with the corresponding size, so that the accumulation duration of correlation operation is prolonged, and finally, the correlation operation result is transmitted to the data transmission unit for data return.
The complex correlation unit further comprises: a register and Multiply Accumulator (MAC), wherein: the register plays a role of transmitting input data to other complex correlation units, and the delay of the input data caused by the introduction of the register is compensated by the input data buffer area so as to ensure the synchronism of the data received by each complex correlation unit; the multiply-accumulate unit (MAC) performs multiplication and addition between the data of one channel output from the FIFO1 and the data of one channel output from the FIFO2, wherein the number of times of addition is small, and preferably, the multiplication can be performed using a look-up table (LUT table) because the number of input bits is low.
Through specific practical experiments, under the specific environment setting that an FPGA with an XCVU095 model is used for realizing distributed module internal units, a copper cable is used as a bottom layer full interconnection layer link, an optical cable is used as a top layer ring structure link, 20 distributed modules are used in total, each distributed module samples signals of 32 channels, the sampling rate of each channel is 1GHz, and the method is started by using parameters with the sampling bit width of 1bit, so that the obtained experimental data are as follows: the real-time processing capacity of each distributed module, namely the calculable correlation operation times per second can reach 4.096Tops, the real-time processing capacity of the double-layer topological structure of the whole radiometer imaging system can reach 81.92Tops, and the throughput rate can reach 100% due to the design of the signal processing mode of the flowing water.
Compared with the prior art, the invention has the remarkable improvement points that:
first, no clogging, no redundancy: because the correlation operation is essentially to realize "two-to-two meeting" of data transmission between distributed nodes, neither blocking (i.e., no "two-to-two meeting") nor redundancy (e.g., the same data meets at different nodes twice, or the transmission line has idle time without data transmission within a certain period of time) can occur. Therefore, the invention ensures 'two-by-two meeting' of data transmission through a transmission closed loop, and transmission redundancy does not occur. Thereby reducing transmission power consumption.
Secondly, the transmission load is balanced: in the double-layer topological structure and the transmission mode, the data transmission quantity of the outer-layer data lines is balanced, and the data transmission quantity of the inner-layer data lines is also balanced. I.e. to equalize the total amount of data transmission to each data line. Thereby achieving the goal of reducing the transmission load on a single line.
Third, the adjustable implementation: because the data transmission bandwidth is in direct proportion to the power consumption, the data transmission quantity of the outer layer data line is large, and the data transmission can be realized by adopting high-power-consumption modes such as optical fibers and the like; and the data transmission quantity of the inner layer data line is small, and the data transmission quantity can be realized by adopting a mode with low power consumption, such as a coaxial cable.
Fourthly, the internal processing mode of the distributed module is as follows: the autocorrelation and cross-correlation operation among all channels in the distributed modules are actually a triangular matrix and a square matrix, the overall calculation requirement of the double-layer framework is a large triangular matrix, the calculated amount of each distributed module is balanced as much as possible through the design of the pulse array, and the calculation redundancy of the system is reduced to the minimum.
The foregoing embodiments may be modified in many different ways by those skilled in the art without departing from the spirit and scope of the invention, which is defined by the appended claims and all changes that come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Claims (13)

1. An information transmission method of a double-layer topological structure of a multi-channel radiometer imaging system is characterized in that a distributed module is realized in a mode of bottom layer full interconnection connection and top layer ring connection, wherein: each distributed module carries out signal preprocessing, data transmission and sending and complex correlation operation processing according to the acquired digital signal information of a plurality of channels and outputs complex correlation operation result information, one core distributed module in the double-layer topological framework collects the correlation operation results of all other distributed modules and outputs the correlation operation results to a subsequent module for image inversion, and meanwhile, the core distributed module receives external control signals and forwards the control signals to other distributed modules;
the bottom layer full interconnection connection refers to that: all distributed modules in the bottom layer are connected by links, and the link set is Ef(ii) a The distributed modules in the bottom layer transmit data acquired by the distributed modules and subjected to time slot division in a small circle data transmission mode;
the top ring-shaped connection means that: all distributed modules in the top layer are connected end to end in a ring link mode, and the link set is Ec(ii) a Data received by each distributed module through small circle data transmission is transmitted among all distributed modules in the top layer in a large circle data transmission mode, so that the data of all distributed modules can meet each other;
the distributed module comprises: AD sampling unit, inter-board synchronization unit, quadrature down conversion unit, data transmission unit, data buffer unit and relevant arithmetic unit, wherein: the AD sampling unit samples an externally input intermediate frequency analog signal into a digital signal through an ADC chip and transmits the digital signal to the inter-board synchronization unit; the inter-board synchronization unit carries out synchronization processing on the digital signals and transmits the synchronized data to the orthogonal down-conversion unit; the orthogonal down-conversion unit carries out orthogonal filtering on the digital signal to obtain an in-phase signal and an orthogonal signal and transmits the in-phase signal and the orthogonal signal to the data transmission unit; data transmission units of adjacent distributed modules in a bottom layer full interconnection layer and a top layer ring structure are respectively connected with each other to realize small-circle data transmission and large-circle data transmission, a data cache unit is connected with the data transmission units and transmits in-phase orthogonal signal data obtained by an orthogonal down-conversion unit and data obtained by small-circle data transmission, a correlation operation unit is connected with the data transmission units and transmits a complex correlation operation result and a control signal, and the data transmission units ensure that the data can be orderly transmitted back through a request frame sending mechanism;
the small circle data transmission and the large circle data transmission are carried out simultaneously, and the two data transmission modes are coupled through a data cache unit arranged in the distributed module.
2. The information transmission method according to claim 1, wherein the two-layer topology is: the distributed modules, namely the vertex set and the link set between the vertices, are in one-to-one correspondence, specifically: h _ L1_ L2 is G ═ V, E, and the set of vertices is V ═ DU, { DU, }t,iI t ∈ N1, i ∈ N2}, where: n1 { [ N | N ∈ N and N<L1}, N2 ═ N | N ∈ N and N<L2, t represents the number of the distributed module in the top ring structure, i represents the number of the distributed module in the bottom full interconnect layer; set of links is E ═ Ef,Ec}, wherein: ef＝{<DUt1,i,DUt2,i>L t1, t2 belongs to N1, i belongs to N2, t1 is not equal to t2 is a link set of a bottom layer full interconnection layer,
Figure FDA0002625212470000011
a link set of a top ring structure;
the number of the links of the double-layer topological structure is
Figure FDA0002625212470000021
Wherein: when the L1 is reduced to 1, the topological structure is degenerated into a ring, when the L2 is reduced to 1, the topological structure is degenerated into a classic full interconnection structure, and the L2 is recommended in a general practical system>2to characterize a ring-type structure.
3. The information transmission method as claimed in claim 1, wherein said small data transmission is:
s1 all vertex DUs of topological structuret,iThe corresponding original data frame P is acquired after the time Tt,i；
S2 reaction of Pt,iEqually dividing the data into L1 data frames according to the sequence of acquisition, and respectively recording the data frames as pt,i 0,pt,i 1,pt,i 2,...pt,i L1-1The upper corner mark can be one-to-one corresponding to the top point number of the bottom full interconnection layer, i.e. Pt,i＝{pt,i t′|t′∈N1}；
S3:DUt,iWill L this1L in a slotted data frame11 other modules addressed to the corresponding underlying full interconnect layer number, leaving p alonet,i tThe transmission process is marked as
Figure FDA0002625212470000022
Simultaneous DUt,iReceiving from L11 data frame with time slot number t sent by other modules, namely:
Figure FDA0002625212470000023
after a small circle of data transmission, all vertex DUst,iGet DU from self and same top ring structure number it′,iThe collection of data frames, i.e. P, having a time-slot number ti t，Pi t＝{p0,i t,p1,i t,p2,i t,...pL1-1,i t}。
4. The information transmission method according to claim 3, wherein the data transmission of the great circle is:
s4: data frame { P) obtained by small circle transmissioni tI ∈ N2} totals the ring structure's neighboring vertices with the same underlying full interconnect layer number t
Figure FDA0002625212470000025
Clockwise and counterclockwise alternating data migration in secondary ring, DUt,iIs DUt,inAnd DUt,ipWherein: subscript in ═ i +1) mod L2, ip ═ i + L2-1) mod L2, i ∈ N2, and for brevity, DU may be usedt,i+1And DUt,i-1Replacing DUt,inAnd DUt,ipTo represent DUt,iDefining the clockwise data direction as DU for the next and previous nodes in the ringt,iFrom DUt,i-1Receive data and forward to DUt,i+1Transmitting data with DU in anti-clockwise directiont,iFrom DUt,i+1Receive data and forward to DUt,i-1Sending data, if the first transmission direction is not clockwise, the data transmission is as follows:
Figure FDA0002625212470000024
5. the information transmission method according to claim 4, wherein the number of transitions N in step S4transWhen the number of the data is odd, the direction of the first data migration is changed into anticlockwise when the large-circle data transmission of the next frame data is carried out, and at the moment, the positive and negative signs in the data transmission expression are just required to be totally reversed, so that the full-duplex bidirectional bandwidth of the ring-shaped structure link is fully utilized; number of migration NtransAnd when the number of the data is even, the problem of changing the direction of the next frame data large circle transmission in the data migration direction does not need to be considered.
6. The information transmission method as claimed in claim 1, wherein the signal is preprocessed at a sampling rate fs(Hz) and a sampling bit width Wadc(bit), when the number of sampling channels integrated by each distributed module is m, the required bidirectional bandwidth of each link of the full interconnection layer is
Figure FDA0002625212470000031
The required bidirectional bandwidth per link of the ring structure is
Figure FDA0002625212470000032
Thereby corresponding to a required total transmission bandwidth of the system of
Figure FDA0002625212470000033
Figure FDA0002625212470000034
Because the bandwidth requirements of the two links are mutually decoupled, when a real-time system needs to expand the scale by increasing the number of L2, only the bandwidth requirement of the ring link needs to be considered, and the bandwidth requirement of the full interconnection link does not need to be considered, so that the scale expandability of the topological structure is embodied.
7. The information transmission method according to claim 6, wherein the transmission power consumption, the transmission bandwidth and the number of links of the small-circle data transmission and the large-circle data transmission are optimized by adjusting values of L1 and L2, specifically: the power consumption of each transmission interface of the bottom full interconnection layer is Ptrans_L1＝P0_L1+Pd_L1·BWL1The power consumption of each transmission interface of the top ring structure is Ptrans_L2＝P0_L2+Pd_L2·BWL2Total transmission power consumption of
Figure FDA0002625212470000035
Wherein: the power consumption of the transmission interface consists of static power consumption and dynamic power consumption, Ptrans＝P0+Pd·BW，P0For static power consumption, corresponds to P0_L1Static power consumption, P, for each transmission interface of the bottom fully interconnected layer0_L2For static power consumption of each transmission interface of the top ring structure, the static power consumption remains unchanged during transmission, PdFor dynamic power consumption per bit, corresponding to Pd_L1Dynamic power consumption per bit, P, for each transmission interface of the underlying fully interconnected layerd_L2For each bit dynamic power consumption of each transmission interface of the top ring structure, the dynamic power consumption is in direct proportion to the bandwidth BW of the transmission of the interface, and the corresponding BWL1Transmission bandwidth, BW, for each transmission interface of the underlying fully interconnected layerL2The factor of 2 in the total transmission power consumption is the transmission bandwidth of each transmission interface of the top ring structureEach link corresponds to two transmission interfaces.
8. The information transmission method according to claim 1, wherein the data transmission unit transmits to different modules according to different sources of the received data, and specifically comprises:
receiving data of same time slot number p from bottom full interconnection layert′,i tIf t 'belongs to N1 and t' is not equal to t }, directly sending the data to a data cache unit for merging;
when receiving the signal from the orthogonal down-conversion unit, the signal is processed and divided into L1One-slot data frame { pt,i t′If t ' belongs to N1 and t ' is not equal to t }, sending the data to a distributed module corresponding to the time slot number t ', thereby completing data transmission of a small circle;
thirdly, when receiving the small circle data from the data cache unit, the small circle data is Pi tThen the data is transmitted to the DUs respectivelyt,i+1And DUt,i-1；
Fourthly, when the small circle of data sent by the data transmission unit from the adjacent distributed module in the top ring structure is received, the small circle of data is Pj tJ ≠ i, and the sender is DUt,i+1Then the data is transmitted to the DUt,i-1When the sender is DUt,i-1Then the data is transmitted to the DUt,i+1Therefore, full duplex work of a transmission interface of the distributed module is met, and large circle data transmission is completed;
when receiving the request frame sent by the data post-processing module or forwarded by the data transmission unit of the adjacent distributed module, reading the operation result of the relevant processing module and transmitting the operation result back to the distributed module DU connected with the data post-processing module0,0Completing the return of the related operation result;
when receiving the control instruction sent by the data post-processing module, executing or forwarding the control instruction to other distributed modules according to the instruction to finish control information instruction forwarding.
9. The information transmission method according to claim 8, wherein the request frame uses a specific sending mechanism to ensure that the core distributed module reads the correlation operation results of all the distributed modules in sequence, and outputs the correlation operation results to the data post-processing module to complete the return of the correlation operation results, and specifically comprises: the destination distributed module periodically sends request frames to ask for correlation operation results until data is received, when the module receiving the request frames is not ready for the correlation operation results needing to be transmitted, the sent request frames are directly discarded, otherwise: i) the distributed module adjacent to the core distributed module directly transmits the related processing operation result or ii) the distributed module not adjacent to the core distributed module transfers the related operation result and forwards the request frame through connecting the intermediate node.
10. The information transmission method according to claim 1, wherein the data buffer unit further comprises: a data buffer (data buffer) and a two-stage FIFO, wherein: the data buffer (data buffer) receives and combines the data frames with the same time slot number of the same full interconnection layer after the small circle transmission sent by the data transmission unit into Pi tThen merging the merged data Pi tSending the data to a data transmission unit for subsequent large circle data transmission, and simultaneously combining the data Pi tSending the data to a correlation processing module for subsequent correlation operation; two-stage FIFO sequential storage data frame of large circle transmission adjacent distributed module sent by data transmission unit
Figure FDA0002625212470000041
Each frame of data is stored in FIFO1 upon receipt, and the data received from the previous frame is stored in sequential order in FIFO2 above FIFO1, noting that when FIFO2 just received P in the data bufferi tStoring data frames P from a data buffer on demandi tSo that the data stored in the FIFO1 and the FIFO2 can be input to the relevant processing modules in time sequence.
11. The information transmission method according to claim 1, wherein the correlation operation unit further comprises an autocorrelator and a cross correlator, wherein: the autocorrelator calculates the correlation operation of the data frames of the same full interconnection layer with the same time slot number after the small circle of transmission sent by the data buffer of the data cache unit, and the cross correlator calculates the correlation operation of the data frames of the adjacent distributed modules of the large circle of transmission sent by the two-stage FIFO of the data cache unit;
the autocorrelator inputs the data packets input into the autocorrelator into two input ports of a correlation operation matrix, and the correlation operation matrix consists of a plurality of complex multiplication accumulators, so that a plurality of correlation operation results can be output in parallel.
12. The information transmission method of claim 11, wherein said cross-correlator further comprises: input data buffer, complex correlation unit array, second grade accumulator unit, wherein: after the input data buffer area is positioned after the input data and before the complex correlation unit array, the input data buffer area is utilized to ensure that time points corresponding to the input data received by the complex correlation units positioned at different positions in the array are the same; the complex correlation unit array is an integral structure of the whole cross-correlator, and the complex correlation units are arranged according to a matrix form to complete correlation operation among a plurality of channels in parallel; the second-stage accumulator unit reads the output result of the multiple correlation units within a certain range through time division multiplexing, further accumulates the output result and stores the accumulated result in the RAM with the corresponding size, so that the accumulation duration of correlation operation is prolonged, and finally, the correlation operation result is transmitted to the data transmission unit for data return.
13. The information transmission method as claimed in claim 1, wherein said complex correlation unit further comprises: a register and Multiply Accumulator (MAC), wherein: the register plays a role of transmitting input data to other complex correlation units, and the delay of the input data caused by the introduction of the register is compensated by the input data buffer area so as to ensure the synchronism of the data received by each complex correlation unit; the multiply-and-accumulate (MAC) unit performs multiplication and addition between the data of one lane output from the FIFO1 and the data of one lane output from the FIFO2, where the number of times of addition is small.
CN202010794926.8A 2020-08-10 2020-08-10 Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system Active CN112242942B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010794926.8A CN112242942B (en) 2020-08-10 2020-08-10 Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010794926.8A CN112242942B (en) 2020-08-10 2020-08-10 Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system

Publications (2)

Publication Number Publication Date
CN112242942A true CN112242942A (en) 2021-01-19
CN112242942B CN112242942B (en) 2021-07-06

Family

ID=74171622

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010794926.8A Active CN112242942B (en) 2020-08-10 2020-08-10 Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system

Country Status (1)

Country Link
CN (1) CN112242942B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113189408A (en) * 2021-04-30 2021-07-30 上海航天测控通信研究所 Related matrix processing system capable of realizing super-large-scale operation through rapid expansion

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050122254A1 (en) * 2003-09-18 2005-06-09 Xytrans, Inc. Multi-channel radiometer imaging system
CN102684831A (en) * 2012-03-21 2012-09-19 上海交通大学 Digital multichannel correlated processing system and output method for buffer module in same
CN103674250A (en) * 2013-12-16 2014-03-26 中国电子科技集团公司第四十一研究所 High-accuracy middle-infrared-band absolute spectral responsivity calibrating device
CN106097333A (en) * 2016-06-08 2016-11-09 上海交通大学 Microwave holography map generalization and reconstructing method and system thereof
US20180019594A1 (en) * 2016-07-12 2018-01-18 The Research Foundation For The State University Of New York Photovoltaic system power output control with retained reserve power output
CN108107061A (en) * 2017-12-20 2018-06-01 电子科技大学 A kind of passage of multi channel imaging radiometer improves system and method
CN108551378A (en) * 2018-02-08 2018-09-18 中国航天时代电子有限公司 A kind of radiofrequency signal remote transmission method
CN109946235A (en) * 2019-02-26 2019-06-28 南京信息工程大学 The multi layer cloud inversion method of wind and cloud 4A meteorological satellite Multichannel Scan Imagery Radiometer
CN110113242A (en) * 2019-05-07 2019-08-09 南京磐能电力科技股份有限公司 Multi-node synchronization sampling and data transmission method in ring-type communication network
CN110165422A (en) * 2019-06-21 2019-08-23 哈尔滨工业大学 Based on the passive millimeter wave near field imaging system for focusing super surface

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050122254A1 (en) * 2003-09-18 2005-06-09 Xytrans, Inc. Multi-channel radiometer imaging system
CN102684831A (en) * 2012-03-21 2012-09-19 上海交通大学 Digital multichannel correlated processing system and output method for buffer module in same
CN103674250A (en) * 2013-12-16 2014-03-26 中国电子科技集团公司第四十一研究所 High-accuracy middle-infrared-band absolute spectral responsivity calibrating device
CN106097333A (en) * 2016-06-08 2016-11-09 上海交通大学 Microwave holography map generalization and reconstructing method and system thereof
US20180019594A1 (en) * 2016-07-12 2018-01-18 The Research Foundation For The State University Of New York Photovoltaic system power output control with retained reserve power output
CN108107061A (en) * 2017-12-20 2018-06-01 电子科技大学 A kind of passage of multi channel imaging radiometer improves system and method
CN108551378A (en) * 2018-02-08 2018-09-18 中国航天时代电子有限公司 A kind of radiofrequency signal remote transmission method
CN109946235A (en) * 2019-02-26 2019-06-28 南京信息工程大学 The multi layer cloud inversion method of wind and cloud 4A meteorological satellite Multichannel Scan Imagery Radiometer
CN110113242A (en) * 2019-05-07 2019-08-09 南京磐能电力科技股份有限公司 Multi-node synchronization sampling and data transmission method in ring-type communication network
CN110165422A (en) * 2019-06-21 2019-08-23 哈尔滨工业大学 Based on the passive millimeter wave near field imaging system for focusing super surface

Non-Patent Citations (7)

* Cited by examiner, † Cited by third party
Title
B. FENG ET AL.: "Joint Millimeter Wave and Microwave Wave Resource Allocation Design for Dual-Mode Base Stations", 《2019 IEEE WIRELESS COMMUNICATIONS AND NETWORKING CONFERENCE (WCNC)》 *
J. R. PIEPMEIER ET AL.: "Multi-Channel Correlator array-fed Microwave Radiometer", 《IGARSS 2019 - 2019 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM》 *
J. RICHTER: "A Multi-Channel Radiometer with Focal Plane Array Antenna for W-Band Passive Millimeterwave Imaging", 《2006 IEEE MTT-S INTERNATIONAL MICROWAVE SYMPOSIUM DIGEST》 *
LING X: "An Area Efficient Real-Time PFFT Architecture Using Parallel Distributed Arithmetic", 《IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS & COMPUTER SCIENCES》 *
宫新保: "RBF网络雷达天线扫描方式识别系统", 《红外与激光工程》 *
王淦泉: "风云四号气象卫星多通道扫描成像辐射计", 《中国光学学会2004年学术大会》 *
陈博洋: "FY-4A成多通道扫描成像辐射计性能分析", 《第34届中国气象学会年会》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113189408A (en) * 2021-04-30 2021-07-30 上海航天测控通信研究所 Related matrix processing system capable of realizing super-large-scale operation through rapid expansion
CN113189408B (en) * 2021-04-30 2023-01-20 上海航天测控通信研究所 Related matrix processing system capable of realizing super-large-scale operation through rapid expansion

Also Published As

Publication number Publication date
CN112242942B (en) 2021-07-06

Similar Documents

Publication Publication Date Title
CN106294234B (en) A kind of data transmission method and device
KR102305470B1 (en) Image signal processing device performing image signal processing in parallel through plurality of image processing channels
JPH09247176A (en) Asynchronous transfer mode exchange system
KR101678916B1 (en) High-speed data bus system and image sensor using the same
CN113807509B (en) Neural network acceleration device, method and communication equipment
CN112242942B (en) Information transmission method of double-layer topological architecture of multi-channel radiometer imaging system
KR101340763B1 (en) Interconnects using self-timed time-division multiplexed bus
CA2325091C (en) Bus switch and bus switch system
CN101286226B (en) Diminished image digital filtering method and device
CN101697486A (en) Two-dimensional wavelet transformation integrated circuit structure
CN102025391A (en) High-speed filtering device and method for implementing multiple spread spectrum code lengths
CN101569181B (en) Solid state image picking-up device and image picking-up signal output circuit
CN201111042Y (en) Two-dimension wavelet transform integrate circuit structure
US6567909B2 (en) Parallel processor system
US11308388B2 (en) Electronic circuit, particularly for the implementation of neural networks with multiple levels of precision
CN111626410B (en) Sparse convolutional neural network accelerator and calculation method
CN113537482B (en) Neural network computing module, method and communication device
CN116974510A (en) Data stream processing circuit, circuit module, electronic chip, method and device
CN113189408B (en) Related matrix processing system capable of realizing super-large-scale operation through rapid expansion
US9449597B2 (en) Ultrasound receiving module, method and system
KR100577005B1 (en) Apparatus and method of selectively performing fast hadamard transform and fast fourier transform
EP2199919A1 (en) Method for processing data using triple buffering
CN108055484B (en) Sequence conversion module, high dynamic range image sequence conversion module and conversion method
CN112766453A (en) Data processing device and data processing method
CN109687883B (en) Channel comprehensive engineering implementation method and device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant