CN102306213B - Anti-single particle irradiating method and anti-single particle irradiating system based on frame data processing - Google Patents
Anti-single particle irradiating method and anti-single particle irradiating system based on frame data processing Download PDFInfo
- Publication number
- CN102306213B CN102306213B CN 201110202501 CN201110202501A CN102306213B CN 102306213 B CN102306213 B CN 102306213B CN 201110202501 CN201110202501 CN 201110202501 CN 201110202501 A CN201110202501 A CN 201110202501A CN 102306213 B CN102306213 B CN 102306213B
- Authority
- CN
- China
- Prior art keywords
- frame data
- frame
- single particle
- bit
- data processing
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Detection And Correction Of Errors (AREA)
Abstract
The invention discloses an anti-single particle irradiating method and an anti-single particle irradiating system based on frame data processing, which mainly solves the problem of function error caused by single-particle irradiation of a traditional frame data processing system. The anti-single particle irradiating system comprises a single frame detecting and restoring module and a frame data processor. The single frame detecting and restoring module is used for restoring the frame data processor. The frame data processor comprises a finite state machine collection unit, a controller protecting and correcting unit, a memory, a memory protecting and correcting unit, a data input interface, a data output interface and a configuration module, wherein the finite state machine collection unit is used for processing various frame data; the controller protecting and correcting unit is used for protecting the finite state machine collection unit; the memory protecting and correcting unit is used for protecting memory data; and the configuration module is used for storing parameter information required by the frame data processor. The invention has the advantages of simple realization, anti-single particle irradiation and the like, and can be used for image compression systems and other correcting and modulating circuits based on frames or groups in space environments.
Description
Technical field
The present invention relates to communication technical field, relate to anti-single particle radiation method for designing and system, be used for the image compression system under space environment and the single particle radiation, or other are based on error correction and the modulation circuit of frame or grouping.
Background technology
Along with the mankind march toward the information age, particularly recently for over ten years, multimedia signal processing technique and the communication technology have obtained fast development.Fusion between the two is irresistible, and greatly to information highway, backbone network, little multimedia recreation to family and individual, personal communication etc. are all had higher requirement to storage and the transmission of information data.The user wishes can both obtain by any way at any time and any place reliable and stable information and represents and transmit, thus digital signal processing technique, VLSI (very large scale integrated circuit) VLSI technology and modern digital communication are proposed huge challenge, also brought the flourish of above technology and related industry.Along with consumer electronics develops to portability and miniaturization, space flight and Military Application change to low storage low-power consumption, and the fast development of integrated circuit technique, and application-specific integrated circuit ASIC and SOC (system on a chip) SoC enter into the market and be ripe gradually.
Along with the development of microelectronic technique, the live width of chip internal constantly reduces, so the area of electron device is more and more little; But live width reduces the chips wire direction density is increased, the bolt-lock effect is more and more serious in cmos circuit.And, the paper Single Event Upset at Ground Level (Vo1.43 that delivers at periodical IEEE Trans.Nuclear Science in 1996 according to American Studies personnel Eugene Normand, No.6, Dec.1996, pp.2742-2750) point out in, integrated circuit and electronic equipment are all very responsive to irradiation in space environment and ground environment.For example communication and signal processing chip just must guarantee reliably working under space radiation environment if be applied to satellite system.Space irradiation mainly comes from cosmic rays, Van Allen radiation belt, solar flare, solar electromagnetic radiation and auroral radiation etc.; The high energy particle that irradiation in the ground environment mainly comes from atmospheric neutron, produced by deep space cosmic radiation source and radioactive material of adopting in the cosmic rays that forms and the circuit production run etc.
Effect of irradiation is discomposition and ionization in two kinds of main causes that electron device produces disabler.Discomposition is that if atom is arranged in crystal structure, this discomposition will change material behavior because thereby the particle hits atom changes its fixed position, and it is similar with it that particle injects the effect that semiconductor devices is produced.Ionization also can cause electric charge to move, and forms electron-hole pair, thereby produces internal electric field and internal current, and internal current may change the function of primary circuit.This effect can not produce damage to circuit, just causes circuit function to get muddled.
After electron device is subjected to effect of irradiation, will produce ionising radiation total dose effect, dose rate effect, single particle effect, displacement effect and high-pressure electrostatic discharge breakdown effect etc., that at present of greatest concern is ionising radiation total dose effect TID and single particle effect SEE.Adopt specific CMOS flow technique, can obtain in theory good ionising radiation accumulated dose index, and in experiment, obtain checking.Single particle effect mainly is divided into locking single particle SEL and single-particle inversion SEU.Locking single particle is that the low frequency channel place that produces between VDD-to-VSS has converged larger electric current, therefore circuit had inevitable destructiveness, and being accompanied by the continuous reduction of physical size, supply current and threshold voltage, the disabler that locking single particle brings constantly increases.Therefore in practical application, from the circuit design angle pin solution is analyzed and proposed to single-particle inversion problem the radiation environment, have very important value.
The state transition that single-particle inversion produces when being high energy charged particles bombardment device, Fig. 8 (a) is depicted as the ball bearing made using of single-bit storage unit, circuit among the figure remains on stable state, i.e. storage " 0 " or storage " 1 ", under each state, have two transistors to be in state of activation, and two other is in blocked state.Fig. 8 (b) has illustrated the single-bit storage unit by the process of a high energy charged particles bombardment and generation bit reversal, particle forms electron-hole pair in the path that punctures substrate, these electron-hole pairs are assembled at source electrode and drain electrode, similar with the effect of enough heavy current pulses, thereby generation and normal signal act on the dummy results of transistor equivalence, i.e. bit reversal.Except the storage unit effect is produced bit reversal, single-particle inversion also might cause the mistake of logical resource, mainly contains following several: (1) changes a bit among the look-up table LUT, thereby changes the combination logic function of realizing; (2) change the configuration of MUX MUX, thereby cause the signal can't correctly forward transmission in each logical block; (3) configuration of change trigger FF, thus the polarity that for example changes reseting signal line or clock line causes the circuit function mistake.
Used on-site programmable gate array FPGA device and special chip is based on random access memory ram mostly at present, the device of this integrated a large amount of storage unit is more responsive to single-particle inversion, and frequency and radiation environment that single-particle inversion occurs are in close relations.According to American Studies personnel Michael Wirthlin at paper The Reliability of FPGA Circuit Designs in the Presence of Radiation Induced Configuration Upsets (llth Annual IEEE Symposium on Field-Programmable Custom Computing Machines, 2003, pp.133-142) estimation in, for integrated 6,000,000 bit storage arrays, model is the FPGA of Xilinx Virtex 1000, single-particle inversion crest frequency under the mal-condition is up to per hour 4.2 times, considerably beyond during the stage of stable development per hour 0.13 time of solar activity.And along with charged particle with the difference of energy and occurrence frequency, the phenomenon that component failure shows is also different, when slight may only in single clock single bit upset occur, may occur then when serious that a plurality of bits continue or intermittent upset within certain period, even the function of whole device is collapsed when more abominable.
In the communication and signal processing system that relate to a plurality of fields such as civilian applications, space flight detection and Military Application, there are a lot of data processing units to be consisted of by frame structure, i.e. input and output are take frame as unit, and they are independently each other.For example for the image compression encoding processing unit that meets the JPEG2000 standard, respectively each frame data is compressed separately processing, output bit flow is made of the N frame, and every frame is made of frame head, frame data and postamble, and concrete form as shown in Figure 9; The Error Correction of Coding unit in the communication system and for example, for block coding method commonly used, such as Reed-Solomon RS code, low-density checksum LDPC code etc., the signal that is encoded also separates one by one.If processing unit is not considered a minute frame design, thereby causes the mistake of control function to go on making mistakes being corrected always by single-particle inversion.Therefore even carried out physical shielding and protection, not have to add the application under space radiation environment for example of processing unit that minute frame designs and still have potential risks, can't provide reliable and stable service for the user.
Summary of the invention
The analysis that the present invention is based on the above-mentioned single particle effect origin cause of formation and electron device is affected; characteristics in conjunction with the frame structure data processing unit; a kind of anti-single particle radiation method and anti-single particle radiation system that processes based on frame data is provided; be intended to from the circuit design angle downlink data upon handover processing unit be protected and error correction; reduce the impact of single particle effect, improve the reliability and stability of processing unit and whole system.
The key problem in technology of realizing the object of the invention is to set up rational single-particle inversion model; by the impact of software phantom analysis list particle effect on the downlink data upon handover processing unit; specifically be exactly to analyze bit reversal to occur in respectively frame head; in frame data and the postamble on the impact of result; on this basis from the circuit design angle to the single frame detection the hardware implementation structure and reseting module; controller circuitry; memory circuitry and configuration module are protected and error correction, so that the hardware configuration of downlink data upon handover processing unit has the characteristic to anti-single particle radiation within the specific limits.Its concrete scheme is described below:
One, the anti-single particle radiation method of processing based on frame data comprises the steps:
(1) mode by reading data produces each pending and separate frame data;
(2) mistake of simulation single particle radiation effect generation generates single particle radiation effect error listing;
(3) utilize single particle radiation effect error listing, correct frame data processing module built-in function is made amendment, form the erroneous frame data processing module;
(4) respectively each frame data is processed separately with correct frame data processing module and erroneous frame data processing module, obtained respectively correct result and the wrong result of each frame;
(5) contrast the correct result of each frame and the difference of wrong result, the mistake that analysis list particle irradiation effect error listing causes correct frame data processing module specifically appears at which part of frame head, frame data or the postamble of wrong result, and provides error reporting;
(6) part of makeing mistakes that error reporting is listed feeds back to the erroneous frame data processing module wrong function is wherein corrected, and obtains the result after this step corrects;
(7) repeating step (5)～(6) are until the result after correcting is identical with the result of correct frame data processing module.
Two, the anti-single particle radiation system that processes based on frame data comprises: single frame detection and reseting module, and the frame data processor,
This single frame detection and reseting module, comprise: frame synchronizing signal is extracted circuit and single frames reset signal generating circuit, frame synchronizing signal extraction circuit detects from each frame data of input and extracts frame synchronizing signal and is transferred to the single frames reset signal generating circuit, the single frames reset signal generating circuit resets to the frame data processor between dynamic stage in frame synchronizing signal, makes all circuit of frame data processor revert to original state;
This frame data processor; comprise: the finite state machine aggregation units; controller protection error correction unit; storer; the storage protection error correction unit; Data Input Interface; data output interface and configuration module; described finite state machine aggregation units respectively with Data Input Interface; data output interface is connected connection with configuration module; and be connected with storer with controller protection error correction unit and be connected; for the treatment of each frame data; described storage protection error correction unit is connected with the finite state machine aggregation units by storer, is used for the data Hamming code of storer is protected and error correction.
The present invention has the following advantages:
(1) the anti-single particle radiation method based on the frame data processing of the present invention's proposition, mistake by the generation of software approach simulation single particle radiation effect, set up rational single-particle inversion model, and the impact that makes mistakes that the downlink data upon handover processing unit is produced of analysis list particle effect, thereby can determine that bit reversal occurs in respectively in frame head, frame data and the postamble impact on result;
(2) the anti-single particle radiation system based on the frame data processing of the present invention's proposition; from the circuit design angle downlink data upon handover processing unit is carried out error correction and protection; detect take every frame data as unit and reset; and the controller in the frame data processor, storer and configuration module protected and error correction; thereby reduce the impact of single particle effect, improve the reliability and stability of downlink data upon handover processing unit and whole system.
Description of drawings
Fig. 1 the present invention is based on the anti-single particle radiation method flow diagram that frame data are processed;
Fig. 2 the present invention is based on the anti-single particle radiation system functional block diagram that frame data are processed;
Fig. 3 is single frame detection of the present invention and reseting module functional block diagram;
Fig. 4 is finite state machine aggregation units functional block diagram of the present invention;
Fig. 5 is controller protection error correction unit functional block diagram of the present invention;
Fig. 6 is storage protection error correction unit functional block diagram of the present invention;
Fig. 7 is configuration module schematic diagram of the present invention;
Fig. 8 is single-bit storage unit circuit and single-particle inversion schematic diagram;
Fig. 9 is the frame format schematic diagram based on the JPEG2000 standard that the present invention adopts.
Embodiment
The frame data processing unit that adopts in the specific embodiment of the invention is the image compression system based on the JPEG2000 standard.The JPEG2000 standard is present general stationary image compression coding standard in the world, and each frame absolute coding is carried out in compression to rest image.The JPEG2000 standard is with the characteristics such as progressive transmission of its good compression performance, support picture quality and resolution, be used widely at aspects such as medical image, remote sensing images and digital picture/transmission of videos, especially in the application such as satellite earth observation that image compression quality is had relatively high expectations, bringing into play increasing technology and strategy function.For other data processing units of processing based on frame data, for example based on image compression unit or packet-based error correction and the modulation circuit of Joint Photographic Experts Group, therefore also occupy important demand in various application, research and design realize having very important significance and being worth for the method for designing of processing based on frame data under space environment and the single particle radiation and counter structure.
Software phantom of the present invention utilizes Microsoft Visual Studio development platform and C/C++ language to finish realization, the hardware implementation structure adopts Xilinx ISE 10.1.03 Integrated Development Environment and Verilog HDL language, is that the FPGA device of XC4Vfx140-11ff1517 is realized in the model of Xilinx company.
With reference to Fig. 1, the anti-single particle radiation method based on the frame data processing that the present invention proposes comprises the steps:
The first step, the mode by reading data produces each pending and separate frame data, and described each frame data are for being used for some two field pictures or packet-based error correction and the modulation sequence of independent processing.
Second step, the mistake that simulation single particle radiation effect produces generates single particle radiation effect error listing:
(2a) some mode bit in the correct frame data processing module or data bit are fixed or are set at random " 1 " or be set to " 0 ", these mode bits or data bit are made a mistake, this mistake is used for the register bit upset that the simulation single-particle inversion causes;
(2b) one or several bit in the correct frame data processing module look-up table is fixed or at random upset, made the corresponding relation of look-up table mistake occur, this mistake is used for the combination logic function mistake that the simulation single-particle inversion causes;
(2c) with stochastic distribution or other probability distribution, in the memory headroom of the correct data cached mapping of frame data processing module, add error code, make data cached making a mistake, this mistake is used for the memory cell data mistake that the simulation single-particle inversion causes;
(2d) revise parameter values or cut-out annexation between the function in correct frame data processing module, function passes is made a mistake, dysfunction even global function that this mistake occurs when serious for the simulation single particle effect lost efficacy;
One or more that (2e) choose in (2a)～(2d) mistake are carried out permutation and combination, generate single particle radiation effect error listing.
The 3rd step, utilize single particle radiation effect error listing, correct frame data processing module built-in function is made amendment, form the erroneous frame data processing module:
(3a) single particle radiation effect error listing is added the frame head part of correct frame data processing module, cause the numerical value on frame head start-stop zone bit or the Information sign position " 1 " " 0 " exchange, bit increase and three kinds of mistakes of bit minimizing to occur, cause correctly to determine frame head start-stop sign bit position implication that can't correction frame originating point information zone bit;
(3b) single particle radiation effect error listing is added the frame data part of correct frame data processing module, cause the numerical value on the frame data effective information position " 1 " " 0 " exchange, bit increase and three kinds of mistakes of bit minimizing to occur, the effective information of frame data representative is made a mistake;
(3c) single particle radiation effect error listing is added the postamble part of correct frame data processing module, cause the numerical value on postamble start-stop zone bit or the Information sign position " 1 " " 0 " exchange, bit increase and three kinds of mistakes of bit minimizing to occur, cause and correctly to determine postamble start-stop sign bit position, implication that can't correction postamble Information sign position.
The 4th step, respectively each frame data is processed separately with correct frame data processing module and erroneous frame data processing module, obtain respectively correct result and the wrong result of each frame:
The sequence I that (4a) each frame data is formed, carry out image Compression based on Joint Photographic Experts Group or JPEG2000 standard by correct frame data processing module function F ():
At first, view data is carried out discrete cosine transform or wavelet transform,
Then, coefficient after the conversion is carried out the entropy coding,
At last, the entropy coding result is packed and export the correct result J=F (I) of each frame;
The sequence S that (4b) each frame data is formed, carry out Error Correction of Coding by correct frame data processing module function G () and process:
At first, treat encoded information bits and divide into groups,
Then, determine the generator matrix that Error Correction of Coding is used,
Then, the code word after the grouping and generator matrix are carried out matrix operation generation Error Correction of Coding check bit,
At last, information bit and check bit are interweaved or merge, and export the correct result T=G (S) of each frame;
The sequence X that (4c) each frame data is formed, undertaken processing with the cascade of Error Correction of Coding based on Joint Photographic Experts Group or the compression of JPEG2000 standard picture by correct frame data processing module function H ():
(4c1) view data is carried out discrete cosine transform or wavelet transform,
(4c2) coefficient after the conversion is carried out the entropy coding,
(4c3) the entropy coding result is packed,
(4c4) information bit after the packing is divided into groups,
(4c5) determine used generator matrix in the cascade processing,
(4c6) will pack code word and cascade after the grouping processed used generator matrix and carried out matrix operation and generate cascade processing check bit,
(4c7) information bit and check bit are interweaved or merge the correct result Y=H (X) of final each frame of output;
The sequence P that (4d) each frame data is formed, process by wrong frame data processing module function E (), this processing is the compression of images capability error that occurs behind the adding single particle radiation effect error listing based on the JPEG2000 standard, or Error Correction of Coding mistake, or based on the cascade processing mistake of JPEG2000 standard picture compression with Error Correction of Coding, obtain the wrong result Q=E (P) of each frame.
The 5th step, contrast the correct result of each frame and the difference of wrong result, the mistake that analysis list particle irradiation effect error listing causes correct frame data processing module specifically appears at which part of frame head, frame data or the postamble of wrong result, and provides error reporting.
In the 6th step, error feedback is corrected:
From Output rusults, the mistake that single particle effect causes is divided into three kinds of situations, it is the bit mistake, bit drop-out and bit increase, wherein the bit mistake comprises random error and burst error, and the wrong normal operation that directly affects decoding software of compressed bit stream, the Recovery image mistake causes decoding, cause that image quality decrease even Recovery image are unavailable, perhaps in decode procedure, lose a large amount of useful datas, and in case makeing mistakes, decoding may directly affect normally carrying out of subsequent decoding, the decoding error accumulation appears, propagate, may cause the decoding software system crash when serious.
The part of makeing mistakes that error reporting is listed feeds back to the erroneous frame data processing module wrong function is wherein corrected, and obtains the result after this step corrects, and carries out as follows:
At first, the frame data analysis module is by the detection to erroneous frame data processed result and correct frame data result, and the misjudgment bit appears at respectively which part of frame head in the compressed bit stream, frame data and postamble,
Then, by the impact on Recovery image of image decompressor software, objective and subjective quality assessment criterion profiling error, affect hardly the Recovery image quality when for example small fault occurs; Perhaps can part correctly recover original image during the moderate mistake, just cause the decline of Quality of recovery; Even fully can't the Recovery image data during gross error,
At last, the part of makeing mistakes that error reporting is listed feeds back to the erroneous frame data processing module wrong function is wherein corrected.
The 7th step repeated for the 5th step and the 6th step, until the result after correcting is identical with the result of correct frame data processing module.
With reference to Fig. 2, the anti-single particle radiation system based on the frame data processing that the present invention proposes comprises: single frame detection and reseting module, and the frame data processor, wherein: single frame detection and reseting module, as shown in Figure 3; The frame data processor, as shown in Figure 4 to 7.
With reference to Fig. 3, single frame detection and reseting module that the present invention proposes comprise: frame synchronizing signal is extracted circuit and single frames reset signal generating circuit.
If contain frame synchronizing signal in the view data of input, frame synchronizing signal extraction circuit therefrom extracts frame synchronizing signal and is transferred to the single frames reset signal generating circuit, and frame synchronizing signal is that high level is effective among Fig. 3, and low level is invalid; If do not contain frame synchronizing signal in the view data of input, then extract frame synchronizing signal corresponding to circuit generation by frame synchronizing signal and be transferred to the single frames reset signal generating circuit.
The single frames reset signal generating circuit resets to the frame data processor between dynamic stage in frame synchronizing signal, single frames reset signal low level is in the effective status that resets among Fig. 3, high level is in disarmed state, the single frames reset signal that produces makes all circuit of frame data processor inside revert to original state before every two field picture valid data input, contingent single-particle mistake just may be limited to this frame inside in this frame image data like this, the capability error of avoiding being caused by single-particle inversion transmits between two continuous frames, leads to errors and can't correct.
This frame data processor, as shown in Figure 4 to 7, it comprises: finite state machine aggregation units, controller protection error correction unit, storer, storage protection error correction unit, Data Input Interface, data output interface and configuration module, the finite state machine aggregation units is connected with configuration module with Data Input Interface, data output interface respectively and is connected, and be connected with storer with controller protection error correction unit and be connected, for the treatment of each frame data; The storage protection error correction unit is connected with the finite state machine aggregation units by storer, is used for the data Hamming code of storer is protected and error correction.
With reference to Fig. 4, the finite state machine aggregation units in the described frame data processor is comprised of some finite state machine cascades, and these finite state machines are finished the processing to each frame data jointly, and wherein each finite state machine comprises:
Current status register is used for the state value of this finite state machine of record under current state;
The output function circuit is used for finishing the logic function under the current state;
Output register is used for the result that the output function circuit produces is deposited and exports;
The NextState commutation circuit is used for after the output function circuit is finished the logic function of current state the state value in the current status register being switched to NextState.
This finite state machine is divided into two kinds of mole type finite state machine and Miller type finite state machines, be respectively mole type and Miller type finite state machine shown in Fig. 4 (a) and Fig. 4 (b), both differences are that mole output of type state machine is only relevant with current state, and the output of Miller type state machine is the function of current state and input signal; No matter both something in commons are mole type or Miller type, and only having current status register is the sequential logic parts with memory capability.
With reference to Fig. 5; controller protection error correction unit in the described frame data processor; realize the control function in the frame data processor is effectively protected by the protection to state machine; this controller protection error correction unit is divided into two kinds of the protection error correction unit of the protection error correction unit of Fig. 5 (a) mole type state machine and Fig. 5 (b) Miller type state machine, and these two kinds protection error correction unit all comprise:
The hamming code logical circuit is used for all state values of current status register are carried out hamming code;
The Hamming decode logic circuit is used for the state value under the current status register current state is carried out the Hamming decoding;
The indication reset circuit of makeing mistakes, be used for judging whether the state value that the Hamming decode logic circuit produces equates with the state value of hamming code logical circuit, if both equate, proceed by the finite state machine original function, if both are unequal, finite state machine is resetted, make it revert to original state;
Described hamming code logical circuit, the Hamming decode logic circuit links to each other with each finite state machine in the finite state machine aggregation units respectively with the indication reset circuit of makeing mistakes, and adopts Hamming code to protect and error correction to finite state machine.
Below the protection error correction unit of Fig. 5 (a) mole type state machine is described in detail.The protection error correction unit of mole type state machine adopts (7 when finite state machine is encoded; 4) Hamming code but be not limited to the block code form of this Hamming code improves the fault-tolerance that the state machine register is affected by single particle effect by the Hamming distance that increases between the code word.In the process of next state transition, the specific implementation function of the protection error correction unit circuit of its mole type state machine is as follows by current state for state in current status register:
Current state obtains net result through output function circuit and output register;
Current state jumps to NextState through the NextState commutation circuit;
The hamming code logical circuit carries out hamming code to the NextState value, and the state in the current status register also becomes NextState by current state simultaneously
The Hamming decode logic circuit is carried out the Hamming decoding to the state value under the current state in the current status register, finishes the protection error correction effect to mole type state machine;
The indication reset circuit of makeing mistakes, judge that whether state value that the Hamming decode logic circuit produces equates with state value in the hamming code logical circuit, if the status error that single particle effect causes is in the error correcting capability of Hamming code, proceed by the finite state machine original function, just can guarantee the correct of whole state machine function; If Hamming code can't be corrected a mistake, then current status register is resetted, avoid current status register to can not get correcting because of single particle effect generation status error always.Fig. 5 (b) is the protection error correction unit of Miller type finite state machine, and protecting the difference of error correction unit with Fig. 5 (a) mole type finite state machine only is the function of current state and input signal in output function.
With reference to Fig. 6, the storage protection error correction unit in the described frame data processor is protected and error correction the data in the storer, and it comprises: check bit counting circuit, check bit checking circuit and comparer.Described check bit counting circuit and check bit checking circuit all are connected with storer is unidirectional, and described comparer is connected with the check bit checking circuit with storer respectively and is connected.The valid data information position M that the check bit counting circuit is treated write store carries out hamming code production check bit K, and with M and K difference write store; Check bit checking circuit sense information position M ' from the storer that is subjected to the single particle radiation effects carries out Hamming code and checks, and generates new check bit K "; Comparer is reader check position K ' from the storer that is subjected to the single particle radiation effects, with K, and " comparing by turn, if K ' and K " equates, the information bit M ' output with reading if both are unequal, will provide the indication that makes mistakes.
This storage protection error correction unit can be alleviated the impact that single particle effect brings storer to a certain extent.In order to save for the storage space of preserving check information, select the form of Hamming code to carry out check bit calculating, for example select to comprise (7,4) form Hamming code of 4 information bits and 3 bit check positions.According to the character of Hamming code, any 1 in 7 bit data is overturn, and Hamming code can be corrected, thereby reaches the purpose of anti-single particle effect.
With reference to Fig. 7, configuration module in the described frame data processor is comprised of three functions identical parameter register group and a decision device, the needed parameter information of storage frame data processor in three register groups, decision device is read respectively corresponding parameter information from three parameter register groups, and according to majority decision mode Output rusults.
The first parameter register group is comprised of the register of some quantity, register quantity is determined by the function of frame data processor, store the needed parameter information of frame data processor in the register, the inner structure of the second parameter register group and the 3rd parameter register group and the first parameter register group are identical, and the function that is about to the first parameter register group has copied again twice.Take in the parameter register group in the parameter information at the frame data processor, decision device is read respectively corresponding data simultaneously from the first parameter register group, the second parameter register group and the 3rd parameter register group, determine Output rusults according to the majority decision mode.Mode by this parameter register group backup, in case the bit in certain parameter register group is subjected to single particle effect to affect the appearance upset and makes a mistake, decision device can be determined correct parameter information by the data of other two parameter register groups, avoids the parameter information of frame data processor mistake in using and obtains wrong result.
For those skilled in the art; after having understood content of the present invention and method; can be in the situation that do not deviate from the principle and scope of the present invention; the method according to this invention is carried out various corrections and the change on form and the details; can make and be not limited to anti-single particle radiation method and the anti-single particle radiation structure of compressing based on the JPEG2000 standard picture among the embodiment; for example other are based on the error correction of frame or packet data processes and anti-single particle radiation method and the anti-single particle radiation structure of modulating system, but these are based on correction of the present invention with change still within claim protection domain of the present invention.
Claims (8)
1. an anti-single particle radiation method of processing based on frame data comprises the steps:
(1) mode by reading data produces each pending and separate frame data;
(2) mistake of simulation single particle radiation effect generation generates single particle radiation effect error listing;
(3) utilize single particle radiation effect error listing, correct frame data processing module built-in function is made amendment, form the erroneous frame data processing module;
(4) respectively each frame data is processed separately with correct frame data processing module and erroneous frame data processing module, obtained respectively correct result and the wrong result of each frame;
(5) contrast the correct result of each frame and the difference of wrong result, the mistake that analysis list particle irradiation effect error listing causes correct frame data processing module specifically appears at which part of frame head, frame data or the postamble of wrong result, and provides error reporting;
(6) part of makeing mistakes that error reporting is listed feeds back to the erroneous frame data processing module wrong function is wherein corrected, and obtains the result after this step corrects;
(7) repeating step (5)～(6) are until the result after correcting is identical with the result of correct frame data processing module;
The mistake that the described simulation single particle radiation of step (2) effect produces generates single particle radiation effect error listing, generates as follows:
(2a) some mode bit in the correct frame data processing module or data bit are fixed or are set at random " 1 " or be set to " 0 ", these mode bits or data bit are made a mistake;
(2b) one or several bit in the correct frame data processing module look-up table is fixed or at random the upset, make the corresponding relation of look-up table mistake occur;
(2c) with stochastic distribution or other probability distribution, in the memory headroom of the correct data cached mapping of frame data processing module, add error code, make data cached making a mistake;
(2d) revise parameter values or cut-out annexation between the function in correct frame data processing module, function passes is made a mistake;
One or more that (2e) choose in (2a)～(2d) mistake are carried out permutation and combination, generate single particle radiation effect error listing.
2. anti-single particle radiation method according to claim 1, wherein the described single particle radiation effect error listing that utilizes of step (3) is made amendment to correct frame data processing module built-in function, as follows operation:
(3a) single particle radiation effect error listing is added the frame head part of correct frame data processing module, cause the numerical value on frame head start-stop zone bit or the Information sign position " 1 " " 0 " exchange, bit increase and three kinds of mistakes of bit minimizing to occur, cause correctly to determine frame head start-stop sign bit position implication that can't correction frame originating point information zone bit;
(3b) single particle radiation effect error listing is added the frame data part of correct frame data processing module, cause the numerical value on the frame data effective information position " 1 " " 0 " exchange, bit increase and three kinds of mistakes of bit minimizing to occur, the effective information of frame data representative is made a mistake;
(3c) single particle radiation effect error listing is added the postamble part of correct frame data processing module, cause the numerical value on postamble start-stop zone bit or the Information sign position " 1 " " 0 " exchange, bit increase and three kinds of mistakes of bit minimizing to occur, cause and correctly to determine postamble start-stop sign bit position, implication that can't correction postamble Information sign position.
3. anti-single particle radiation method according to claim 1, wherein the correct frame data processing module of the described usefulness of step (4) and erroneous frame data processing module are processed separately each frame data respectively, obtain respectively correct result and the wrong result of each frame, carry out as follows:
The sequence I that (4a) each frame data is formed, carry out image Compression based on Joint Photographic Experts Group or JPEG2000 standard by correct frame data processing module function F ():
At first, view data is carried out discrete cosine transform or wavelet transform,
Then, coefficient after the conversion is carried out the entropy coding,
At last, the entropy coding result is packed and export the correct result J=F (I) of each frame;
The sequence S that (4b) each frame data is formed, carry out Error Correction of Coding by correct frame data processing module function G () and process:
At first, treat encoded information bits and divide into groups,
Then, determine the generator matrix that Error Correction of Coding is used,
Then, the code word after the grouping and generator matrix are carried out matrix operation generation Error Correction of Coding check bit,
At last, information bit and check bit are interweaved or merge, and export the correct result T=G (S) of each frame;
The sequence X that (4c) each frame data is formed, undertaken processing with the cascade of Error Correction of Coding based on Joint Photographic Experts Group or the compression of JPEG2000 standard picture by correct frame data processing module function H ():
The first step is carried out discrete cosine transform or wavelet transform to view data,
Second step carries out the entropy coding to coefficient after the conversion,
The 3rd step, the entropy coding result is packed,
The 4th step, the information bit after the packing is divided into groups,
The 5th step, determine used generator matrix in the cascade processing,
The 6th goes on foot, and the code word after packing is divided into groups and cascade are processed used generator matrix and carried out matrix operation generation cascade processing check bit,
The 7th step interweaved information bit and check bit or merges, and output is the correct result Y=H (X) of each frame finally;
The sequence P that (4d) each frame data is formed, process by wrong frame data processing module function E (), this processing is the compression of images capability error that occurs behind the adding single particle radiation effect error listing based on Joint Photographic Experts Group or JPEG2000 standard, or Error Correction of Coding mistake, or process mistake based on the compression of Joint Photographic Experts Group or JPEG2000 standard picture and the cascade of Error Correction of Coding, obtain the wrong result Q=E (P) of each frame.
4. anti-single particle radiation system that processes based on frame data comprises: single frame detection and reseting module, and the frame data processor,
Described single frame detection and reseting module, comprise: frame synchronizing signal is extracted circuit and single frames reset signal generating circuit, frame synchronizing signal extraction circuit detects from each frame data of input and extracts frame synchronizing signal and is transferred to the single frames reset signal generating circuit, the single frames reset signal generating circuit resets to the frame data processor between dynamic stage in frame synchronizing signal, makes all circuit of frame data processor revert to original state;
Described frame data processor, comprise: finite state machine aggregation units, controller protection error correction unit, storer, storage protection error correction unit, Data Input Interface, data output interface and configuration module, described finite state machine aggregation units is connected with configuration module with Data Input Interface, data output interface respectively and is connected, and be connected with storer with controller protection error correction unit and be connected, for the treatment of each frame data; Described storage protection error correction unit is connected with the finite state machine aggregation units by storer, is used for the data Hamming code of storer is protected and error correction.
5. anti-single particle radiation according to claim 4 system, it is characterized in that the finite state machine aggregation units in the frame data processor, be comprised of some finite state machine cascades, these finite state machines are finished the processing to each frame data jointly, and wherein each finite state machine comprises:
Current status register is used for the state value of this finite state machine of record under current state;
The output function circuit is used for finishing the logic function under the current state;
Output register is used for the result that the output function circuit produces is deposited and exports;
The NextState commutation circuit is used for after the output function circuit is finished the logic function of current state the state value in the current status register being switched to NextState.
6. anti-single particle radiation according to claim 4 system is characterized in that the controller protection error correction unit in the frame data processor, comprising:
The hamming code logical circuit is used for all state values of current status register are carried out hamming code;
The Hamming decode logic circuit is used for the state value under the current status register current state is carried out the Hamming decoding;
The indication reset circuit of makeing mistakes, be used for judging whether the state value that the Hamming decode logic circuit produces equates with the state value of hamming code logical circuit, if both equate, proceed by the finite state machine original function, if both are unequal, finite state machine is resetted, make it revert to original state;
Described hamming code logical circuit, the Hamming decode logic circuit links to each other with each finite state machine in the finite state machine aggregation units respectively with the indication reset circuit of makeing mistakes, and adopts Hamming code to protect and error correction to finite state machine.
7. anti-single particle radiation according to claim 4 system is characterized in that the storage protection error correction unit in the frame data processor adopts the form of Hamming code that the data in the storer are protected and error correction, and this storage protection error correction unit comprises:
The check bit counting circuit carries out hamming code production check bit K for the valid data information position M that treats write store, and with M and K write store;
The check bit checking circuit is used for carrying out Hamming code from the storer sense information position M ' that is subjected to the single particle radiation effects and checks, and generates new check bit K ";
Comparer equates for " comparing by turn, if K ' and K " from being subjected to the storer reader check position K ' of single particle radiation effects, with K, and the information bit M ' output with reading if both are unequal, will provide the indication that makes mistakes.
8. anti-single particle radiation according to claim 4 system, it is characterized in that the configuration module in the frame data processor is comprised of three functions identical parameter register group and a decision device, the needed parameter information of storage frame data processor in three register groups, decision device is read respectively corresponding parameter information from three parameter register groups, and according to majority decision mode Output rusults.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110202501 CN102306213B (en) | 2011-07-19 | 2011-07-19 | Anti-single particle irradiating method and anti-single particle irradiating system based on frame data processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110202501 CN102306213B (en) | 2011-07-19 | 2011-07-19 | Anti-single particle irradiating method and anti-single particle irradiating system based on frame data processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102306213A CN102306213A (en) | 2012-01-04 |
| CN102306213B true CN102306213B (en) | 2013-01-23 |
Family
ID=45380074
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110202501 Expired - Fee Related CN102306213B (en) | 2011-07-19 | 2011-07-19 | Anti-single particle irradiating method and anti-single particle irradiating system based on frame data processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102306213B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9047988B2 (en) * | 2012-11-20 | 2015-06-02 | International Business Machines Corporation | Flash interface error injector |
| CN103236271A (en) * | 2013-03-26 | 2013-08-07 | 深圳市国微电子有限公司 | Memory for reinforcing single event upset based on triple modular redundancy, and reinforcing method |
| CN103218272B (en) * | 2013-04-25 | 2015-09-23 | 西安空间无线电技术研究所 | A kind of spaceborne digital signal processor upset reinforcement means |
| CN103530196A (en) * | 2013-09-24 | 2014-01-22 | 北京控制工程研究所 | Method for preventing single event upset of FPGA |
| CN103617811B (en) * | 2013-12-03 | 2017-08-29 | 中国科学院微电子研究所 | Error correction circuit of SRAM type memory |
| CN107395327A (en) * | 2017-07-19 | 2017-11-24 | 上海航天测控通信研究所 | A kind of high reliability LDPC encoder suitable for satellite communication |
| CN107894898A (en) * | 2017-11-28 | 2018-04-10 | 中科亿海微电子科技（苏州）有限公司 | Refresh device, implementation method and the fpga chip with error correction on SRAM type FPGA pieces |
| US10348302B1 (en) * | 2018-05-31 | 2019-07-09 | Bae Systems Information And Electronic Systems Integration Inc. | Radiation-hardened latch circuit |
| CN112798943B (en) * | 2021-01-16 | 2022-03-08 | 西安电子科技大学 | Real-time FPGA error detection method based on standard input and output verification |
| CN113296820B (en) * | 2021-06-18 | 2023-01-24 | 上海航天测控通信研究所 | Satellite-borne single event effect resisting reinforcement method based on SRAM FPGA |
| CN114692564B (en) * | 2022-04-18 | 2024-07-26 | 北京泽石科技有限公司 | Verification system and method of LDPC error correction module, storage medium and electronic equipment |
| CN117474061B (en) * | 2023-12-26 | 2024-03-19 | 华中师范大学 | Anti-radiation low-delay neural network reasoning acceleration chip |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1363437A1 (en) * | 2001-08-22 | 2003-11-19 | Telecommunications Advancement Organization of Japan | COMMUNICATION QUALITY ESTIMATION METHOD, COMMUNICATION QUALITY ESTIMATION APPARATUS, AND COMMUNICATION SYSTEM |
| CN1516461A (en) * | 2003-01-10 | 2004-07-28 | 孙顺姬 | Digital dynamic image compression method based on variable image plaid |
| CN1716211A (en) * | 2004-07-02 | 2006-01-04 | 中国科学院上海微系统与信息技术研究所 | Data error detects and corrects the positive and negative coding structure of intersection of usefulness and the method for decoding |
| CN1980386A (en) * | 2006-07-20 | 2007-06-13 | 西安交通大学 | 1/5 small-wave core optimizing design computing method based on JPEG2000 picture compression |
-
2011
- 2011-07-19 CN CN 201110202501 patent/CN102306213B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1363437A1 (en) * | 2001-08-22 | 2003-11-19 | Telecommunications Advancement Organization of Japan | COMMUNICATION QUALITY ESTIMATION METHOD, COMMUNICATION QUALITY ESTIMATION APPARATUS, AND COMMUNICATION SYSTEM |
| CN1516461A (en) * | 2003-01-10 | 2004-07-28 | 孙顺姬 | Digital dynamic image compression method based on variable image plaid |
| CN1716211A (en) * | 2004-07-02 | 2006-01-04 | 中国科学院上海微系统与信息技术研究所 | Data error detects and corrects the positive and negative coding structure of intersection of usefulness and the method for decoding |
| CN1980386A (en) * | 2006-07-20 | 2007-06-13 | 西安交通大学 | 1/5 small-wave core optimizing design computing method based on JPEG2000 picture compression |
Non-Patent Citations (1)
| Title |
|---|
| 雷杰、周有喜、吴成柯、李云松、孔繁锵."一种新的高效干涉多光谱图像压缩算法".《光子学报》.2009,第38卷(第6期),第1534-1538页. |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102306213A (en) | 2012-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102306213B (en) | Anti-single particle irradiating method and anti-single particle irradiating system based on frame data processing | |
| Neuberger et al. | A multiple bit upset tolerant SRAM memory | |
| Rao et al. | Protecting SRAM-based FPGAs against multiple bit upsets using erasure codes | |
| Tambatkar et al. | Error detection and correction in semiconductor memories using 3D parity check code with hamming code | |
| Li et al. | Design and characterization of SEU hardened circuits for SRAM-based FPGA | |
| EP2975522A2 (en) | Digital error correction | |
| Kishani et al. | HVD: horizontal-vertical-diagonal error detecting and correcting code to protect against with soft errors | |
| KR101667400B1 (en) | Apparatus and method for generating and detecting single event upset | |
| Liu et al. | Low redundancy matrix-based codes for adjacent error correction with parity sharing | |
| Silva et al. | Extended matrix region selection code: An ECC for adjacent multiple cell upset in memory arrays | |
| Sharma et al. | An HVD based error detection and correction of soft errors in semiconductor memories used for space applications | |
| CN106328209A (en) | Storage single event multiple bit upset fault-tolerance method and circuit | |
| Dang et al. | Parity-based ECC and mechanism for detecting and correcting soft errors in on-chip communication | |
| Satyanarayana et al. | Design and implementation of error detection and correction system for semiconductor memory applications | |
| Dang et al. | An adaptive and high coding rate soft error correction method in network-on-chips | |
| Neto et al. | Evaluating fault coverage of bulk built-in current sensor for soft errors in combinational and sequential logic | |
| CN104378120A (en) | Hsiao coding check matrix generation method for continuous MBU detection | |
| Kamalakannan et al. | Implementation of error correction technique based on decimal matrix code | |
| CN202838976U (en) | EDAC (Error Detection And Correction)-module-based fault-tolerant device of triple modular redundancy to storage | |
| Maheswari et al. | Random and Triple burst error correction code with low redundancy for Network-on-Chip link | |
| Shahriyar et al. | An efficient error correction approach by using successive parity generation | |
| Sunita et al. | Matrix code based multiple error correction technique for n-bit memory data | |
| Ruiz et al. | Zero-Space In-Weight and In-Bias Protection for Floating-Point-based CNNs | |
| Dutta et al. | Combinational logic circuit protection using customized error detecting and correcting codes | |
| Vaisakhi et al. | Novel Double Modular Redundancy Based Fault-Tolerant FIR Filter for Image Denoising. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130123 Termination date: 20180719 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |