CN114298721A - Portable credible device of tracing to source towards chinese-medicinal material - Google Patents
Portable credible device of tracing to source towards chinese-medicinal material Download PDFInfo
- Publication number
- CN114298721A CN114298721A CN202111622717.6A CN202111622717A CN114298721A CN 114298721 A CN114298721 A CN 114298721A CN 202111622717 A CN202111622717 A CN 202111622717A CN 114298721 A CN114298721 A CN 114298721A
- Authority
- CN
- China
- Prior art keywords
- data
- module
- traditional chinese
- block chain
- 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.)
- Pending
Links
Images
Abstract
The invention discloses a portable credible traceability device for traditional Chinese medicinal materials, which comprises a federal learning module, a block chain module, a networking module, a power supply module, a display module, a printing module, an acquisition module, a sensing layer, a transmission layer, a data layer, a service layer and an application layer, wherein the traditional Chinese medicinal materials are traced to prevent similar production items from being repeatedly invested and executed in each management department and each traditional Chinese medicinal material production enterprise, so that the burden of the production enterprise is reduced, the resource sharing is realized, and the social investment and the management operation cost are saved; the traditional Chinese medicine tracing can provide an effective monitoring means for relevant management departments of the traditional Chinese medicines, the quality safety condition of the traditional Chinese medicines is mastered on the whole, a real scientific basis is provided for decision making, and the occurrence rate of quality safety accidents of the traditional Chinese medicines is reduced through whole-process supervision, so that the market order is maintained, the management level and the productivity of production enterprises are improved, the competitiveness of the enterprises is enhanced, and the quality safety consciousness and the cognition degree of consumers on the traditional Chinese medicines are enhanced.
Description
Technical Field
The invention relates to a trusted traceability device, in particular to a portable trusted traceability device for traditional Chinese medicinal materials.
Background
The traditional Chinese medicine is a material basis for inheritance and development of the traditional Chinese medicine industry, and the quality of the traditional Chinese medicine is a strategic resource which is concerned with the growth and the decline of the traditional Chinese medicine industry and the relation between China and civilian life. The traditional Chinese medicine traceability protection and service system construction is promoted by advanced means, the method is an important measure for benefiting livelihood, promoting consumption, stably increasing and promoting supply side structural reform, has great significance for deepening medical and health system reform, improving people health level, developing strategic emerging industry, increasing farmer income, promoting ecological civilization construction, improving supply chain efficiency and product quality safety guarantee level, promoting circulation transformation upgrading and innovation development, constructing an informatization monitoring and supervision system and creating a market environment for safe consumption; driven by economic benefits, the safety incidents of the traditional Chinese medicinal materials are gradually frequent in recent years, the number of counterfeit substitutes in the market of the traditional Chinese medicinal materials is increasing, and the phenomenon of heavy metal pollution caused by pesticide residues in the traditional Chinese medicinal materials is serious. Along with the increasingly deep coverage of scenes by the fusion application of the national security and the new technology, the method is more and more protected, provides technical guarantee for the further development and effective implementation of the traditional Chinese medicine material traceability, and has a rooting basis by the high fusion with a supply chain transaction system. Therefore, the portable credible tracing method and device for the traditional Chinese medicinal materials are invented, the complementation and fusion of the blockchain technology and the federal learning are utilized, the training precision of a local model is guaranteed, meanwhile, the parameter uploading efficiency is improved, the problem of data privacy protection is solved by adopting a state-secret algorithm, and the overall safety is improved;
federal learning is proposed by Google in 2016 for the first time, and is used for solving the problem that an android mobile phone terminal user updates a frequent word model in an input method locally, the design target of the Federal learning is to guarantee information safety during big data exchange, protect terminal data and personal data privacy and develop efficient machine learning among multiple participants or multiple computing nodes on the premise of guaranteeing legal compliance. In the model, all training data are stored locally by participants, the model is trained locally, then the model obtained by training is updated and transmitted to the cloud, and other participants download the update to own mobile equipment, so that the accuracy of the training model is improved.
Currently federal learning techniques are receiving wide attention and deep participation in the industry, and a large number of companies and organizations are actively laying out in terms of frameworks, products/applications, and standardization.
(1) Technical framework aspects
Currently, internet enterprises mainly actively carry out related research works, and mainly comprise large-scale enterprises such as Google, Facebook, micro-public, ali, ant golden clothes, Baidu, Jingdongduo, byte jumping and the like, and emerging enterprises such as Tongdun, Fudu and blue image technologies and the like. Mainstream federal learning open source frameworks include FATE (micro member), PaddleFL (hundred degrees), Fedlearner (byte jitter), TensorFlow FL (Google), PySyft (facebook), and the like.
(2) Aspect of standardization
The whole is still in the starting stage. Currently, various organizations in the industries such as 3GPP, ITU-T, CCSA, etc. are actively advancing the relevant standardization work. In 3 months 2021, the Federal learning architecture and application Specifications (IEEE P3652.1) were promulgated by the Ministry of China in IEEE. In recent two years, the technical workgroups of domestic CCSA TC1, TC8, TC601 and the like develop related research subjects and standard items aiming at the technologies of federal learning, privacy calculation, multi-party safety calculation and the like.
(3) Aspect of application
The application scenes of the federal learning technology are rich, and main internet enterprises, financial enterprises and scientific research institutions in the industry explore and apply the federal learning technology in multiple fields of finance, e-commerce, medical treatment, internet of things, communication and the like.
(4) Aspects of business model
Due to the fact that data distribution of all parties is uneven, and data value and contribution degree are different, an evaluation method of the participation degree and the contribution degree of a data owner to a combined model needs to be researched, a fair cooperation and sharing win-win incentive mechanism is established to attract more data owners to evaluate the participation degree and the contribution degree of the combined model, a fair cooperation and sharing win-win incentive mechanism is established to attract more data owners to participate in federal learning, and therefore a cross-industry and cross-field data circulation ecology with healthy and sustainable development is established. The federal learning has an incentive mechanism, the contribution of participants to the model can be evaluated, and the participants need to continuously participate in the process of federal learning, so that the long-term sustainable operation of the federal is realized.
At present, all countries in the world strive for laying out block chains, open up new tracks of international competition, and seize the high points of a new round of industrial innovation so as to improve the international competitiveness of the countries, but the countries are both domestic and foreign, and the block chain development roads are all searched ahead, mainly from the aspects of a block chain bottom layer, a middle layer, an application layer and the like.
(1) Bottom layer key technology
Including BFT consensus algorithms, atom-chain-crossing techniques, sub-chain techniques, etc. The BFT consensus algorithm comprises the following steps: an autonomous controllable algorithm is not formed in China, and optimization and adjustment are mostly carried out on open source codes. If the fun chain and the well communication adopt RBFT algorithm, but the strategy is slightly different; atom chain-crossing technology: synchronous exploration at home and abroad shows that although products of some domestic enterprises are stably operated, the performance efficiency, stability and applicability are far inferior to those of the same foreign enterprises. Such as the Relay technique of Xunliexin, contract Cross-chain technique of Shanghai Huo Yi; the subchain technology comprises the following steps: besides Hangzhou secret apes and Shanghai chain views, domestic enterprises almost do not explore a sub-chain technology, and foreign enterprises mostly adopt a layered design idea to guarantee the safety of the sub-chain by deploying a high-safety main chain and improve the processing speed of the main chain by using the sub-chain.
(2) Intermediate layer key technology
Including hash locking, distributed private key control, private data authorization access, etc. Hash locking: relatively mature products appear in the early foreign research; although domestic institutions are realized, large-scale application and inspection are not available. Distributed private key control: the foreign technology is relatively perfect, and the domestic safety and reliability are slightly insufficient; and (3) authorizing access to the private data: domestic enterprise research is early, but access authorized by private data based on national password is not popularized and applied on a block chain on a large scale.
(3) Application layer key technology
Including distributed applications, intelligent contracts, etc. Intelligent contract: the intelligent contract language is mainly composed of identity, JavaScript, Wasm, Move and the like, is a leading intelligent contract language abroad, and does not form an autonomous intelligent contract language at home. Distributed application: the DApp and the DeFi technology in China are relatively mature, the types are far richer than those in foreign countries, and the fields of electronic invoices, electronic certificates, insurance, judicial law and the like are concerned. Such as Tengcong science and technology, electronic invoices of Dongkong stocks, Hangzhou Internet court of ant golden clothes, etc.;
the traditional Chinese medicinal material tracing mode is based on the technology of the Internet of things basically, and related data information of the Chinese medicinal material is collected through wireless equipment of the Internet of things and then transmitted to a central database. The traditional Chinese medicine information tracing system is basically established on the basis of the technology of the Internet of things and layered based on the functional structure of the traditional Chinese medicine information tracing system, and is shown in the attached figure 1. The sensing layer collects relevant information of the traditional Chinese medicinal materials in the processes from production to sale by applying the Internet of things sensing technology, and the collected data are uploaded to the network layer and then transmitted to the application layer through the wireless equipment. The functions of the sensing layer of the internet of things system can be positioned, application modes of the sensing layer in the traditional Chinese medicine material tracing system can be classified, and the functions can be mainly divided into the following five aspects, namely an object for information acquisition, main content information of the object for information acquisition, a method for information acquisition, a marking mode selected in the information acquisition process and a mode for storing acquired data. The main function of the network layer is information transmission, and data information is transmitted to the application layer by using a WLAN or a 4G/5G network. The application layer is mainly used for classifying and processing a large amount of data acquired in the traditional Chinese medicine tracing process, and the function of each data can be played. In the operation process of the traditional Chinese medicine material tracing system, the application layer can intelligently analyze and process the uploaded data by using a credit data mining technology. The application layer realizes management of the traditional Chinese medicine from the production process to the sales process. After the system acquires the uploaded data, relatively complete data tracing can be provided.
At present, the traditional Chinese medicine material tracing device is in a starting stage, the core technology in the field is continuously developed, but the safety of the traditional Chinese medicine material tracing device is less in the technical field. The existing traceability in the field of traditional Chinese medicinal materials is mainly a traceability centralized system used in enterprises based on a bar code technology, an acquisition mode lacks a necessary supervision link, information tampering is difficult to avoid traceability information being protected by a stakeholder, invalid effective measures are used for supervising change records of the information, and the requirements of quality traceability and transaction integrity realization are difficult to achieve. If the absolute credibility of the source data in the logistics process cannot be guaranteed, the alliance chain system with higher computing performance is meaningless;
therefore, the existing traditional Chinese medicinal material tracing device has the following defects:
(1) the tracing information is easy to be tampered
The tracing technical solution commonly adopted at present adopts a centralized architecture in both a label class and a system integration scheme class. The centralized traceability system cannot self-prove, which is a fundamental problem in the current traceability industry. The system database of the traditional traceability industry is mainly a cloud database provided by a database center built by an enterprise or a system technology service side, and the data control management right is centralized on the enterprise side. Because it is itself a stakeholder on the circulation chain, there is a possibility that the enterprise will tamper with the ledger and disable the traceability process. When problematic traditional Chinese medicinal materials appear or the ledger information is unfavorable, source tracing data can be modified or deleted, and the vulnerability risk of intentionally avoiding self responsibility exists. Meanwhile, the potential safety hazard that a hacker attacks the system database exists due to the external risk, and the market of the true data and false traditional Chinese medicinal materials by stealing the data can imitate the bad means
(2) Insufficient sharing of data information across a supply chain
The coverage range of tracing the source of the traditional Chinese medicinal materials is limited at present, most of the tracing of the traditional Chinese medicinal materials only extends to the deep processing stage of products, and the field of tracing of raw materials and the field of tracing of logistics transportation are not involved. The root of the method lies in that information of each supplier, logistics provider, manufacturer and retailer in a supply chain is closed, respective management systems operate independently, a multi-party common trust system and a privacy protection system are difficult to establish, the information islanding phenomenon is serious, a centralized traceability means cannot extend along the supply chain, and the authenticity and the public trust of the traditional traceability method are reduced.
(3) Source tracing information responsibility subject is difficult to determine
The construction of a traceability system aims to achieve the goal of being accountable, but the traditional traceability industry cannot automatically confirm the responsibility main body and the root of the problem, the traditional traceability industry still needs to rely on the law enforcement of a supervision department and the quality safety control inside an enterprise, and the system needs to be managed by people when the responsibility is really determined. Therefore, the phenomena of asymmetric data and untimely supervision exist, when quality accidents and safety problems occur to traditional Chinese medicinal materials, a production enterprise, as a responsibility main party, needs to bear corresponding social responsibility and legal responsibility, but the production enterprise has the sole control right of traceability data, and cannot avoid the situation that the enterprise chooses to modify or delete the traceability data in order to avoid responsibility, so that a specific responsibility main body cannot be confirmed, and finally, the traditional Chinese medicinal materials with the problems cannot be subjected to targeted processing.
(4) User trust is difficult to establish
At present, the quantity of main safety indexes of Chinese medicinal materials in China is basically equivalent to that of developed countries in the west, but the proportion of mandatory standard quantity is lower. The low standard constraint causes the poor integrity of market operation, and the problems of weak administrative supervision, vulnerability of supervision systems and the like. Counterfeit and inferior traditional Chinese medicinal materials which are difficult to distinguish and trace are difficult to trace, so that the rights and interests of consumers are violated, the life safety of people is threatened even in the field of the traditional Chinese medicinal materials, the safety accidents of the traditional Chinese medicinal materials are frequent, the confidence of Chinese consumers on the safety of the traditional Chinese medicinal materials is seriously struck, and the trust of the users is difficult to establish.
Disclosure of Invention
The invention aims to:
(1) series supply chain body, construction technology trust foundation stone
The method is characterized in that a plurality of main bodies are involved in a supply chain, the problem of cooperative trust among the main bodies cannot be effectively solved in the original centralized form, a block chain technology is used for establishing a alliance network for manufacturers, dealers, logistics service providers, retailers, government regulatory agency detection mechanisms and the like, alliance members in the whole network form a consensus mechanism, after data are written into the block chain, synchronization of all nodes is automatically completed, and once the data are linked, the data cannot be tampered unilaterally, so that the possibility that the data are manually interfered in the transmission process is greatly reduced, a mutual trust mechanism is established on a technical level, and a trust foundation is provided for cooperation among the main bodies in the supply chain.
(2) Quickening information circulation and providing enterprise decision basis
In supply chain management, the bullwhip effect has always been a big problem to the operation of the supply chain, that is, when the supply chain is composed of a plurality of links, the downstream fluctuation will gradually become larger to be transmitted to the upstream. Through the block chain alliance network, data are written into the block chain, each node automatically completes data synchronization, and data information stored in all the nodes is completely consistent, so that the problem of information asymmetry can be effectively solved, and an effective basis is provided for decision making of enterprises on and off the supply chain.
(3) Reduce the difficulty of pursuing responsibility and assist government supervision
The government supervision department, as a monitoring information in the block chain alliance chain, can synchronously acquire data written in the block chain by each main body, can quickly locate an interest source once a problem occurs, check the destination of the problem related to a product, quickly start emergency treatment measures and avoid further expansion of the event. Through the data source of the block chain, the responsible party with problems can be quickly found out, and corresponding management is carried out. Therefore, safety control of the consumer goods is realized in a quick and effective technical mode, legal rights and interests of consumers are guaranteed, and construction and attack of a national quality control system are assisted.
(4) Transmit credit value and improve enterprise financing efficiency
Supply chain finance mainly provides financing requirements for upstream and downstream small and medium-sized enterprises around core enterprises, and common development and continuous operation of the enterprises in the supply chain are realized through effective transmission of supply chain system information, resources and the like. In the original mode, except for a first-level enterprise and a second-level enterprise which are closely related to core enterprises, other enterprises are difficult to transmit credit, and information of all enterprises in the whole supply chain is completely consistent through a block chain alliance network, so that the evaluation basis of the credit of the small and medium-sized enterprises is greatly improved, a better trust basis is provided for financing of the small and medium-sized enterprises, and the problems in the background art are solved.
In order to achieve the purpose, the invention provides the following technical scheme:
towards portable credible device of tracing to source of chinese-medicinal material, including federal study module, block chain module, networking module, power module, display module, print module, collection module, perception layer, transmission layer, data layer, service layer, application layer, safety architecture and supervision and examination framework, federal study module includes inquiry module, acquisition module, check module and training module, and the operation of local training of federal study is in the user side, according to the local model parameter of data study of user side, and the block chain then operates in the service side, receives and stores up the model parameter of federal study, block chain module includes encryption module, storage module and authentication module, collection module is used for gathering the data information of chinese-medicinal material, and the networking module is used for being connected microprocessor and appointed platform.
As a further scheme of the invention: the federal learning module is used for carrying out local model training on traditional Chinese medicine data, can realize a virtual common model through the parameter exchange mode under the encryption mechanism under the premise that the data can not go out of the local, guarantee the privacy of traditional Chinese medicine data, the block chain module is used for verifying staff and equipment, adopts the state secret SM2 certificate to carry out validity identity authentication, after the global training of the federal learning module is accomplished, carries out encryption protection and storage in each node of block chain network to global model parameter through the state secret algorithm SM4 algorithm, obtains the relevant block chain database of traditional Chinese medicine from the internet and carries out traceability comparison, obtains traditional Chinese medicine kind information.
As a still further scheme of the invention: the printing module is used for generating a bar code according to the Chinese medicinal material information acquired by the tracing module, assigning codes and controlling the portable printer to print.
As a still further scheme of the invention: the system comprises a power module, a sensing layer, a transmission layer and a card sender, wherein the power module is used for converting received solar energy into electric energy to be stored in a storage battery, the sensing layer is used for providing a basic acquisition device access function and comprises a fixed reader-writer, a handheld reader-writer and a card sender, 3 reader-writers and corresponding labels, basic reading and writing operations of the labels generated by the sensing layer are main components of block chain data in the system, the transmission layer is used for providing high-reliability and high-safety transmission of sensing data, and the sensing layer data are accessed into a gateway through a wired network.
As a still further scheme of the invention: the data layer is used for providing the most basic data processing functions, such as data encryption, hash function operation, digital signature and other mechanisms, is the most basic data processing support of the federal learning and block chain technology, and the service layer is used for providing a basic layer for services such as calculation, storage and the like for the service layer from bottom to top; a virtual machine, a high-level language editor, etc. for intelligent contract implementation; the core block chain related services are provided for the tracing application, high availability, high convenience, reliable and reliable data access, accurate data calculation and safe data management of the services are guaranteed, and the services are guarantee for the capability of the tracing application.
As a still further scheme of the invention: the application layer is used for providing functions of an application interface of the acquisition device, such as warehousing and warehousing, inventory, anti-counterfeiting query and the like, the safety architecture is used for functions including authentication and identity management, authorization, safety strategy management, privacy protection and the like related to system safety, and the supervision and examination architecture is used for functions including supervision support, examination realization and the like which are responsible for normal and accurate operation of the system.
Compared with the prior art, the invention has the beneficial effects that:
(1) economic benefits
The traditional Chinese medicine material tracing can prevent similar production items from being repeatedly invested and executed in each management department and each traditional Chinese medicine material production enterprise, further reduce the burden of the production enterprises, realize resource sharing, save social investment and management operation cost, and can also supervise and control the quality safety of the traditional Chinese medicine materials through a traditional Chinese medicine material tracing system, reduce the quality safety risk and create more economic benefits;
(2) social benefits
The traditional Chinese medicine tracing can provide an effective monitoring means for relevant management departments of the traditional Chinese medicines, the quality safety condition of the traditional Chinese medicines is mastered on the whole, a real scientific basis is provided for decision making, and the occurrence rate of quality safety accidents of the traditional Chinese medicines is reduced through whole-process supervision, so that the market order is maintained, the management level and the productivity of production enterprises are improved, the competitiveness of the enterprises is enhanced, and the quality safety consciousness and the cognition degree of consumers on the traditional Chinese medicines are enhanced.
Drawings
Fig. 1 is a traditional traceability system architecture diagram of a portable trusted traceability device for traditional Chinese medicinal materials.
Fig. 2 is a traditional Chinese medicine material traceability level architecture diagram of the portable credible traceability device for traditional Chinese medicine materials.
Fig. 3 is a traditional Chinese medicine material traceability acquisition system architecture diagram of the portable trusted traceability device for traditional Chinese medicine materials.
Fig. 4 is a traditional Chinese medicine material traceability flowchart of the portable trusted traceability device for traditional Chinese medicine materials.
FIG. 5 is a structural framework diagram of a traditional Chinese medicine material traceability device of a portable credible traceability device for traditional Chinese medicine materials
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 5, in the embodiment of the present invention, a portable trusted traceability device for traditional Chinese medicinal materials includes a federal learning module, a block chain module, a networking module, a power module, a display module, a printing module, an acquisition module, a sensing layer, a transmission layer, a data layer, a service layer, an application layer, a security architecture and a supervision and review architecture, where the federal learning module includes a query module, an acquisition module, a verification module and a training module, the federal learning local training is run at a user side, local model parameters are learned according to data of the user side, the block chain is run at the service side, and the block chain receives and stores model parameters of federal learning, and the method includes the following steps:
the method comprises the following steps: the improved data acquisition device which needs to have the function of executing programming is adopted, the acquired data is from the information record of the complete life cycle of the data of planting production, processing and packaging, transportation and circulation, storage and sale and the like of the traditional Chinese medicinal materials, and the improved data acquisition device has the unique identification which can not be reconfigured by hardware and the functions of key control, configuration and data acquisition and non-local storage, so that information leakage is effectively prevented;
step two: the method comprises the following steps that a user carries out local training to obtain a corresponding local model, participated users find model parameters locally according to held Chinese medicinal material data through an algorithm based on gradient descent to minimize a loss function, and the model parameters are uploaded to a platform or a designated server after homomorphic encryption;
step three: the server collects model parameters trained by the client for global training, and the acquisition device is required to be ensured to operate to automatically execute an agreed intelligent contract of the block chain, and data fusion and operation audit can be performed;
step four: the updated model parameters of the server are stored in a transaction form through a block chain, data are packaged strictly according to a block chain data block format, and then data records are stored in each block chain node through an encryption algorithm and a timestamp adding mode;
step five: on the premise of not influencing the safety of the blockchain, a block chain encryption flow is optimized by adopting a state secret algorithm, so that the safer and more reliable blockchain realizes safe and credible decentralized point-to-point communication, safe transmission and privacy protection in a distribution system without a credible environment through technical means such as time stamps, information encryption and consensus mechanisms;
step six: the blockchain generates a transaction block, and each participating node collects transactions (model parameters) from the user layer and packages the transactions into blocks through cryptographic signatures. Determining the block weight through a running consensus mechanism among all nodes of the block chain, obtaining the node of the block weight, broadcasting the block to the whole network, and adding the block chain after the block chain passes the authentication;
step seven: and aggregating the model parameters of the nodes according to the records on the block chain, and updating the whole model. Further issuing the model to each participating user, thereby starting a new round of training and learning;
the block chain module comprises an encryption module, a storage module and an authentication module, the acquisition module is used for acquiring data information of traditional Chinese medicinal materials, the networking module is used for connecting the microprocessor with a specified platform, the federal learning module is used for carrying out local model training on the traditional Chinese medicinal material data, a virtual common model can be realized through a parameter exchange mode under an encryption mechanism on the premise that the data is not local, the privacy of the traditional Chinese medicinal material data is guaranteed, the block chain module is used for verifying workers and equipment, validity identity authentication is carried out by adopting a state secret SM2 certificate, after the global training of the federal learning module is finished, global model parameters are encrypted and protected through a state secret SM4 algorithm and stored in each node of the block chain network, and a block chain database related to the traditional Chinese medicinal materials is obtained from the internet and subjected to source tracing comparison, the system comprises a printing module, a power module, a sensing layer, a fixed reader-writer, a handheld reader-writer, a card sender, 3 readers-writers and corresponding labels, wherein the power module is used for converting received solar energy into electric energy to be stored in a storage battery, the sensing layer is used for providing a basic acquisition device access function and comprises a fixed reader-writer, a handheld reader-writer and a card sender, basic reading-writing operation of the labels generated by the sensing layer is a main component of block chain data in the system, the transmission layer is used for providing high-reliability and high-safety transmission of the sensing data, the data of the sensing layer is accessed to a gateway through a wired network, the data layer is used for providing the most basic data processing function for the data layer, and the data processing support is the most basic data processing support of the federal learning and block chain technology, such as data encryption, hash function operation, digital signature and other mechanisms, the service layer is used for providing services such as calculation, storage and the like for the service layer from bottom to top; a virtual machine, a high-level language editor, etc. for intelligent contract implementation; the method comprises the steps that core block chain related services are provided for tracing applications, high availability, high convenience, credible and reliable data access, accurate data calculation and safe data management of the services are guaranteed, the services are guaranteed by capability of the tracing applications, an application layer is used for providing functions of an application interface of an acquisition device, such as warehousing and warehousing, inventory, anti-fake inquiry and the like, a safety framework is used for functions including authentication and identity management, authorization, safety strategy management, privacy protection and the like related to system safety, and a supervision and audit framework is used for functions including supervision support and audit realization and the like which are responsible for normal and accurate operation of the system;
the hardware part of the traditional Chinese medicine material traceability system comprises a block chain gateway and an automatic data acquisition device, wherein the block chain gateway comprises a fixed reader-writer, a card sender, a handheld reader-writer, a contact base thereof and the like; the software part comprises a handheld reader-writer task management list running on the contact base, reader-writer management middleware running on the gateway and service management software; the system consists of a reader-writer and label layer responsible for data acquisition, a gateway layer responsible for data receiving and transmission, a server layer responsible for data storage and application support, and an application layer responsible for user interface and comprehensive display;
when in use, the traditional Chinese medicinal materials to be traced are placed in front of the camera of the acquisition equipment, the camera is controlled to acquire images of the traditional Chinese medicinal materials, the image information is transmitted to the federal learning module for training, the block chain module encrypts and stores the image information, the basic information such as the types and the drug effects of the traditional Chinese medicinal materials is acquired by comparing the data with the traditional Chinese medicinal material database acquired from the block chain module, the printing module is utilized to generate corresponding bar codes and assign codes to the related information of the traditional Chinese medicinal materials, the part of bar code information is transmitted to the inside of the portable printer, the portable printer is controlled to print, the bar codes containing the related information of the traditional Chinese medicinal materials are printed on the paster, and the paster is attached to the corresponding traditional Chinese medicinal materials or the outer package of the traditional Chinese medicinal materials, so that the tracing and assigning codes of the traditional Chinese medicinal materials are completed, the device has small and compact integral structure and is convenient to carry, the solar cell panel is high in portability, can be used for supplementing electric power when used outdoors, is high in practicability and can be used in various scenes;
the data acquisition traceability system comprises the following working procedures:
the method comprises the following steps: the application layer initiates a business process, and corresponding instructions are issued by business platform support software;
step two: splitting an instruction by service software in the gateway and issuing a reader-writer instruction which accords with a protocol to a reader-writer;
step three: the reader-writer executes the command to communicate with the tag through a protocol, and returns a corresponding result;
step four: the gateway is responsible for sorting and fusing relevant model parameters, uploading the relevant model parameters to a block chain server for uplink storage, and simultaneously feeding back relevant execution results to an application end;
step five: and the gateway carries out the on-link data auditing service at regular time, checks the authenticity and accuracy of the on-link data, and feeds back the checking result to the application terminal in real time.
The federal learning framework proposed in the traceability scene mainly comprises two modules of local learning and distributed parameter updating. In local training, the tracing device obtains a local model by executing a gradient descent algorithm added with a noise mechanism. And then sends its local model to other devices through a distributed update mechanism. In addition, the traceability device in federal learning stores all data locally and carries out data processing locally, and privacy of traditional Chinese medicine data can be guaranteed.
The proposed asynchronous federated learning scheme includes three phases: the method comprises the steps of local differential privacy gradient descent training, distributed parameter random updating and convergence speed increasing.
1) Local differential privacy: gradient descent training
As an efficient iterative optimization method, the proposed scheme utilizes gradient descent to minimize the loss function. The gradient descent minimizes the objective function by updating the parameters towards a negative gradient of the objective function. The goal of local training is to find the optimal model parameters by adjusting the model parameters to the negative direction of the gradient defined in equation (1);
in the t-th iteration, the vehicle viCalculating a local update model (parameter vector) w according to equation (2)i(t)
Wherein alpha istIs the training step proceeding in the direction of the negative gradient.
Local Differential Privacy (LDP) is a recently proposed technique that can provide higher Privacy assurance for data. Different from the traditional differential privacy technology which provides privacy guarantee in the data analysis stage, the local differential privacy concerns the privacy problem of the data acquisition device in the data acquisition process. In order to protect the privacy of updated parameters and realize local differential privacy, the proposed scheme applies a gaussian mechanism on the model updated by each tracing device, and perturbs the parameters by adding noise, as shown in formula (3).
Wherein
Is the added Gaussian noise with a mean of 0 and a standard deviation of σ · Sf. By embedding local differential privacy into the federated learning process, a traceback device viTraining noisy local parameter update w% in iteration ti(t)。
To achieve epsilon differential privacy, the total privacy budget should be limited to epsilon. Tracing device v with total iteration number of TiThe privacy budget (cost) should be allocated to each iteration, where
For example, given a fixed T, the privacy budget may be equally distributed in each iteration: epsilontT is equal to or more than 1 and is equal to or less than T. In iteration t, if the accumulated privacy cost exceeds the total budget
The entire learning process should be stopped.
By optimizing the stackThe accuracy of the generation process and the privacy cost of privacy protection are comprehensively considered, so that the privacy budget is dynamically allocated to each iteration. In the gradient optimization process of the model, the initial privacy budget is epsilon0. If the gradient contributes to the global model is larger, the training process is close to convergence, and more noise (smaller privacy cost epsilon) needs to be added in the training to protect the privacy of the model. So the privacy budget of iteration t is εt＝ε01- α. Otherwise, if the gradient contributes less to the overall model, then one needs to assign-a greater privacy cost, εt＝ε0(1+ β), i.e., less noise is added to the gradient parameters.
2) Updating distributed parameters: random gossip (sub-gossip) update mechanism
One of the challenges of deploying federated learning in a traceable network is the mobility of the traceable devices. The synchronization update mechanism between the centralized server and the client is difficult to maintain due to the movement of the tracing device. In order to alleviate the mobility problem of the tracing device and get rid of the trouble of a centralized server, the section provides a distributed random gossip updating scheme.
In each iteration, the participating tracing devices viRandom selection algorithm is operated based on communication and computing resources to select a subset traceability device by sampling
And (6) updating. In traditional federal learning, all participating users send updated parameters to a centralized server, and in the proposed federal learning, a tracing device viOnly the parameter w% is addedi(t) updating the vehicle set obtained by sampling to improve the reliability of the federal learning. Since the temporal complexity of the update is O (log (| V |)), V is thereforesIs usually taken to be a small value to reduce redundant transmission of parameters and reduce the overhead of computation and communication resources. To balance transmission delay and computational cost, V may be dynamically adjusted based on communication conditions and computational capabilities of the receiving vehiclesThe size of (2). Update mechanism due to gossipConvergence consistency can be eventually achieved, thus updating the skilful w%i(t) can be distributed to all relevant tracing devices within O (log (n)) time. In addition, the Mean Absolute Error (MAE) of the model calculated by equation (4) is also w%i(t) are broadcast together to other participating users.
(4) Wherein a lower MAE indicates a better quality of the trained model.
The proposed random gossip update mechanism prevents malicious participants from receiving the information from a certain tracing device viContinuously updating parameters to reduce the pass analysis viTo infer risk of its data information;
for a participating vehicle vjIt collects all the received update parameters and computes an aggregate model mi(t)＝mi(t-1)+∑w％i(t)。
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes in the embodiments and/or modifications of the invention can be made, and equivalents and modifications of some features of the invention can be made without departing from the spirit and scope of the invention.
Claims (7)
1. Towards portable credible device of tracing to source of chinese-medicinal material, including federal study module, block chain module, networking module, power module, display module, print module, collection module, perception layer, transmission layer, data layer, service layer, application layer, safety framework and supervision and examination framework, its characterized in that: the federal learning module comprises a query module, an acquisition module, a check module and a training module, the federal learning local training is operated on a user side, local model parameters are learned according to data of the user side, a block chain is operated on a service side, the federal learning model parameters are received and stored, the block chain module comprises an encryption module, a storage module and an authentication module, the acquisition module is used for acquiring data information of traditional Chinese medicinal materials, and the networking module is used for connecting a microprocessor and a designated platform.
2. The portable trusted traceability device for traditional Chinese medicinal materials as claimed in claim 1, wherein: the federal learning module is used for carrying out local model training on traditional Chinese medicine data, can realize a virtual common model through the parameter exchange mode under the encryption mechanism under the premise that the data can not go out of the local, guarantee the privacy of traditional Chinese medicine data, the block chain module is used for verifying staff and equipment, adopts the state secret SM2 certificate to carry out validity identity authentication, after the global training of the federal learning module is accomplished, carries out encryption protection and storage in each node of block chain network to global model parameter through the state secret algorithm SM4 algorithm, obtains the relevant block chain database of traditional Chinese medicine from the internet and carries out traceability comparison, obtains traditional Chinese medicine kind information.
3. The portable trusted traceability device for traditional Chinese medicinal materials as claimed in claim 1, wherein: the printing module is used for generating a bar code according to the Chinese medicinal material information acquired by the tracing module, assigning codes and controlling the portable printer to print.
4. The portable trusted traceability device for traditional Chinese medicinal materials as claimed in claim 1, wherein: the system comprises a power module, a sensing layer, a transmission layer and a card sender, wherein the power module is used for converting received solar energy into electric energy to be stored in a storage battery, the sensing layer is used for providing a basic acquisition device access function and comprises a fixed reader-writer, a handheld reader-writer and a card sender, 3 reader-writers and corresponding labels, basic reading and writing operations of the labels generated by the sensing layer are main components of block chain data in the system, the transmission layer is used for providing high-reliability and high-safety transmission of sensing data, and the sensing layer data are accessed into a gateway through a wired network.
5. The portable trusted traceability device for traditional Chinese medicinal materials as claimed in claim 1, wherein: the data layer is used for providing the most basic data processing functions, such as data encryption, hash function operation, digital signature and other mechanisms, is the most basic data processing support of the federal learning and block chain technology, and the service layer is used for providing a basic layer for services such as calculation, storage and the like for the service layer from bottom to top; a virtual machine, a high-level language editor, etc. for intelligent contract implementation; the core block chain related services are provided for the tracing application, high availability, high convenience, reliable and reliable data access, accurate data calculation and safe data management of the services are guaranteed, and the services are guarantee for the capability of the tracing application.
6. The portable trusted traceability device for traditional Chinese medicinal materials as claimed in claim 1, wherein: the application layer is used for providing functions of an application interface of the acquisition device, such as warehousing and warehousing, inventory, anti-counterfeiting query and the like, the safety architecture is used for functions including authentication and identity management, authorization, safety strategy management, privacy protection and the like related to system safety, and the supervision and examination architecture is used for functions including supervision support, examination realization and the like which are responsible for normal and accurate operation of the system.
7. The operation method of the portable trusted traceability device for traditional Chinese medicinal materials, as claimed in claim 1, is characterized in that: the method comprises the following steps:
the method comprises the following steps: the improved data acquisition device which needs to have the function of executing programming is adopted, the acquired data is from the information record of the complete life cycle of the data of planting production, processing and packaging, transportation and circulation, storage and sale and the like of the traditional Chinese medicinal materials, and the improved data acquisition device has the unique identification which can not be reconfigured by hardware and the functions of key control, configuration and data acquisition and non-local storage, so that information leakage is effectively prevented;
step two: the method comprises the following steps that a user carries out local training to obtain a corresponding local model, participated users find model parameters locally according to held Chinese medicinal material data through an algorithm based on gradient descent to minimize a loss function, and the model parameters are uploaded to a platform or a designated server after homomorphic encryption;
step three: the server collects model parameters trained by the client for global training, and the acquisition device is required to be ensured to operate to automatically execute an agreed intelligent contract of the block chain, and data fusion and operation audit can be performed;
step four: the updated model parameters of the server are stored in a transaction form through a block chain, data are packaged strictly according to a block chain data block format, and then data records are stored in each block chain node through an encryption algorithm and a timestamp adding mode;
step five: on the premise of not influencing the safety of the blockchain, a block chain encryption flow is optimized by adopting a state secret algorithm, so that the safer and more reliable blockchain realizes safe and credible decentralized point-to-point communication, safe transmission and privacy protection in a distribution system without a credible environment through technical means such as time stamps, information encryption and consensus mechanisms;
step six: the blockchain generates a transaction block, and each participating node collects transactions (model parameters) from the user layer and packages the transactions into blocks through cryptographic signatures. Determining the block weight through a running consensus mechanism among all nodes of the block chain, obtaining the node of the block weight, broadcasting the block to the whole network, and adding the block chain after the block chain passes the authentication;
step seven: and aggregating the model parameters of the nodes according to the records on the block chain, and updating the whole model. The model is further issued to each participating user, thereby starting a new round of training and learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111622717.6A CN114298721A (en) | 2021-12-28 | 2021-12-28 | Portable credible device of tracing to source towards chinese-medicinal material |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111622717.6A CN114298721A (en) | 2021-12-28 | 2021-12-28 | Portable credible device of tracing to source towards chinese-medicinal material |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114298721A true CN114298721A (en) | 2022-04-08 |
Family
ID=80969001
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111622717.6A Pending CN114298721A (en) | 2021-12-28 | 2021-12-28 | Portable credible device of tracing to source towards chinese-medicinal material |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114298721A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114741734A (en) * | 2022-06-09 | 2022-07-12 | 湖南大学 | Drug anti-counterfeiting traceability cloud chain data multi-party safe computing method |
-
2021
- 2021-12-28 CN CN202111622717.6A patent/CN114298721A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114741734A (en) * | 2022-06-09 | 2022-07-12 | 湖南大学 | Drug anti-counterfeiting traceability cloud chain data multi-party safe computing method |
| CN114741734B (en) * | 2022-06-09 | 2022-09-09 | 湖南大学 | Drug anti-counterfeiting traceability cloud chain data multi-party safe computing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lin et al. | A survey of application research based on blockchain smart contract | |
| Aggarwal et al. | Blockchain for smart communities: Applications, challenges and opportunities | |
| Zheng et al. | bcBIM: A blockchain-based big data model for BIM modification audit and provenance in mobile cloud | |
| Shahid et al. | Blockchain-based agri-food supply chain: A complete solution | |
| CN112132198B (en) | Data processing method, device and system and server | |
| Zhang et al. | Construction site information decentralized management using blockchain and smart contracts | |
| Hsiao et al. | Employing blockchain technology to strengthen security of wireless sensor networks | |
| Joshi et al. | Adoption of blockchain technology for privacy and security in the context of industry 4.0 | |
| Mathur et al. | A survey on role of blockchain for IoT: Applications and technical aspects | |
| Li et al. | A novel system for medical equipment supply chain traceability based on alliance chain and attribute and role access control | |
| Bordel et al. | Digital food product traceability: using blockchain in the international commerce | |
| Xiao et al. | Blockchain for secure-GaS: Blockchain-powered secure natural gas IoT system with AI-enabled gas prediction and transaction in smart city | |
| Saranya et al. | Proof of transaction (PoTx) based traceability system for an agriculture supply chain | |
| Wei | Blockchain-based data traceability platform architecture for supply chain management | |
| CN113011883A (en) | Data processing method, device, equipment and storage medium | |
| Li et al. | An exploratory study on the design and management model of traditional chinese medicine quality safety traceability system based on blockchain technology | |
| Qureshi et al. | Blockchain applications for the Internet of Things: Systematic review and challenges | |
| Hou et al. | IoT and blockchain-based smart agri-food supply chains | |
| Li et al. | The design of power grid data management system based on blockchain technology and construction of system security evaluation model | |
| Lee et al. | SPChain: a smart and private blockchain-enabled framework for combining GDPR-compliant digital assets management with AI models | |
| Shen et al. | Blockchains for artificial intelligence of things: A comprehensive survey | |
| Kiruthika et al. | Fusion of IoT, blockchain and artificial intelligence for developing smart cities | |
| Sathish et al. | A survey on Blockchain mechanisms (BCM) based on internet of things (IoT) applications | |
| Xie et al. | Dual-chain blockchain in agricultural E-commerce information traceability considering the viniar algorithm | |
| CN114298721A (en) | Portable credible device of tracing to source towards chinese-medicinal material |
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 |