Peer-to-peer network for collaborative computing and peer-to-peer computing network for non-specific feature recognition

Abstract

The present invention relates to a peer-to-peer network for collaborative computing and a peer-to-peer computing network for non-specific feature recognition. There is no primary or secondary relationship among all node devices, there is no fixed connection path between node devices, node devices only receive the calculation results of other node devices, and send out the result data of their own calculations. The discovery of events and / or the corresponding execution devices do not rely on a single node device for recognition and control, but are jointly confirmed through collaborative computing by multiple node devices in the peer-to-peer network. The present invention fundamentally changes the nature of traditional information-based single-point security sensitivity, with reliable data sources and calculation processes, high execution efficiency, low hardware requirements, and high recognition accuracy. It realizes non-specific feature recognition of a target without the need for specific feature recognition and without the need to obtain the identity information of the target.

Description

Technical Field

[0001] The present invention relates to the fields of network technology and data processing technology. More specifically, it relates to a peer-to-peer network for collaborative computing and a peer-to-peer computing network for non-specific feature recognition. Background Art

[0002] For identification and monitoring over a large geographical area, such as the monitoring scope at the metropolitan level, when performing identity identification or event confirmation through feature recognition, including identity identification or event confirmation of objects such as vehicles, road surfaces, public facilities, pedestrians, events, environments, weather, and noises, not only is the accuracy low, but also the credibility of the data cannot be guaranteed.

[0003] In the prior art, identity confirmation or recognition usually adopts a point-by-point method for identity confirmation or identity recognition, that is, recognition is only performed at one location or repeated recognition is performed at multiple locations.

[0004] For example, the method for confirming or recognizing a person's identity is: performing user password verification or biometric recognition at the location where recognition is required; the disadvantages are: on the one hand, there are significant security risks. When the password is stolen or the face is imitated, whether recognition is performed at one location or repeated recognition is performed at multiple locations, there is a risk of identity impersonation; on the other hand, it requires the user to actively output information, and the user experience is poor.

[0005] For example, the method for confirming or recognizing a vehicle's identity is: feature recognition; the disadvantages are: the recognition result is prone to disconnection, the data source is not completely credible, the requirement for the recognition technology is high, but the recognition accuracy is low.

[0006] Moreover, the single-point recognition method requires high-speed recognition operations at each point, which has high requirements for hardware. As the data volume increases, the execution efficiency will also be reduced.

[0007] Regarding the prior art of metropolitan-level intelligent management, taking the city brain as a typical example, from the overall architecture perspective, it is still a traditional information-based single-point aggregation type computing mode; the city brain is to set up a centralized service platform and multiple stations with different functions and different functions. Between stations and between stations and the service platform, they are connected in a non-peer-to-peer manner with a relatively fixed connection and layer-by-layer aggregation relationship; each station directly or indirectly sends the processing result or original data of the event to the service platform, and the service platform processes and coordinates the whole network to achieve the so-called purpose of intelligent management. However, in the technical solution of the city brain, the functions of each station are actually still the technical means of single-point recognition, and the coordination between each station is actually the overall management of the centralized platform.

[0008] In addition to the above-mentioned deficiencies of single-point recognition in each station, the technical solution of the city brain as a whole also has the following deficiencies:

[0009] Weak anti - attack ability: including being extremely vulnerable to single - point attacks and easy leakage of privacy data;

[0010] Difficulty in ensuring data security and authenticity: including easy data tampering, false information, high costs for discovering and correcting incorrect information, great governance difficulty, and hard to hold someone accountable for data problems, and hard to respond in a timely manner;

[0011] Existence of data islands: Data between departments and sites cannot be fully shared, with high sharing costs and high network security risks, thus restricting the maximum utilization of information efficiency. Summary of the Invention

[0012] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a peer - to - peer network for collaborative computing and a peer - to - peer computing network for non - specific feature recognition, which fundamentally changes the nature of single - point security sensitivity in traditional informatization. The data source and calculation process are trustworthy, with high execution efficiency, low hardware requirements, and high recognition accuracy. Without the need for specific feature recognition and without obtaining the identity information of the target, non - specific feature recognition of the target can be achieved.

[0013] The technical solution of the present invention is as follows:

[0014] A peer - to - peer network for collaborative computing includes multiple node devices, and there is no primary - secondary relationship among all node devices; for a certain node device, the raw data collected is processed to obtain result data, and the result data is propagated to other node devices; other node devices that receive the result data use the result data as one of the collected raw data, and the result data has an impact on the result data of other node devices; based on this, multiple node devices in the peer - to - peer network perform collaborative computing to achieve the discovery of events and / or respond to corresponding execution devices.

[0015] Preferably, when the result data of collaborative computing can determine a certain event, the discovery of the event is completed.

[0016] Preferably, for a certain node device, its result data is selectively sent to specific other node devices.

[0017] Preferably, some or all node devices are provided with at least one type of sensor for collecting different corresponding types of perception data; the current node device receives the result data output by other node devices; for the current node device, the perception data collected is combined with the result data from other node devices, and the result data of the current node device is calculated and sent to other node devices; the node devices in the peer - to - peer network perform collaborative computing with the collection of perception data and the calculation of result data.

[0018] Preferably, the node device is deployed with a data processing model for calculating result data; the data processing model adjusts model parameters according to the result data received each time.

[0019] Preferably, if the node device is deployed with an identification model, when the node device senses an event, it uses the identification model to preliminarily process the raw data to obtain identification data corresponding to the event; the identification data is used as the input of the data processing model.

[0020] Preferably, the data processing model loaded on the node device performs adaptive perception calculation on the data processing model of the previous node device. Specifically, when a certain node device calculates and processes the result data from other node devices, the adaptive perception calculation includes performing verification calculation on the result data of the current node device and other related node devices and / or verification calculation on the historical result data of the current node device to obtain an adaptability value.

[0021] Preferably, when the data processing model of a certain node device cannot adapt to the sensors and actuators connected thereto, the influence of the adaptability value of its adaptive perception calculation on the current result data of the current node device is as follows: in the current result data of the current node device, a preset flag bit is assigned; or, when the result data obtained by other related node devices performing comprehensive calculation on the current result data of the current node device according to their own data processing models can enable subsequent node devices to use the result data as input, after the result data is calculated by other node devices, it drives the adjustment of the data processing model of the current node device.

[0022] Preferably, in the calculation of the current result data or several data transmissions and result data calculations, the data processing model of the node device covers the following modes:

[0023] When it is found that the adaptability value of the data processing model of a certain node device exceeds the degree threshold, the node device outputs result data to add the node device connected to the model training facility to the receiving node list. When the node device connected to the model training facility receives such data, its result data calculated in that instance will include a drive command for driving the model training facility.

[0024] The model training facility is regarded as a type of sensor, and the signals it emits are added to the connected node devices. After the result data of the node devices is calculated and emitted, the corresponding node devices in the peer-to-peer network calculate the result data and then drive the data storage module of the connected sensors to send the sensed data specified by the model training facility to a specific location through establishing a file transfer channel between the node devices in the peer-to-peer network or in other network communication modes to participate in model training and improvement. The improved data processing model will be deployed to the node devices that the original data processing model cannot adapt to in the same way.

[0025] Preferably, or, the model training facility drives multiple node devices to generate a new data processing model by means of federated computing using their own sensed data and updates the data processing model.

[0026] Preferably, the degree threshold is a set value or a dynamic value; if it is a dynamic value, it is determined by a preset program or by a dynamically updated data processing model.

[0027] Preferably, the principle of data transfer between node devices is: when the result data of the current node device changes or the change amount of the result data reaches a preset threshold, the latest result data is sent to other node devices;

[0028] Or, the current node device continuously sends the result data of each time to other node devices at a certain frequency.

[0029] Preferably, all node devices encrypt the result data they calculate based on an encryption consensus mechanism to obtain an encrypted result, and then send the encrypted result to other node devices.

[0030] Preferably, the encryption consensus mechanism includes one or more consensus mechanisms, and different consensus mechanisms correspond to changing the encryption algorithm structure and parameters of the node devices.

[0031] Preferably, the discovery of events by the peer-to-peer network includes the content of the event, the location where the event occurs, and the corresponding response handling.

[0032] Preferably, node devices communicate with standard-sized data packets.

[0033] Preferably, the peer-to-peer network deploys a QoS mechanism, and the QoS mechanism gives priority to ensuring the transmission quality of the result data between node devices.

[0034] Preferably, the networking mode of the peer-to-peer network is one or a combination of several of the 4G mode, 5G mode, or MESH mode.

[0035] Preferably, the MESH mode is based on the LTE standard and communicates at the physical layer of the LTE standard; data is carried in a customized frame structure and interaction is carried out using a dedicated wireless communication protocol.

[0036] A peer-to-peer computing network for non-specific feature recognition. Based on the peer-to-peer network, node devices are provided with data acquisition devices and computing modules. The data acquisition device includes at least one type of sensor; node devices located at different acquisition positions acquire at least one point sample of the object to be recognized, and the point sample is the sensed data corresponding to the sensor type; without the need to obtain the identity information of the object to be recognized, collaborative computing is carried out between node devices to determine that each unique object to be recognized is itself, realizing non-specific feature recognition.

[0037] Preferably, in the peer-to-peer network, for a certain point sample of a certain object to be recognized, in the result data transmitted from the node device that acquired the point sample to other node devices, subsequent node devices adjust their sensing attention according to the characteristics of the point sample, or report the characteristics of the point sample for subsequent node devices to adjust their sensing attention; if subsequent other node devices do not detect the characteristics of the point sample, but it can be determined from the characteristics of other point samples that the characteristics of the undetected point sample still belong to the object to be recognized, then continue to express the characteristics of the undetected point sample in the result data of the current node device and transmit it to other node devices.

[0038] Preferably, the method of reporting the characteristics of the point sample for subsequent node devices to adjust their sensing attention is: for the result data expressing the characteristics of the point sample provided by the previous node device, or the characteristics of the point sample adjust the parameters of the data processing model of the subsequent node device, so that the subsequent node device improves its computing power for identifying the characteristics of the point sample; or, the subsequent node device uses the sensing attention model to match the characteristics of the received point sample or the result data expressing the characteristics of the point sample for computing power adjustment.

[0039] Preferably, when a node device processes the result data output by several previous node devices, based on the data processing model, when the objects to be recognized described by several previous node devices can be determined to be the same target through certain common point sample characteristics, the point sample characteristics and other information described by each node device are merged into the same target.

[0040] Preferably, when the result data received by a node device indicates that before the current node device received the result data this time, the flag used to identify the object to be recognized is different from the flag used by other node devices to identify the object to be recognized, and the flag assigned by other node devices for the object to be recognized is updated, then the flag used by the current node device to identify the object to be recognized before receiving the result data this time is converted.

[0041] Preferably, the method for converting the flag used to identify the object to be recognized by the current node device before receiving the result data for the current time is as follows:

[0042] Replace the flag used to identify the object to be recognized by the current node device before receiving the result data for the current time with the flag assigned to the object to be recognized by the latest other node device;

[0043] Alternatively, record the conversion relationship between the flag used to identify the object to be recognized by the current node device before receiving the result data for the current time and the flag assigned to the object to be recognized by the updated other node device, and perform the conversion when the result data received by the current node device for the current time needs to be referenced;

[0044] Alternatively, the node device deploys a conversion model, and performs corresponding conversions on the flags of multiple objects to be recognized according to the input original data or result data.

[0045] Preferably, for one or more point samples collected successively by node devices at different acquisition positions, if the characteristic values of a certain one or more point samples at different acquisition positions respectively meet the preset similar conditions or are determined by a specific model to have a correlation reaching the threshold, and are unique at each acquisition position, it is determined that the point samples of this type at different acquisition positions have relevance.

[0046] Preferably, for one or more point samples collected simultaneously by node devices at different acquisition positions, if the node devices at different acquisition positions collect the same spatial field, when there is only a single object to be recognized in the spatial field, or the collected point samples can correctly point to one of the multiple objects to be recognized to which they belong, then for a certain object to be recognized, the one or more point samples collected by the node devices at different acquisition positions have relevance.

[0047] Preferably, the data acquisition device of the node device includes one or a combination of an image acquisition device, an electromagnetic induction device, a temperature measurement device, a vibration frequency sensing device, and a lidar. The data collected by the above devices and the three-dimensional point cloud collected by the lidar or the point cloud generated from the images collected by multiple image acquisition devices are jointly calculated to obtain three-dimensional points with data; the image color, contour, line, reflectivity, motion trend, electromagnetic characteristics, temperature, temperature change trend, vibration frequency, and vibration frequency change trend based on two-dimensional perception are used as additional attributes of the corresponding three-dimensional points to form a three-dimensional point cloud with attributes; the corresponding relationship between each region of the three-dimensional point cloud with attributes and each part or associated part of the 3D appearance of the object to be recognized is determined by combining the change characteristics of electromagnetic induction, temperature law, vibration frequency, motion correlation, and reflectivity.

[0048] Preferably, when it is determined that the identity information of the object to be recognized needs to be obtained, an identity information acquisition command is triggered, and the identity information acquisition command is used as one of the inputs to participate in the calculation of the result data of the node device. By driving the node device in the peer-to-peer network that is connected with the barrier-free data acquisition condition capable of obtaining the identity information of the object to be recognized to respond to the corresponding result data, the identity information of the object to be recognized is obtained.

[0049] Preferably, the peer-to-peer network verifies the authenticity of the identity information of the object to be recognized, and then determines its authority; among them, the node device in the peer-to-peer network that can obtain the identity information does not provide the identity information, and only expresses the verification result in the result data of the node device according to the verification requirement of the authenticity of the identity information in the received result data.

[0050] Preferably, the peer-to-peer network verifies the relationship between the true identity and the ownership of the requester of the demand, and determines whether its true identity has the execution authority for the current demand. Specifically, the peer-to-peer network directly performs collaborative operations on the actual behavior, trajectory, and data to determine the relationship between its true identity and the execution authority of the current demand. Other node devices perform subsequent calculations on the relationship between the true identity they obtain and the execution authority of the current demand, and express the relationship between the true identity of the requester of the demand and the execution authority of the current demand in the result data and transmit it to the subsequent node devices in the next layer. Among them, it is not necessary to output the specific identity information of the requester of the demand to other node devices.

[0051] Preferably, the node device in the peer-to-peer network that can obtain the identity information does not provide the identity information, and drives the information source device that provides the identity information to establish an encrypted file transmission channel or an encrypted information transmission channel in other network communication modes between the input terminal of the node device that needs to obtain the identity information; and uses the identity information as one of the inputs of the node device.

[0052] Preferably, the data acquisition device includes one or more of an image acquisition device, an audio acquisition device, a temperature measurement device, a vibration frequency sensing device, a lidar, a chemical sensor, and an electromagnetic induction device.

[0053] The beneficial effects of the present invention are as follows:

[0054] In the peer-to-peer network for collaborative computing according to the present invention, there is no primary or secondary relationship among all node devices, there is no fixed connection path among node devices, a node device only receives the calculation results of other node devices, sends out the result data of its own calculation, and the discovery of an event and / or the corresponding execution device does not rely on a single node device for identification and control, but is jointly confirmed through collaborative computing by multiple node devices in the peer-to-peer network. Furthermore, the present invention does not rely on single-point identification, distributes the computing function throughout the entire network, reduces the software and hardware requirements for single-point computing, has high execution efficiency, and greatly improves the anti-attack ability; there is a relatively symmetric information state among node devices, which can immunize the problem of illegal data tampering. Even if a single node device is physically cracked and the data it sends out is tampered with, because the whole-network computing is a super-highly redundant complex calculation and a super-multi-dimensional verification, the data sent out by a single node device being tampered with does not affect the whole-network computing result, and the faulty and tampered node device can be quickly located, ensuring the credibility of the whole-network computing result. Furthermore, the contradiction between data sharing and information security among departments can be solved.

[0055] The result data transmitted among node devices is the processing result of information, rather than the information itself. Furthermore, the original data collected (i.e., sensed data) does not need to be stored. A node device only receives the calculation results output by other node devices and sends out its own calculation results. The information contained in a single calculation result is not sufficient to restore any event and target information. A definite result can be obtained only through collaborative computing jointly by the calculation results on the entire peer-to-peer network, the elements of the multi-dimensional data matrix, and the corresponding relationship between the physical space and facilities. The collaborative computing has less dependence on the information transmitted by a small number of node devices, and thus can fundamentally change the nature of single-point security sensitivity in traditional informatization.

[0056] In the present invention, as one of the node devices, the execution device responds and executes based on the calculation result obtained through collaborative computing, with high response efficiency, avoiding illegal responses such as false execution or non-execution when it should execute due to network attacks. To avoid hijacking, the present invention can also use multiple node devices to cooperate in controlling the execution device, further improving the immunity to hijacking attacks.

[0057] The peer-to-peer computing network for non-specific feature recognition according to the present invention is based on the peer-to-peer network for collaborative computing. Without the need for specific features or obtaining specific identity information, each unique object to be recognized can be determined as itself, realizing non-specific feature recognition. The present invention conducts target recognition, identity confirmation, ownership calculation, or event monitoring through non-specific feature recognition, not only with high accuracy of recognition results, but also precise location recognition. The present invention can recognize identities and ownership relationships without relying on specific features, protecting privacy while solving long-standing urban problems such as transportation, education, medical convenience, epidemic prevention, people's livelihood services, emergency response, public security, anti-terrorism, community management and services, market behavior, work safety, and civilized behavior.

[0058] On the other hand, the present invention adopts non-specific feature recognition, which can effectively prevent risks caused by theft or imitation of specific features, greatly enhancing security. The present invention adopts a non-contact passive method to perform non-intrusive recognition of the identity of the object to be recognized, greatly enhancing the convenience of execution. Based on the peer-to-peer network described above, the coverage range can be easily arranged at the level of hundreds of meters or hundreds of kilometers, suitable for various levels of geographical ranges.

[0059] For the recognition of the target, confirmation of identity and ownership, or monitoring of events by the present invention, since the original data is not transmitted and the original data can be not stored, there is no risk of leakage of privacy information; correspondingly, the contradiction between public security and privacy protection can be solved. Specific embodiments

[0060] The present invention will be further described in detail below in conjunction with the embodiments.

[0061] In order to solve the deficiencies of the traditional information-based single-point aggregation computing mode in the prior art, the present invention provides a peer-to-peer network for collaborative computing and a peer-to-peer computing network for non-specific feature recognition, solving the contradiction between public security and privacy protection, solving the contradiction between data sharing and information security among departments, and providing a safer and more convenient intelligent service while eliminating false information and fraud behaviors to the greatest extent. Without relying on specific features for identity recognition and protecting privacy, the present invention solves long-standing urban problems such as transportation, education, medical convenience, epidemic prevention, people's livelihood services, emergency response, public security, anti-terrorism, community management and services, market behavior, work safety, and civilized behavior.

[0062] The peer-to-peer network for collaborative computing according to the present invention includes multiple node devices. There is no primary or secondary relationship among all node devices, forming a decentralized network and computing architecture. Different from the traditional information-based single-point aggregation computing mode, there is no fixed or preset path relationship for the data transfer direction among the node devices of the present invention. In the peer-to-peer network of the present invention, for a certain node device, the original data collected is processed to obtain result data, and the result data is propagated to other node devices; other node devices that receive the result data use the result data as one of the original data collected, and the result data has an impact on the result data of other node devices through the result data. For the convenience of description, the aforementioned "certain node device" is hereinafter referred to as the "current node device", and the "other node devices" are referred to as "subsequent node devices". Then one of the impacts is that the result data calculated by the subsequent node devices is not completely determined by the original data collected by itself, but is jointly determined by the result data output by the current node device; among them, the result data output by the current node device may change the data processing model and parameters, etc. used by the subsequent node devices to calculate the result data, thereby affecting the result data of the subsequent node devices. For example, if there is a correlation between the result data output by the current node device and the original data collected by the subsequent node device, it is necessary to consider the impact of the result data output by the current node device on the accuracy of the result data of the subsequent node device; specifically, for the perception of a certain specific target, if the result data calculated only based on the original data collected by the subsequent node device can only reflect the real-time (including real-time location and time) single-point result judgment of the target within the perception range of the subsequent node device; while the result data output by the current node device reflects the direct perception data and result judgment of the target at other locations and other times, or other perception data and result judgments that are indirectly related, which helps to improve the accuracy and comprehensiveness of the result data of the subsequent node device, including the superposition calculation of the same dimension and the correlation reference of different dimensions.

[0063] Since there is no master-slave relationship among the node devices in the peer-to-peer network, point-to-point transmission can be carried out between the node devices. Furthermore, for the calculation result corresponding to a certain perception data of a certain target reflected in the result data output by a certain node device, the information is relatively symmetric among other node devices that receive the result data; other node devices use the received result data as input, combine the perception data of their own sensors, calculate their own result data, and the own result data naturally covers the received result data and the information reflected by their own sensors, and transmits it to other node devices in the next layer. Furthermore, for a certain perception data of a certain target, the information is in a relatively symmetric state among all node devices, which can immunize the impact on the result data caused by the tampering and forgery of the calculation process and calculation result of a single node device, and at the same time become a means to discover faulty or tampered node devices and node devices with non-compliant performance, fundamentally solving the fundamental hidden danger of traditional information technology, that is, false information, forged information, and wrong information caused by information asymmetry, and then becoming the entry point for fraud and cyber attacks, with poor accuracy, long time consumption, poor credibility, and poor adaptability in complex comprehensive applications. Furthermore, it can truly become the information infrastructure for urban comprehensive management and the infrastructure for the digital economy. Different from the blockchain technology that still uses each node to calculate independently, determine the result, and focuses on the technical solution of storing the original data, the present invention focuses on the peer-to-peer collaborative calculation among node devices. Through peer-to-peer collaborative calculation, each node device can adjust its own data processing model (i.e., the algorithm for calculating the result data) and parameters when processing data. The adjustment is the feedback of all node devices to their own adjustments, so that the calculations of all node devices become a whole. Each node device no longer completes the calculation independently, but all node devices complete the calculation together. After the data processing model of the node device is adjusted, it is an objective adjustment and will affect the next data processing.

[0064] Among them, for a certain node device, according to the implementation requirements, its result data can be selectively sent to specific other node devices. Specifically, in implementation, the result data of a certain node device does not necessarily need to be received by all other node devices, and some node devices can be selected as the subsequent node devices in the next layer to receive.

[0065] In the present invention, some or all of the node devices are provided with at least one type of sensor for collecting perception data of different corresponding types; the node devices in the peer-to-peer network collect any perception data (i.e., raw data) that can be perceived within their perception range through all the sensors they are provided with; at the same time, any node device also receives the result data output by other node devices, calculates its own result data and outputs it; then, continuous perception is carried out in the entire peer-to-peer network, and data collection, input, output, calculation, etc. are performed. Based on this, multiple node devices in the peer-to-peer network perform collaborative computing to achieve the discovery of events and / or the response of corresponding execution devices. In the present invention, as one of the node devices, the execution device responds and executes based on the calculation result obtained by collaborative computing, with high response efficiency, avoiding illegal responses such as false execution caused by network attacks or non-execution when it should be executed. To avoid hijacking, the present invention can also use multiple node devices to cooperate in controlling the execution device to further improve the immunity to hijacking attacks.

[0066] Specifically, taking a certain node device as the current node device, in combination with the data transfer of its previous node device and subsequent node device (the previous node device and subsequent node device in the present invention are only used to describe the front and back relationships with the current node device during the current calculation and data transmission process, and do not indicate an inevitable front and back relationship and priority relationship between them), correspondingly, the current node device receives the result data output by other node devices (including the previous node device), and the subsequent node device receives the result data output by other node devices (including the current node device). For the current node device, the collected perception data is combined with the result data from other node devices (including the previous node device) to calculate the result data of the current node device, and is sent to other node devices (including the subsequent node device). Similarly, the working process of the subsequent node device is the same as that of the current node device, and the previous node device also receives the result data of the previous node device of the previous node device and performs the same working process as the current node device; that is, the node devices in the peer-to-peer network perform the same working process. Furthermore, the node devices in the peer-to-peer network perform collaborative computing as the perception data is collected and the result data is calculated. Among them, for the result data output by a certain node device, only the subsequent layer of node devices receives it and uses it as input, and the result data of the subsequent layer of node devices will cover the result data of the previous layer of node devices (including the aforementioned certain node device).

[0067] Among them, when the result data of collaborative computing can determine a certain event, the discovery of the event is completed. In this embodiment, the discovery of events by the peer-to-peer network includes the content of the event, the location where the event occurs, and the corresponding response handling, etc.

[0068] In a peer-to-peer network, all events are processed synchronously, and it is not necessarily required to explicitly generate phased result outputs such as what events are discovered and what the specific content of the events is; in a peer-to-peer network, all that is clear is the sensing of sensors and the corresponding execution devices to respond. All other intermediate processes are processed simultaneously through collaborative computing. That is, during the operation of the present invention, the intermediate process of event discovery is imperceptible. As the collaborative computing progresses and the node devices obtain the result data, the corresponding execution devices automatically respond and execute.

[0069] Furthermore, in addition to using the collaborative computing of the peer-to-peer network to achieve the discovery of a target or an event, in this embodiment, if an identification model is deployed on the node device, the single-point identification function of the node device can be realized; furthermore, when the node device senses an event, the identification model is used to preliminarily process the raw data to obtain the identification data corresponding to the event; the identification data is used as the input of the data processing model. The node device realizes the single-point identification function through the identification model, which can replace the collaborative computing in application scenarios with a small number of node devices and low requirements, and can be used as a functional supplement to the present invention.

[0070] In the present invention, a data processing model is deployed on the node device, and the data processing model is used to calculate and obtain the result data. During the collaborative computing process of each node device, due to the different physical installation locations, connected sensors and execution devices, functional differentiation gradually forms. When each node device processes data (collecting raw data, receiving the result data output by the previous node device, and calculating its own result data), it may change the architecture and parameters of its own data processing model. That is, the data processing model adjusts the model parameters according to the result data received each time, reflecting the mutual influence between node devices through the output result data; just like human neurons, which will gradually change some of their biological characteristics according to the electrical signals conducted by other neurons, thus reflecting human memory and learning as a whole. The said influence includes improving the accuracy of the result data and continuously increasing the perception degree of the target or event.

[0071] In order to accurately reflect the mutual influence between node devices and determine whether the data processing model of the current node device is suitable for the sensors and actuators it is connected to, in the present invention, the data processing model loaded on the node device performs adaptive perception calculation on the data processing models of the previous node devices. Specifically, when a certain node device performs calculation processing on the result data from other node devices, the adaptive perception calculation includes performing verification calculation on the result data of the current node device and other related node devices and / or verification calculation on the historical result data of the current node device to obtain an adaptability value or serve as the input to the adaptability judgment model. In the present invention, based on collaborative calculation, the result of the verification calculation is reflected in the result data of the node device, so that the result data of a series of nodes can drive the model training facility to make a greater degree of adjustment to the data processing model of the problem node device. The implementation manner of the adaptability value can be used as an implementation manner to replace the aforementioned adjustment manner based on collaborative calculation.

[0072] When the data processing model of a certain node device cannot adapt to the sensors and actuators it is connected to, the influence of the adaptability value of its adaptive perception calculation on the current result data of the current node device is as follows: in the current result data of the current node device, assign a value to a preset flag bit; or, when other related node devices perform comprehensive calculation on the current result data of the current node device according to their own data processing models, and the obtained result data can be used as the input for subsequent node devices, after the result data is calculated by other node devices, it drives the adjustment of the data processing model of the current node device.

[0073] For the data processing model, the adaptive perception calculation is covered in other calculations (including the calculation of result data, the adjustment of the data processing model, etc.). That is to say, only one round of calculation is performed, which includes various other calculations and the adaptive perception calculation, that is, the adaptive perception calculation is performed simultaneously with the calculation of the result data.

[0074] In the present invention, the adjustment of the data processing model occurs in the current result data calculation or several data transmissions and result data calculations. Specifically, the data processing model of the node device should cover the following modes:

[0075] Mode 1: When it is found that the adaptability value of the data processing model of a certain node device exceeds the degree threshold, the result data output by the node device will add the node device connecting to the model training facility to the receiving node list. When the node device connecting to the model training facility receives such data, the drive command for driving the model training facility will be included in the result data of its current calculation. Among them, according to different implementation requirements, the degree threshold is a set value or a dynamic value. If the degree threshold is a dynamic value, it is determined by a preset program or by a dynamically updated data processing model. In the present invention, based on collaborative computing, during the calculation process of the result data by the node device, the model training facility can be driven to adjust the data processing model of the problem node device. The implementation method of the drive command can be used as an implementation method to replace the aforementioned adjustment method based on collaborative computing.

[0076] Mode 2: The model training facility is regarded as a type of sensor, and the signal it emits is added to the connected node device. After the result data of the node device is calculated and sent out, the corresponding node device in the peer-to-peer network, after calculating the result data, will drive the data storage module of the connected sensor, and send the perception data specified by the model training facility to a specific location through establishing a file transfer channel between the node devices in the peer-to-peer network or establishing a file transfer channel in other network communication modes to participate in model training and improvement. The improved data processing model will be deployed to the node devices that the original data processing model cannot adapt to in the same way.

[0077] As another implementation method, in the present invention, the model training facility can also drive multiple node devices to generate a new data processing model in a coalition computing manner using their own perception data and update the data processing model.

[0078] In the present invention, different data transfer principles can be selected according to different implementation requirements, including application scenarios, implementation types and quantities of sensors, etc.; specifically, the principle of data transfer between node devices is: when the result data of the current node device changes, or the change amount of the result data reaches a preset threshold, the latest result data will be sent to other node devices; or, the current node device continuously sends the result data of each time to other node devices at a certain frequency.

[0079] To ensure the further credibility of the data source and the calculation process, in the present invention, all node devices encrypt the result data obtained by their calculations based on an encryption consensus mechanism to obtain an encrypted result, and then send the encrypted result to other node devices. The encryption consensus mechanism includes one or more consensus mechanisms, and different consensus mechanisms correspond to changing the encryption algorithm structure and parameters of the node device.

[0080] Node devices communicate with each other using standard-sized data packets (i.e., result data or calculation results). In the present invention, the node devices of the peer-to-peer network are similar to human neurons. Each neuron does not transmit specific data that directly describes external events. Similarly, the node devices do not output raw data, but process the raw data obtained by the connected sensors and data acquisition devices into standard-sized data packets (i.e., result data or calculation results, similar to the nerve impulses of neurons) according to their own data processing models (similar to the biological characteristics of nerve cells). The amount of information contained in a single data packet is not enough to restore any event and target information. The calculation results, multi-dimensional data matrix elements, and physical space and facility correspondences on the entire peer-to-peer network must be jointly calculated in a collaborative manner to obtain a definite result. The collaborative calculation is not very dependent on the data output by a few node devices, and the collaborative calculation simultaneously processes the needs received or initiated by all node devices. It is a collaborative verification calculation of super-multi-dimensional related information, which can fundamentally change the nature of traditional information technology that is sensitive to single-point security.

[0081] In order to ensure the integrity of data and the effective execution of collaborative computing, in the present invention, the peer-to-peer network deploys a QoS mechanism, and the QoS mechanism is to give priority to ensuring the transmission quality of the result data between node devices.

[0082] In the specific implementation, the networking mode of the peer-to-peer network is one or a combination of 4G mode, 5G mode or MESH mode, so as to adapt to different application scenarios, comprehensively implement feasibility, cost considerations and other factors, and achieve the best solution. Among them, the MESH mode is based on the LTE standard, and communicates at the physical layer of the LTE standard; it carries data with a customized frame structure and interacts with a dedicated wireless communication protocol. The frame structure is customized to be suitable for peer-to-peer network computing, and a proprietary wireless communication protocol developed for urban agglomeration peer-to-peer network computing is adopted to further improve its security and reliability. In addition, the wireless algorithm fully adapts to the multipath channel environment based on consensus mechanism control required for peer-to-peer network computing. The communication distance in the city is 100 meters to 10 kilometers, and the omnidirectional antenna can also achieve efficient transmission of 120 kilometers in the field. In this embodiment, the communication distance of the Mesh network is: the distance between node devices indoors is 50-150 meters, the distance between node devices outdoors is 50 meters to 120 kilometers, and the number of node devices that can be accessed by each node is 65535. In addition, when networking in 4G mode or 5G mode, there is no limit on the communication distance, and the number of accessible node devices depends on the computing power of the computing chip and the communication delay.

[0083] In the peer-to-peer network for collaborative computing according to the present invention, the result data calculated and output by the node devices can be implemented as a state corresponding to the sensed data (i.e., the original data), which can be represented by a state value. Furthermore, the node devices do not need to store and transmit the original data. Based on the technical characteristics of the peer-to-peer network, it can be applied to various usage scenarios that provide targeted services or controls for a certain target or event. Based on this, since the result data transmitted between the node devices is the processing result of information, rather than the information itself, the original data (i.e., the sensed data) collected can be not stored. The node devices only receive the calculation results output by other node devices and send out their own calculation results. The amount of information contained in a single calculation result is not sufficient to restore any event and target information. It is necessary to jointly perform collaborative computing with the calculation results on the entire peer-to-peer network, the elements of the multi-dimensional data matrix, and the corresponding relationship between the physical space and facilities to obtain a definite result. The collaborative computing has less dependence on the information transmitted by a small number of node devices, and thus can fundamentally change the nature of the traditional information-based single-point security sensitivity.

[0084] In the present invention, since the state evolution of the result data output by the previous node devices is reflected in the result data output by each node device, based on the result data received by the current node device, the behavior, attributes, state, or event of the target when it was sensed by the previous node devices can be deduced. For example, when it is necessary to find the position of target a 15 minutes ago, the position corresponding to the node device that sensed target a at the current moment can be obtained, and then the position where target a is located can be inferred; then, according to the transmission path of the result data, it can be deduced back 15 minutes to determine the position where target a was located 15 minutes ago (determined by the node device that sensed target a); furthermore, the node device does not need to store the original data about target a. That is, based on the present invention, the search for target a can be achieved without identifying the original data, but by first deducing the node device that sensed target a, and when necessary, obtaining the original data about target a at the moment when it needs to be searched from the storage device connected to the node device.

[0085] Based on the peer-to-peer network described above, the present invention also provides a peer-to-peer computing network for non-specific feature recognition. The node device is provided with a data acquisition device (in specific implementation, it may include one or more of an image acquisition device, an audio acquisition device, a temperature measurement device, a vibration frequency sensing device, a lidar, a chemical sensor, and an electromagnetic induction device), and an operation module. The data acquisition device includes at least one type of sensor. The operation module calculates and obtains result data based on a data processing model. Node devices set at different acquisition positions (i.e., located at different physical installation positions) acquire at least one point sample of the object to be recognized, and the point sample is the sensed data corresponding to the sensor type. Based on the peer-to-peer network described above, the present invention can perform collaborative computing among node devices without obtaining the identity information of the object to be recognized, and determine that each unique object to be recognized is itself, realizing non-specific feature recognition. The "non-specific feature recognition" described in the present invention is different from the "recognition" in the usual sense in the strict concept definition. The "recognition" in the usual sense means determining the concrete image or specific identity information of the target, such as who exactly (including name, specific information indicating the identity of the target), what it is (such as a car, a person, etc.). However, the "recognition" in the "non-specific feature recognition" described in the present invention means determining each unique target as itself; that is, for a certain object to be recognized, its existence is unique. After the present invention realizes "non-specific feature recognition", it is determined that the object to be recognized is itself, rather than other objects to be recognized. The result of "non-specific feature recognition" does not need to determine the specific features of the object to be recognized, nor does it need to determine the identity information or concrete image of the object to be recognized. For example, for a person regarded as the object to be confirmed A, and for an object regarded as the object to be confirmed B, after realizing "non-specific feature recognition", it is not necessary to recognize that the object to be confirmed A is a person and who the specific identity is, nor to recognize that the object to be confirmed B is an object and what specific item it is; instead, it is necessary to determine that the object to be confirmed A is the object to be confirmed A itself, and the object to be confirmed B is the object to be confirmed B itself. Then, corresponding services or controls can be performed on the object to be confirmed A or the object to be confirmed B.

[0086] When an event is detected within the sensing range of the peer-to-peer computing network, based on the technical concept of the present invention, usually it is not directly detected by a certain node device, but after the data collected by multiple node devices are subjected to collaborative computing across the network, it is determined what sensing data changes occur at which sensing positions of which nodes; among them, all node devices in the entire network can calculate and obtain an effective recognition result. In more cases, there will not be an intermediate result data indicating clearly what event occurs at what position, but as the event occurs, the collaborative computing of each node device will drive the execution device that needs to make a targeted response to the event to execute the corresponding operation.

[0087] In a peer-to-peer network, for a certain point sample of a certain object to be recognized, among the result data transmitted from the node device that collects the point sample to other node devices, it can enable subsequent node devices to adjust their perceptual attention according to the characteristics of the point sample (it is not necessarily required that the characteristics of the point sample are included in the result data, but the characteristics of the point sample participate in the calculation of the previous node device, so that when the result data of the previous node device is used as the input of the data processing model of the subsequent node device, the data processing model of the subsequent node device can achieve the effect of adjusting perceptual attention in the calculation); or, report the characteristics of the point sample for subsequent node devices to adjust their perceptual attention (directly state the characteristics of the point sample in the result data). If other subsequent node devices do not detect the characteristics of the point sample, but it can be determined from the characteristics of other point samples that the undetected characteristics of the point sample still belong to the object to be recognized, then continue to state the undetected characteristics of the point sample in the result data of the current node device and transmit it to other node devices. For example, the previous node device senses the color on the object to be recognized A. When the current node device does not sense the color on the object to be recognized A, but it can be determined from the perceptual data of other node devices that in addition to other objects to be recognized, there is also the object to be recognized A, then the unperceived color on the object to be recognized A is still stated in the result data of the current node device.

[0088] In this embodiment, the method of reporting the characteristics of the point sample for subsequent node devices to adjust their perceptual attention is as follows: adjust the parameters of the data processing model of the subsequent node device for the characteristics of the point sample provided by the previous node device, so that the subsequent node device can improve its computing power for identifying the characteristics of the point sample; or, the subsequent node device uses the perceptual attention model to match the received characteristics of the point sample for computing power adjustment.

[0089] Among them, the above-mentioned "characteristics" has a different meaning from "feature recognition" in the prior art. "Feature recognition" in the prior art usually refers to information that can determine the identity of the target, while the "characteristics" of the present invention represents a kind of perceived perceptual data belonging to the object to be recognized, such as coordinates, color on the object to be recognized, etc. The "non-specific feature recognition" of the object to be recognized cannot be directly completed only through the "characteristics" of single-point perception.

[0090] In this embodiment, the method for reporting the features of the point sample for subsequent node devices to adjust the sensing attention is as follows: for the result data expressing the features of the point sample provided by the previous node device (in the present invention, usually not the features of the point sample itself, but the features of the point sample are expressed in the result data), or the features of the point sample (i.e., the features of the point sample itself), adjust the parameters of the data processing model of the subsequent node device, so that the subsequent node device improves the computing power for identifying the features of the point sample; or, the subsequent node device uses the sensing attention model to match the received features of the point sample or the result data expressing the features of the point sample for computing power adjustment.

[0091] When the node device processes the result data output by several previous node devices, based on the data processing model, when the objects to be recognized described by several previous node devices can be determined as the same target through some common point sample features, the point sample features and other information described by each node device are merged into the same target. For example, the point sample features in the physical space that almost completely overlap at the same time can be determined to be the same target.

[0092] When the result data received by the node device indicates that before the current node device received the result data this time, the flag used to identify the object to be recognized is different from the flag used by other node devices to identify the object to be recognized, and the flag assigned by other node devices for the object to be recognized is updated, then the flag used by the current node device to identify the object to be recognized before the current node device received the result data this time is converted. Specifically, the method for converting the flag used by the current node device to identify the object to be recognized before the current node device received the result data this time is as follows:

[0093] Replace the flag used by the current node device to identify the object to be recognized before the current node device received the result data this time with the latest flag assigned by other node devices for the object to be recognized; this is a relatively simple implementation manner provided by the present invention.

[0094] Or, record the conversion relationship between the flag used by the current node device to identify the object to be recognized before the current node device received the result data this time and the updated flag assigned by other node devices for the object to be recognized, and perform the conversion when it is necessary to refer to the result data received by the current node device this time; this is a relatively complex implementation manner provided by the present invention.

[0095] Or, the node device deploys a conversion model, and performs corresponding conversions on the flags of multiple objects to be recognized according to the input original data or result data; this is a more complex implementation manner provided by the present invention.

[0096] In the present invention, in order to improve the effectiveness of "non-specific feature recognition", for one or more point samples successively collected by node devices at different collection positions, if the feature values of one or more point samples at different collection positions respectively meet the preset similar conditions or are determined by a specific model to have a correlation reaching a threshold, and are unique at each collection position, then it is determined that the point samples of this type at different collection positions have a correlation.

[0097] On the other hand, for one or more point samples simultaneously collected by node devices at different collection positions, if the node devices at different collection positions collect the same spatial field, when there is only a single object to be recognized in the spatial field, or the collected point samples can correctly point to one of the multiple objects to be recognized, then for a certain object to be recognized, one or more point samples collected by the node devices at different collection positions have a correlation.

[0098] In the present invention, the data collection device of the node device includes one or a combination of an image acquisition device, an electromagnetic induction device, a temperature measurement device, a vibration frequency sensing device, and a lidar. The data collected by the above devices (that is, one or a combination of an image acquisition device, an electromagnetic induction device, a temperature measurement device, and a vibration frequency sensing device) and the three-dimensional point cloud collected by the lidar or the point cloud generated according to the images collected by multiple image acquisition devices are jointly calculated to obtain a three-dimensional point with data; the image color, contour, line, reflectivity, motion trend, electromagnetic characteristics, temperature, temperature change trend, vibration frequency, and vibration frequency change trend based on two-dimensional perception are used as additional attributes of the corresponding three-dimensional points to form a three-dimensional point cloud with attributes; in combination with the change characteristics of electromagnetic induction, temperature law, vibration frequency, motion correlation (different motion correlations presented by different materials such as ropes and fabrics), and reflectivity, the corresponding relationship between each region of the three-dimensional point cloud with attributes and each part or associated part of the 3D appearance of the object to be recognized is determined. This embodiment uses the attributes of the three-dimensional point cloud with attributes and their correlations to judge the relationships between points, and the corresponding relationships between each region to which the relevant points belong and each part or associated part of the 3D appearance of the object to be recognized, and can more accurately judge the point sample characteristics belonging to the object to be recognized, improving the efficiency and accuracy of "non-specific feature recognition".

[0099] In the process of "non-specific feature recognition", the present invention can also obtain the identity information of the object to be recognized when necessary. Specifically, when it is determined that the identity information of the object to be recognized needs to be obtained, an identity information acquisition command is triggered, and the identity information acquisition command is used as one of the inputs to participate in the calculation of the result data of the node device. By driving the node devices in the peer-to-peer network that are connected with barrier-free data acquisition conditions capable of obtaining the identity information of the object to be recognized to respond to the corresponding result data, the identity information of the object to be recognized is obtained. Similarly to the corresponding response execution of the execution device, the acquisition of identity information is the result of collaborative calculation, that is, the determination to obtain identity information triggers the acquisition of identity information, rather than being triggered additionally through a specific request command. Based on the present invention, if the permission calculation is triggered by a request command, in most cases, the permission calculation can be completed without obtaining the identity information. Only when it is found in a few cases that the permission calculation cannot be completed without obtaining the identity information, a determination to obtain the identity information is generated according to the implementation requirements. For example, through collaborative calculation, it is known that the identity information of a certain person exists in the code-scanning registration systems at several locations, the express pickup registration systems at several locations, or the consumption registration systems at several locations, and the person has given prior authorization or the query permission has been obtained in accordance with the law, then the peer-to-peer network can drive the node devices connected to these systems in a barrier-free data acquisition manner, and the relevant information obtained is sent to the peer-to-peer network through each node device to obtain information comparison and provide accurate identity information. Based on this, the present invention can also minimize the possibility of tampering with a certain system to counterfeit an identity.

[0100] In the present invention, for calculating whether a requester has the permission corresponding to the requirement, it is not necessary to first determine its specific identity information; because in most cases in a peer-to-peer network environment, it may not be necessary to obtain the identity information to verify whether the true identity of any requester has this right. The peer-to-peer network verifies the relationship between the true identity of the requester and the right. Among them, the true identity does not refer to its identity data, but refers to the physical entity of the requester in the area covered by the peer-to-peer network. The observation results of multiple authorized operations performed by this physical entity often continue to the node device for this verification in the form of calculation result data between node devices; and then determine whether its true identity has the execution permission for the current requirement, rather than declaring its identity. In most cases, this process does not require obtaining its true identity information. Since the peer-to-peer network can directly perform collaborative operations on actual behaviors, trajectories, and data, that is, it does not need to obtain the specific identity information of the requester, but only needs to verify through the true identity and other rights to determine the relationship between its true identity and the execution permission of the current requirement, including being authorized or not. Other node devices perform subsequent calculations on the relationship between the true identity obtained and the execution permission of the current requirement, and express the relationship between the true identity of the requester and the execution permission of the current requirement in the result data, and transmit it to the subsequent node devices of the next layer, without outputting the specific identity information of the requester to other node devices.

[0101] Specifically, the peer-to-peer network verifies the authenticity of the identity information of the object to be identified, and then determines its permission; among them, the node devices in the peer-to-peer network that can obtain identity information may not provide identity information (according to implementation requirements, identity information can also be provided), and only express the verification result in the result data of this node device according to the verification requirement of the authenticity of the identity information in the received result data. That is, in the present invention, in the case where the node devices that can obtain identity information do not provide identity information, the verification result is only expressed in the result data of this node device according to the verification requirement of the authenticity of the identity information in the received result data.

[0102] When the node devices in the peer-to-peer network that can obtain identity information do not provide identity information, it drives the information source device that provides identity information to establish an encrypted file transmission channel or an encrypted information transmission channel in other network communication modes between the input terminal of the node device that needs to obtain identity information; and uses the identity information as one of the inputs of the node device.

[0103] When necessary, in order to meet the requirements of other traditional computing modes for raw data, such as the evidence preservation required for traditional proof, in this embodiment, the node device can be additionally provided with a data storage device for storing the raw data sensed by the sensor.

[0104] In specific implementation, the node device may also be provided with functions such as leakage protection in its power supply device. The node device may also provide various communication interfaces, including optical fiber interfaces, wireless communication interfaces, etc.; it may also provide data interfaces for external storage devices. The node device may be powered by solar energy or mains power. When the node device is implemented outdoors, it can be installed on poles such as street lamps (without a cross arm, mounted on the main pole, or integrated into the lampshade); in a pole-free area, such as indoors, it can be wall-mounted or integrated into the ceiling.

[0105] When the present invention is implemented indoors and outdoors, the node device, as an artificial intelligence facility installed in public spaces, can be used as the digital economic infrastructure of the urban agglomeration to provide 24-hour uninterrupted seamless coverage. Through collaborative computing across node devices, the vehicle identity recognition accuracy can reach nearly 100% at any location within the covered area, and the location recognition accuracy is related to the sensor accuracy.

[0106] In the architecture of the peer-to-peer computing network of the present invention, all node devices are of the same type and have the same function. Each node device adjusts its own data processing model in real time and dynamically according to the consensus mechanism of the entire network. The raw data collected by the data collection devices (including sensors, cameras, etc.) connected to each node device is processed and encrypted by the node device according to its own data processing model to generate a byte-level processing and encryption result (i.e., the result data), and this result data will be sent to other node devices (the calculation and encryption results output by other node devices received by the current node device at the same time also belong to one of the raw data collected by the current node device). Therefore, the effects generated by the raw data sensed by each sensor will spread among a large number of peer node devices in a power-law manner. If each node device sends its result data to 100 surrounding node devices, after four unit time periods, hundreds of millions of node devices will be affected by the event sensed by this sensor. In this computing mode, the information is relatively symmetric, immune to tampering and forgery, fundamentally solving the fundamental hidden dangers of traditional information technology, that is, false information, forged information, and incorrect information caused by information asymmetry, which then become the entry points for fraud and cyberattacks, as well as problems such as long complex comprehensive application cycles, poor accuracy, and poor adaptability. Thus, it truly becomes the information infrastructure for urban comprehensive management and the digital economic infrastructure of the city.

[0107] The above embodiments are only used to illustrate the present invention and are not intended to limit the present invention. As long as changes, modifications, etc. are made based on the technical essence of the present invention, they will fall within the scope of the claims of the present invention.

Claims

1. A peer-to-peer network for collaborative computing, characterized in that, It includes multiple node devices, and there is no primary or secondary relationship among all node devices; for a certain node device, it processes the collected original data to obtain result data, and propagates the result data to other node devices; other node devices that receive the result data use the result data as one of the collected original data, and the result data affects the result data of other node devices; based on this, multiple node devices in the peer-to-peer network perform collaborative computing to achieve the discovery of events and / or respond to corresponding execution devices; Some or all node devices are provided with at least one type of sensor for collecting different corresponding types of perception data; The current node device receives the result data output by other node devices; For the current node device, it combines the collected perception data with the result data from other node devices, calculates the result data of the current node device, and sends it to other node devices; The node devices in the peer-to-peer network perform collaborative computing along with the collection of perception data and the calculation of result data.

2. The peer-to-peer network for collaborative computing according to claim 1, wherein When the result data of the collaborative computing can determine a certain event, the discovery of the event is completed.

3. The peer-to-peer network for collaborative computing according to claim 1, characterized in that, For a certain node device, it selectively sends its result data to specific other node devices.

4. The collaborative computing peer-to-peer network according to claim 1, wherein The node device deploys a data processing model, and the data processing model is used to calculate the result data; The data processing model adjusts the model parameters according to the result data received each time.

5. The peer-to-peer network for collaborative computing according to claim 4, wherein If the node device deploys an identification model, when the node device perceives an event, it uses the identification model to perform preliminary processing on the original data to obtain identification data corresponding to the event; the identification data is used as the input of the data processing model.

6. The collaborative computing peer-to-peer network according to claim 1 or 4, wherein The data processing model loaded on the node device performs adaptive perception calculation on the data processing models of the previous node devices. Specifically, when a certain node device calculates and processes the result data from other node devices, the adaptive perception calculation includes performing verification calculation on the result data of the current node device and other related node devices and / or verification calculation on the historical result data of the current node device to obtain an adaptive value.

7. The collaborative computing peer-to-peer network according to claim 6, wherein When the data processing model of a certain node device cannot adapt to the sensors and execution devices it is connected to, the influence of the adaptive value of its adaptive perception calculation on the current result data of the current node device is: in the current result data of the current node device, a preset flag bit is assigned; Or, other related node devices perform comprehensive calculation on the current result data of the current node device according to their own data processing models. When the obtained result data enables subsequent node devices to use the result data as input, after the result data is calculated by other node devices, it drives the adjustment of the data processing model of the current node device.

8. The peer-to-peer network for collaborative computing according to claim 6, wherein In the calculation of the current result data or several data transmissions and result data calculations, the data processing model of the node device covers the following modes: After it is found that the adaptability value of the data processing model of a certain node device exceeds the degree threshold, the result data output by the node device will add the node device connecting to the model training facility to the receiving node list. When the node device connecting to the model training facility receives such data, the drive command for driving the model training facility will be included in the result data calculated in that instance. The model training facility is regarded as a type of sensor, and the signal it emits is added to the connected node device. After the result data of the node device is calculated and emitted, the corresponding node device in the peer-to-peer network will drive the data storage module of the connected sensor to send the sensed data specified by the model training facility to a specific location through establishing a file transfer channel between the node devices in the peer-to-peer network or in other network communication modes to participate in model training and improvement. The improved data processing model will be deployed to the node devices that the original data processing model cannot adapt to in the same way.

9. The collaborative computing peer-to-peer network according to claim 8, wherein, Alternatively, the model training facility drives multiple node devices to generate a new data processing model in a federated computing manner using their own sensed data through federated computing and updates the data processing model.

10. The collaborative computing peer-to-peer network according to claim 8 or 9, characterized in that, The degree threshold is a set value or a dynamic value; if it is a dynamic value, it is determined by a preset program or by a dynamically updated data processing model.

11. The peer-to-peer network for collaborative computing according to claim 1, wherein The principle for data transfer between node devices is: when the result data of the current node device changes, or the change amount of the result data reaches a preset threshold, the latest result data will be sent to other node devices; Alternatively, the current node device continuously sends the result data of each instance to other node devices at a certain frequency.

12. The peer-to-peer network for collaborative computing according to claim 1, wherein All node devices encrypt the result data calculated by them based on an encryption consensus mechanism to obtain an encrypted result, and then send the encrypted result to other node devices.

13. The collaborative computing peer-to-peer network according to claim 12, wherein The encryption consensus mechanism includes one or more consensus mechanisms, and different consensus mechanisms correspond to changing the encryption algorithm structure and parameters of the node device.

14. The peer-to-peer network for collaborative computing according to claim 1, wherein The discovery of events by the peer-to-peer network includes the content of the event, the location where the event occurs, and the corresponding response handling.

15. The peer-to-peer network for collaborative computing according to claim 1, wherein Node devices communicate with each other using data packets of standard size.

16. The collaborative computing peer-to-peer network according to claim 15, wherein The peer-to-peer network deploys a QoS mechanism, and the QoS mechanism gives priority to ensuring the transmission quality of the result data between node devices.

17. The collaborative computing peer-to-peer network according to claim 1, characterized in that, The networking mode of the peer-to-peer network is one or a combination of several of the 4G mode, 5G mode, or MESH mode.

18. The collaborative computing peer-to-peer network according to claim 17, wherein The described MESH mode is based on the LTE standard and communicates at the physical layer of the LTE standard; it carries data with a customized frame structure and uses a dedicated wireless communication protocol for interaction.

19. A peer-to-peer computing network for non-specific feature recognition, characterized in that, Based on the peer-to-peer network according to any one of claims 1 to 18, the node device is provided with a data acquisition device and an operation module. The data acquisition device includes at least one type of sensor; the node devices arranged at different acquisition positions acquire at least one point sample of the object to be recognized, and the point sample is the sensed data corresponding to the sensor type. In the case where it is not necessary to obtain the identity information of the object to be recognized, the node devices perform collaborative calculation to determine that each unique object to be recognized is itself, realizing non-specific feature recognition.

20. The peer-to-peer computing network for non-specific feature recognition according to claim 19, wherein In a peer-to-peer network, for a certain point sample of a certain object to be recognized, in the result data transmitted from the node device that collects the point sample to other node devices, subsequent node devices adjust the sensing attention according to the characteristics of the point sample, or report the characteristics of the point sample for subsequent node devices to adjust the sensing attention; if subsequent other node devices do not detect the characteristics of the point sample, but it can be determined from the characteristics of other point samples that the characteristics of the undetected point sample still belong to the object to be recognized, then continue to express the characteristics of the undetected point sample in the result data of the current node device and transmit it to other node devices.

21. The peer-to-peer computing network for non-specific feature recognition according to claim 20, characterized in that, The method of reporting the characteristics of the point sample for subsequent node devices to adjust the sensing attention is as follows: for the result data expressing the characteristics of the point sample provided by the previous node device, or the characteristics of the point sample adjust the parameters of the data processing model of the subsequent node device, so that the subsequent node device improves the computing power for identifying the characteristics of the point sample; or, the subsequent node device uses the sensing attention model to match the characteristics of the received point sample or the result data expressing the characteristics of the point sample for computing power adjustment.

22. The peer-to-peer computing network for non-specific feature recognition according to claim 20, wherein When a node device processes the result data output by several previous node devices, based on the data processing model, when the objects to be recognized described by several previous node devices can be determined as the same target through certain common point sample characteristics, the point sample characteristics and other information described by each node device are merged into the same target.

23. The peer-to-peer computing network for non-specific feature recognition according to claim 22, wherein When the result data received by the node device indicates that before the current node device receives the result data this time, the flag used to identify the object to be recognized is different from the flag used by other node devices to identify the object to be recognized, and the flag assigned by other node devices to the object to be recognized is updated, then the flag used by the current node device to identify the object to be recognized before receiving the result data this time is converted.

24. The peer-to-peer computing network for non-specific feature recognition according to claim 22, characterized in that, The method of converting the flag used by the current node device to identify the object to be recognized before receiving the result data this time is as follows: Replace the flag used by the current node device to identify the object to be recognized before receiving the result data this time with the latest flag assigned by other node devices to the object to be recognized; Or, record the conversion relationship between the flag used by the current node device to identify the object to be recognized before receiving the result data this time and the updated flag assigned by other node devices to the object to be recognized, and perform the conversion when the result data received by the current node device this time needs to be referenced; Or, the node device deploys a conversion model, and performs corresponding conversions on the flags of multiple objects to be recognized according to the input original data or result data.

25. The peer-to-peer computing network for non-specific feature recognition according to claim 20, wherein For one or more point samples collected successively by node devices at different collection positions, if the characteristic values of one or more point samples at different collection positions respectively meet the preset similar conditions or are determined by a specific model to have a correlation reaching the threshold, and are unique at each collection position, then it is determined that the point samples of this type at different collection positions have relevance.

26. The peer-to-peer computing network for non-specific feature recognition according to claim 20, wherein For one or more point samples simultaneously collected by node devices at different collection positions, if the node devices at different collection positions collect data from the same spatial field, when there is only a single object to be recognized in the spatial field, or the collected point samples can correctly point to one of the multiple objects to be recognized, then for a certain object to be recognized, the one or more point samples collected by the node devices at different collection positions are relevant.

27. The peer-to-peer computing network for non-specific feature recognition according to claim 26, wherein The data acquisition device of the node device includes one or a combination of an image acquisition device, an electromagnetic induction device, a temperature measurement device, a vibration frequency sensing device, and a lidar. The data collected by the above devices and the three-dimensional point cloud collected by the lidar or the point cloud generated from the images collected by multiple image acquisition devices are jointly calculated to obtain three-dimensional points with data. The image color, contour, lines, reflectivity, motion trend, electromagnetic characteristics, temperature, temperature change trend, vibration frequency, and vibration frequency change trend based on two-dimensional sensing are used as additional attributes of the corresponding three-dimensional points to form a three-dimensional point cloud with attributes. The corresponding relationship between each region of the three-dimensional point cloud with attributes and each part or associated part of the 3D appearance of the object to be recognized is determined by combining the change characteristics of electromagnetic induction, temperature law, vibration frequency, motion correlation, and reflectivity.

28. The peer-to-peer computing network for non-specific feature recognition according to claim 19, wherein When it is determined that the identity information of the object to be recognized needs to be obtained, an identity information acquisition command is triggered. The identity information acquisition command is used as one of the inputs to participate in the calculation of the result data of the node device. By driving the node devices in the peer-to-peer network that are connected to the barrier-free data acquisition conditions capable of obtaining the identity information of the object to be recognized to respond to the corresponding result data, the identity information of the object to be recognized is obtained.

29. The peer-to-peer computing network for non-specific feature recognition according to claim 28, wherein The peer-to-peer network verifies the authenticity of the identity information of the object to be recognized, and then determines its permissions. Among them, the node devices in the peer-to-peer network that can obtain the identity information do not provide the identity information, but only express the verification result in the result data of the node device according to the verification requirement of the authenticity of the identity information in the received result data.

30. The peer-to-peer computing network for non-specific feature recognition according to claim 29, characterized in that The peer-to-peer network verifies the relationship between the true identity and ownership of the requester of the demand to determine whether its true identity has the execution permission for the current demand. Specifically, the peer-to-peer network directly performs collaborative operations on the actual behavior, trajectory, and data to determine the relationship between its true identity and the execution permission of the current demand. Other node devices perform subsequent calculations on the relationship between the true identity obtained by them and the execution permission of the current demand, and express the relationship between the true identity of the requester of the demand and the execution permission of the current demand in the result data and transmit it to the subsequent node devices at the next layer. Among them, it is not necessary to output the specific identity information of the requester of the demand to other node devices.

31. The peer-to-peer computing network for non-specific feature recognition according to claim 29, wherein The node devices in the peer-to-peer network that can obtain the identity information do not provide the identity information, but drive the information source device that provides the identity information to establish an encrypted file transmission channel or an encrypted information transmission channel in other network communication modes between the input terminal of the node device that needs to obtain the identity information; use the identity information as one of the inputs of the node device.

32. The peer-to-peer computing network for non-specific feature recognition according to any one of claims 19 to 31, characterized in that, The data acquisition device includes one or more of an image acquisition device, an audio acquisition device, a temperature detection device, a vibration sensor, a lidar, a chemical sensor, and an electromagnetic induction device.