AI glasses data low-power-consumption storage management method and system

Abstract

The invention relates to the technical field of intelligent wearable equipment, and discloses an AI glasses data low-power-consumption storage management method and system, and the method comprises the steps: obtaining a multi-source signal data unit which is collected by AI glasses and is provided with a block identification code; performing transmission processing to obtain cloud data; performing data recombination according to the cloud data to obtain an intermediate data stream; performing integrity verification according to the intermediate data stream to obtain a complete environment data stream; performing multi-level distribution of target data according to the complete environment data stream to obtain a multi-level distribution state, and performing distribution verification on the multi-level distribution state to obtain initial response distribution; performing consistency verification according to the initial response distribution to obtain a consistency response set; and carrying out storage configuration optimization according to the consistency response set to obtain a final data storage scheme. According to the method, efficient collaboration among multi-layer architectures can be realized, and the reliability of data transmission and the storage management efficiency are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of intelligent wearable devices, and in particular to an AI glasses data low-power storage management method and system. BACKGROUND

[0002] At present, data storage and management in the field of intelligent wearable devices is an important link for promoting technological progress and improving user experience, especially the data collected by intelligent glasses, which includes various forms such as images and sounds, which puts forward very high requirements for data storage efficiency and power consumption control.

[0003] In one prior art, when dealing with intelligent glasses data storage and management, the multi-modal data stream collected by AI glasses is directly transmitted and stored without discrimination, and there is a lack of dynamic judgment of data value, resulting in a large amount of useless information occupying storage space; meanwhile, in the data flow conversion process of the multi-level architecture of edge devices, near-field devices and cloud, the accuracy of data filtering is not high, the priority of data processing is difficult to unify, and the coordination is insufficient, and there is a general phenomenon of waste of computing resources.

[0004] In summary, the prior art has the problem of insufficient coordination and data processing efficiency of AI glasses multi-source heterogeneous data processing, resulting in power waste. SUMMARY

[0005] The application provides an AI glasses data low-power storage management method and system to realize efficient coordination between multi-level architectures and improve the reliability of data transmission and the efficiency of storage management.

[0006] In a first aspect, to solve the above technical problems, the application provides an AI glasses data low-power storage management method, comprising: obtaining multi-source environment signal data collected by edge devices on AI glasses, and performing data blocking to obtain multi-source signal data units; performing transmission processing on the multi-source signal data units to obtain a blocking sending sequence, and performing data transmission according to the blocking sending sequence to obtain cloud data; recombining the cloud data to obtain a recombined intermediate data stream; performing integrity verification on the intermediate data stream to obtain a complete environment data stream; performing multi-level distribution of target data according to the complete environment data stream to obtain a multi-level distribution state, and performing distribution verification on the multi-level distribution state to obtain an initial response distribution; performing consistency verification according to the initial response distribution to obtain a consistent response set; performing storage configuration optimization according to the consistent response set to obtain a final data storage scheme.

[0007] In an alternative embodiment, the multi-source environmental signal data collected by the edge device on the AI glasses is acquired, and data blocking is performed to obtain a multi-source signal data unit, comprising: Acquiring multi-source environmental signal data collected by the edge device on the AI glasses; Performing data blocking on the multi-source environmental signal data to obtain a data blocking unit, and adding blocking identification coding to the data blocking unit to obtain the multi-source signal data unit with blocking identification coding.

[0008] In an alternative embodiment, the transmission processing of the multi-source signal data unit is performed to obtain a blocking sending sequence, comprising: Acquiring a transmission channel capacity state; Performing preliminary grouping on the multi-source signal data unit in combination with the transmission channel capacity state to obtain a preliminary grouping sequence; Performing timing marking on the preliminary grouping sequence to obtain a data set with timing identification; Performing priority sorting on the data set to obtain the sorted blocking sending sequence.

[0009] In an alternative embodiment, the data reorganization according to the cloud data is performed to obtain a reorganized intermediate data stream, comprising: Extracting the blocking identification coding in the cloud data, and performing preliminary sorting in combination with the blocking identification coding to obtain a preliminary arrangement sequence; If the receiving time interval of the preliminary arrangement sequence exceeds a preset window threshold, temporarily buffering the data of the preliminary arrangement sequence to obtain a to-be-processed data set; Performing timestamp priority selection on the to-be-processed data set to obtain a candidate data group meeting the conditions; Performing data combination operation according to the candidate data group to obtain the intermediate data stream.

[0010] In an alternative embodiment, the integrity check according to the intermediate data stream is performed to obtain a complete environmental data stream, comprising: Extracting the field identification of the intermediate data stream, and performing integrity check on the field identification, if the integrity of the field identification is insufficient, performing loss detection on the intermediate data stream to determine the missing data; Performing data retransmission according to the missing data to obtain supplementary data; Performing sequence calibration in combination with the supplementary data and the intermediate data stream to obtain a data group meeting the conditions; Performing batch integration processing on the data group to obtain the complete environmental data stream.

[0011] In an optional implementation, the multi-level distribution of the target data according to the complete environment data stream obtains a multi-level distribution state, and the method comprises: Based on the complete environment data stream, a target response sequence extraction and classification is performed to obtain a response fragment set; The response fragment set is hierarchically divided to determine the content distribution range of each node; According to the content distribution range, data pushing is performed to obtain the multi-level distribution state.

[0012] In an optional implementation, the consistency check according to the initial response distribution obtains a consistent response set, and the method comprises: The response content data of the initial response distribution is extracted, and a preliminary difference comparison is performed on the response content data to obtain a preliminary conflict node list; Obtain each node priority information, and combine the each node priority information and the preliminary conflict node list to perform priority sorting to obtain a sorted priority sequence; According to the priority sequence, the response content data is updated to obtain a data update state, and the data update state is periodically compared to obtain the final consistent response set.

[0013] In an optional implementation, the storage configuration optimization according to the consistent response set obtains a final data storage scheme, and the method comprises: Obtain the storage state and the scene adaptation state in the consistent response set, and perform storage structure division according to the storage state and the scene adaptation state to obtain a storage configuration scheme; According to the storage configuration scheme, data synchronization verification is performed to determine the final data storage scheme.

[0014] In a second aspect, the application provides an AI glasses data low-power storage management system, comprising: A multi-source data acquisition module is configured to acquire multi-source environment signal data collected by edge devices on AI glasses, and perform data blocking to obtain multi-source signal data units; A cloud data transmission module is configured to perform transmission processing on the multi-source signal data units to obtain a block sending sequence, and perform data transmission according to the block sending sequence to obtain cloud data; An intermediate data reorganization module is configured to perform data reorganization according to the cloud data to obtain a reorganized intermediate data stream; A data integrity verification module is configured to perform integrity verification according to the intermediate data stream to obtain a complete environment data stream; A multi-layer distribution response module is configured to perform multi-layer distribution of target data according to the complete environment data stream, obtain a multi-layer distribution state, and perform distribution verification on the multi-layer distribution state to obtain an initial response distribution; A response consistency verification module is configured to perform consistency verification according to the initial response distribution to obtain a consistent response set. A final storage optimization module is configured to perform storage configuration optimization according to the consistent response set to obtain a final data storage scheme.

[0015] In a third aspect, the present application further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the AI glasses data low-power storage management method according to any one of the above.

[0016] In a fourth aspect, the present application further provides a computer readable storage medium comprising a stored computer program, wherein the computer program controls a device where the computer readable storage medium is located to execute the AI glasses data low-power storage management method according to any one of the above when the computer program is running.

[0017] Compared with the prior art, the present application has the following beneficial effects: (1) The present application preliminarily blocks and adds identification coding to the data features of the multi-source sensors of AI glasses, detects transmission restrictions, sends in batches and records timestamps, recombines and verifies in the cloud according to the identification and sequence markers, and ensures the accuracy and consistency of key data (such as visual recognition results and navigation instructions) at all levels through the end-edge-cloud multi-layer node synchronous distribution mechanism combined with consistency verification. This integrated processing for the AI glasses end-edge-cloud collaborative link significantly improves the overall efficiency and reliability from data collection to intelligent response, and guarantees the smoothness of user experience in complex mobile scenarios.

[0018] (2) The present application adds lightweight block identification coding to the video stream, audio stream, and sensor data stream generated by AI glasses and performs adaptive packet transmission, which significantly reduces the continuous load of large-scale continuous data (such as high-definition video recording) on the wireless module, reduces the waste and retransmission in the transmission process, directly reduces the key power consumption of the glasses side, prolongs the available time of single charging, and solves the short battery life problem caused by blind data stream transmission in existing solutions.

[0019] (3) In the AI glasses scene (such as real-time translation, object recognition, health monitoring), the application performs multi-level distribution and distribution verification on data, ensures that high-priority data (such as user active interaction instructions, system alarms) can be delivered to edge or cloud processing nodes with low delay and high reliability and quickly feedback, and allows non-urgent data (such as historical environment logs) to be processed asynchronously in the background. This unifies the data processing priority in the complex scene of AI glasses, avoids the mismatch and waste of computing and communication resources, and makes the system resources more focused on user immediate needs. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a kind of AI glasses data low-power storage management method flow chart provided by the first embodiment of the application; Figure 2 is a kind of AI glasses data low-power storage management system structure schematic diagram provided by the second embodiment of the application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0022] Reference Figure 1 The first embodiment of the application provides a kind of AI glasses data low-power storage management method, comprising the following steps: S11, the multi-source environmental signal data collected by the edge device on AI glasses is acquired, and data block is obtained to obtain multi-source signal data unit; S12, the multi-source signal data unit is processed, and the block sending sequence is obtained, and data transmission is carried out according to the block sending sequence, and cloud data is obtained; S13, according to the cloud data, data reorganization is carried out, and the intermediate data stream after reorganization is obtained; S14, according to the intermediate data stream, integrity check is carried out, and complete environmental data stream is obtained; S15, according to the complete environmental data stream, multi-level distribution of target data is carried out, and multi-level distribution state is obtained, and the multi-level distribution state is verified, and initial response distribution is obtained; S16, according to the initial response distribution, consistency check is carried out, and consistent response set is obtained; S17, according to the consistent response set, storage configuration optimization is carried out, and the final data storage scheme is obtained.

[0023] In step S11, the multi-source environmental signal data collected by the edge device on the AI glasses is obtained, and data blocking is performed to obtain a multi-source signal data unit, including: Obtaining multi-source environmental signal data collected by an edge device on AI glasses; Performing data blocking on the multi-source environmental signal data to obtain a data blocking unit, and adding a blocking identification code to the data blocking unit to obtain the multi-source signal data unit with the blocking identification code.

[0024] It should be noted that this step is to convert continuous, massive, and heterogeneous original environmental signals into standardized, structured, and easily processed basic units. The edge device refers to various sensors on the AI glasses, such as environmental sensors, cameras, microphones, and gyroscopes. The multi-source environmental signal data is the original data collected by the edge device without processing, including but not limited to temperature, humidity, PM2.5 concentration, and decibel value.

[0025] Specifically, the data blocking unit is a data packet of a limited size obtained by cutting the continuous multi-source environmental signal data according to a preset blocking rule, which creates conditions for stream processing and transmission. The specific preset blocking rule is based on the typical environmental data sampling rate of 1 Hz and the power consumption test results, and the continuous multi-source signal data is cut according to a time interval of every 5 seconds. The blocking identification code is the timestamp and source identification of each blocking result. The multi-source signal data unit is the data blocking unit with the blocking identification code. For example, on October 10, 2025, at 14:30:20, the data accumulation time of edge device TEMP-CM202 reaches 5 seconds, and the data in the past 5 seconds is immediately divided into a data unit, and each data unit is attached with a blocking identification code containing the source device number, collection timestamp, and sequence number 001, such as TEMP-CM202-20251010-143020-001.

[0026] In step S12, the multi-source signal data unit is processed for transmission, a blocking sending sequence is obtained, and data transmission is performed according to the blocking sending sequence to obtain cloud data.

[0027] It should be noted that this step is to safely, efficiently, and orderly transmit the multi-source signal data unit that has been preprocessed on the edge device to the cloud. The blocking sending sequence is a sequence list used to guide data sending after preliminary grouping, time sequence marking, and priority sorting; the cloud data is a data set composed of multiple multi-source signal data units that have been successfully transmitted and stored on the cloud server.

[0028] In an implementation, the transmission processing on the multi-source signal data units to obtain a block transmission sequence comprises: obtaining a transmission channel capacity state; preliminarily grouping the multi-source signal data units according to the transmission channel capacity state to obtain a preliminary grouping sequence; performing timing marking on the preliminary grouping sequence to obtain a data set with timing marking; performing priority sorting on the data set to obtain the block transmission sequence after sorting.

[0029] It is worth noting that the transmission channel capacity state is a real-time performance indicator of the network communication link between the edge device and the cloud server at the moment of initiating data transmission, such as the current bandwidth occupancy rate. For example, the network monitoring module actively probes the network backhaul channel every 100 milliseconds. At 14:30:20, the system calculates the current bandwidth occupancy rate to be only 200 KB / s, which is far below the preset capacity threshold of 500 KB / s, and immediately determines that it is in a low capacity state, triggering the preliminary data grouping process. The preset capacity threshold is set based on the typical bandwidth of IEEE 802.11 standard and the measured power consumption data, and is used as a reference value to determine the current load state of the network transmission channel. Here, 60% of the available bandwidth is taken.

[0030] Further, the preliminary grouping sequence is a set of multiple batches of data groups formed by logically dividing the queue of data units to be transmitted according to their inherent sequence identifiers under the condition of sufficient transmission channel capacity. Specifically, the system extracts the mixed signal data units collected from 14:30:15 to 14:30:20 within 5 seconds from the cache, a total of 1000 data points, including 500 PM2.5 and 500 humidity. According to the sequence number marked in order, the PM2.5 data within the continuous 5 seconds is grouped into a preliminary grouping sequence, marked as PM25-EDGE01-143020-001 group, containing 500 data points; the humidity data is grouped into HUMI-EDGE01-143020-001 group, also containing 500 data points.

[0031] In this embodiment, the data set with timing marking is a set of multiple batches of data groups with transmission timestamps attached to the preliminary grouping sequence. For example, when performing timing marking on the preliminary grouping sequence, the transmission time of each data unit in the preliminary grouping sequence is accurate to milliseconds, and the system will mark the specific transmission time for each unit, such as a data unit expected to be transmitted at 14:30:25 on October 10, 2023, 123 milliseconds, then a timing mark 2023-10-10-14-30-25-123 is formed.

[0032] Specifically, the priority sorting is to sort the data set with timing identifier according to the business logic, generate the sorted chunk sending sequence, and ensure that the most important data can be sent to the cloud in priority and quickly, trigger the response (such as alarm) in time, and improve the real-time performance and reliability of the system. The specific business logic examples are abnormal data triggering alarm (such as detecting violent shaking), user interaction instruction data, which need low delay, and are given high priority; environmental data stream (such as real-time video frame), non-urgent historical data or log data, which can tolerate certain delay, and are given low priority. Exemplarily, in the specific scenario of air quality monitoring, PM2.5 concentration data involves real-time alarm of environmental standard exceeding, and is given high priority; humidity data is only used for trend analysis and can accept certain delay, and is given low priority. When performing priority sorting, the 500 units of the PM2.5 group are placed in the front as a whole, and the humidity group is placed in the back, to generate the optimized chunk sending sequence.

[0033] It is worth noting that when transmitting data according to the chunk sending sequence, it is necessary to ensure that the transmission state of the data is normal. Exemplarily, the system starts sending according to the chunk sending sequence, first transmits 500 PM2.5 data points, and judges the sending state in real time through the TCP acknowledgement mechanism of the receiving end. If an ACK is not received for a unit, it is automatically retried. Through this processing, the whole process realizes the closed-loop control from capacity awareness to priority scheduling, and significantly improves the response ability and management precision in the high-frequency signal environment.

[0034] In step S13, the cloud data is recombined to obtain a recombined intermediate data stream, including: The chunk identifier code in the cloud data is extracted, and preliminary sorting is performed in combination with the chunk identifier code to obtain a preliminary arrangement order; If the receiving time interval of the preliminary arrangement order exceeds a preset window threshold, the data of the preliminary arrangement order is temporarily buffered to obtain a to-be-processed data set; The to-be-processed data set is subjected to timestamp priority selection to obtain a candidate data group meeting the condition; According to the candidate data group, a data combination operation is performed to obtain the intermediate data stream.

[0035] It should be noted that the purpose of this step is to restore the data blocks received by the cloud, which may be out of order and have delays, into a continuous data stream with correct timing. Among them, the block identification code is added in step S11, which contains information such as source device number, acquisition timestamp, serial number, etc. The preliminary arrangement order is obtained by sorting the data units of the cloud data according to the block identification code. Illustratively, analyze the block identification code of each data unit, extract the timing information such as serial numbers 001, 002, 003, ignore the actual arrival order of the cloud, and strictly arrange in ascending order according to the serial number timing information to restore the original order of the data at the edge of the collection, providing a correct reference for subsequent processing.

[0036] Further, the preset window threshold is a preset time tolerance parameter, which is a scale for judging whether the data unit is continuously arrived or seriously delayed. The setting is based on the average value of historical network delay plus a safety margin, for example, the average value of historical network delay is 3 seconds, and after increasing the safety margin, a 5-second threshold is set to ensure data continuity, and it can be dynamically configured according to the deployment environment of the device; The set of data to be processed is a set of data packets temporarily buffered when the packet delay is detected, that is, the interval between arrivals exceeds the time window threshold. It contains all the data packets that have arrived in order from the last delay point to the current delay point. Illustratively, if the arrival time of a certain data unit exceeds the time interval of the previous unit, for example, the previous one arrives at 14:30:20, and this one arrives at 14:30:32, the interval is 12 seconds>5 seconds threshold, then it is judged that the unit is seriously delayed. At this time, in order to avoid data out of order caused by network delay, the delayed unit will not be waited for, but the data units that have arrived in order before will be combined into a data group and processed first; The seriously delayed unit is put into temporary buffer and waits to form the next candidate data group with the data arriving later, that is, the set of data to be processed. This ensures low-delay processing, which avoids blocking the recombination and subsequent analysis of the entire data stream due to extreme delay of individual data blocks, and ensures the real-time response capability of the system as a whole.

[0037] Specifically, the candidate data group is a subset of internally continuous and complete data packets selected from the set of data to be processed and requiring final splicing. The timestamp priority selection refers to when there are multiple candidate data groups in the set of data to be processed, the system compares the acquisition timestamp and the sending timestamp of the data comprehensively, dynamically adjusts their priority in the recombination process, and obtains the candidate data group that meets the recombination condition. In this embodiment, the system can preferentially process the data group with the earliest acquisition timestamp to process old data as soon as possible; or when the network is restored, the delayed data in the temporary cache is given a higher priority because of its earlier acquisition timestamp and is inserted at the front of the current processing queue to correct the timing deviation as soon as possible, further optimize the accuracy and efficiency of recombination, and ensure that the data can not only be processed in time, but also maximize the correctness of its global timing, which is particularly important in complex scenarios with high concurrency and unstable network.

[0038] It is worth noting that the intermediate data stream is a continuous data sequence formed by splicing the candidate data groups in their correct global order. The data combination operation refers to splicing all data units in the candidate data group into a continuous intermediate data stream according to their correct preliminary arrangement order. Illustratively, the data units are read from the candidate data group in sequence, and the payloads (such as PM2.5 concentration values) are extracted according to the order encoded by the identified block identifier, and then combined with the continuously arriving data units to splice a seamless data stream, i.e., to obtain the intermediate data stream.

[0039] In step S14, integrity verification is performed on the intermediate data stream to obtain a complete environmental data stream, including: Extracting the field identifier of the intermediate data stream and performing integrity verification on the field identifier, if the integrity of the field identifier is insufficient, performing missing detection on the intermediate data stream to determine missing data; According to the missing data, data retransmission is performed to obtain supplementary data; Combining the supplementary data and the intermediate data stream to perform sequence calibration to obtain a data group that meets the conditions; Batch integration processing is performed on the data group to obtain the complete environmental data stream.

[0040] It should be noted that this step is to identify and repair the integrity loss that may occur in the transmission and recombination process of the data stream, to ensure that the data relied on by the subsequent analysis is complete. Among them, the field identifier is a metadata tag embedded in each data unit of the intermediate data stream, which is used to describe the data fields that the unit should contain, such as PM2.5 concentration value, temperature value, etc. The lack of integrity means that one or more necessary fields specified by the field identifier are missing in the actual data unit through verification. The missing data is the specific data unit or field range that is identified as lacking integrity and is missing in the data stream. Illustratively, in a specific scenario of air quality warning, by parsing the intermediate data stream, reading the field identifier of each data unit, it is found that the temperature value fields of units 5 to 7 are marked as NULL, which does not meet the requirement that all units in this scenario must contain PM2.5, temperature, and humidity three elements, triggering the missing detection process immediately, and then comparing the list with the fields contained in the actual data unit to accurately locate the missing range as "temperature data loss in 14:30:04 to 14:30:06 for a total of 3 units". By specifying the missing range, an effective and targeted retransmission request is initiated to the edge device.

[0041] Further, data retransmission is the operation process of the cloud actively initiating a request to the edge device to request it to resend specific data units after identifying the missing data. The supplementary data is the previously missing data unit reacquired from the edge device through the retransmission mechanism. Illustratively, the cloud generates a retransmission request instruction according to the determined missing data range, which contains the device ID, the time range of the missing data, the data type, and is sent to the corresponding edge device through the communication link; the edge device receives the request, retrieves the corresponding data unit from the local cache, and re-sends it to the cloud; the cloud receives these retransmitted data units to form a supplementary data set; wherein the maximum number of retries is set, such as 3 times, to avoid wasting a lot of communication and computing resources due to local errors.

[0042] In this embodiment, sequential calibration is a preprocessing step before data integration, which confirms the time sequence position of newly arrived supplementary data and existing intermediate data stream, and ensures that the overall time sequence logic after integration is correct. The data group that meets the conditions is the supplementary data that can be arranged in the intermediate data stream in the correct time sequence after sequential calibration. Illustratively, by reading the time stamps (or sequence numbers) of adjacent data units in the supplementary data and the intermediate data stream, the correct position of the supplementary data in the stream is found (for example, the time stamp of the supplementary data is T, then it is inserted between the data units with time stamps T-1 and T+1 in the stream). Check if the continuity of the data units before and after the insertion is destroyed (such as whether the time stamps are still continuous). After ensuring that there is no error, mark this section of data as a data group that meets the conditions.

[0043] Specifically, the complete environment data stream is a high-quality data stream without missing data from the starting point to the end point, with correct time sequence, and directly available for subsequent system analysis and processing. Batch integration processing is the final operation of splicing and merging a plurality of calibrated, continuous and qualified data groups into a single, coherent and complete data entity. Exemplarily, the system regards the qualified data groups as a whole, splices the end of one data group with the beginning of the next data group seamlessly according to the global time sequence, removes temporary markers that may be generated in the reorganization and calibration process, and finally outputs a pure complete environment data stream.

[0044] In step S15, multi-level distribution of target data is performed according to the complete environment data stream, a multi-level distribution state is obtained, and distribution verification is performed on the multi-level distribution state, and an initial response distribution is obtained.

[0045] It should be noted that this step is to intelligently distribute the processed data to different parts of the system, realizing the conversion from a unified data stream to a multi-level distribution for specific tasks, and preparing for subsequent collaborative processing. Among them, the multi-level distribution is a data distribution process including target response sequence extraction and classification, hierarchical division, data pushing and the like. The multi-level distribution state is a state set recorded by the system after the pushing instruction is issued, about whether each node successfully receives the data allocated to it, such as node A: received, node B: receiving, node C: no response.

[0046] In an implementation mode, the multi-level distribution of target data according to the complete environment data stream to obtain the multi-level distribution state comprises: Based on the complete environment data stream, target response sequence extraction and classification are performed to obtain a response fragment set; Hierarchical division is performed on the response fragment set to determine the content allocation range of each node; According to the content allocation range, data pushing is performed to obtain the multi-level distribution state.

[0047] It should be noted that the target response sequence is a data sub-sequence with independent business meaning identified and intercepted from the complete environment data stream based on a preset business event rule. The response fragment set is a more fine-grained data fragment set with category labels obtained by dividing and classifying the target response sequence according to data type and time dimension.

[0048] In this embodiment, the preset business event rule is used to trigger the extraction of the target response sequence. For example, in the air quality monitoring scenario, the rule can be defined as "when the PM2.5 concentration exceeds the threshold for 5 consecutive minutes, all the environmental data in this 5-minute period is defined as a target response sequence". After obtaining a target response sequence, it is parsed and fragmented to form a response fragment set. Specifically, the data in the sequence can be divided into PM2.5 concentration fragments, humidity fragments, etc. according to the data type; or a longer target response sequence (such as a pollution event lasting for one hour) is cut into multiple equal-length short-term trend fragments, such as one fragment every 10 minutes. This process aims to decompose complex business data sequences into structured data units that are easier to distribute to different levels of nodes.

[0049] In this embodiment, the system presets a multi-level processing node architecture, for example, including edge gateway nodes for real-time response, regional cloud nodes for regional data aggregation and short-term analysis, and central cloud nodes for global data storage and deep learning.

[0050] The hierarchical division process, i.e. according to the attributes of each data fragment in the response fragment set, determines the target node level to which it should be distributed according to the preset mapping rule. The fragment attributes include its data type and business priority. For example, the mapping rule can define that the fragments belonging to "core indicator data" and "high business priority" (such as PM2.5 over-limit alarm fragments) are assigned to edge gateway nodes to ensure the lowest response delay; the fragments belonging to "short-term trend data" and "medium business priority" (such as 1-hour PM2.5 trend fragments) are assigned to regional cloud nodes; and all raw data fragments are assigned to central cloud nodes for archiving. Through this division, the content distribution range of each level node is clear.

[0051] Exemplarily, the first level node needs to store the core indicator data of all monitoring points in a certain region, and the second level node needs to store the auxiliary indicator data. In addition, real-time alarm data is sent to edge gateway nodes, short-term trend data is sent to regional cloud nodes, PM2.5 over-limit alarm fragments are stored in edge gateway nodes, and 1-hour PM2.5 trend fragments are stored in regional cloud nodes. This process avoids all data flowing to a single node and reduces delay and load through hierarchical processing.

[0052] Further, the multi-level distribution state is a set of states recorded by the system after the push instruction is issued, indicating whether each node successfully receives the data assigned to it. Data pushing is a process of storing data in corresponding nodes according to the content distribution range. Specifically, different response fragments are actively pushed to corresponding nodes according to the content distribution range during data pushing. After pushing, the system waits for the confirmation message of the node and updates the receiving state of each node according to the confirmation message, and finally summarizes the multi-level distribution state, such as node A: received, node B: receiving, and node C: no response.

[0053] It is worth noting that distribution verification is an independent quality inspection link, which aims to check the multi-level distribution state according to the preset delivery strategy and confirm whether the data has been successfully delivered to all target nodes.

[0054] The delivery strategy includes a target node list and a successful receiving standard. For example, the strategy can be defined as follows: all nodes divided into "edge gateway nodes" must successfully receive 100% of the "core index data" fragments within their content distribution range; at the same time, the delay of the data receiving timestamp of any node and the sending timestamp should not exceed the preset threshold (such as 10 minutes). Compare the multi-level distribution state with the delivery strategy. If it is found that there are nodes that do not meet the successful receiving standard, such as receiving only 80% of the specified fragments or exceeding the delay threshold, it is determined that the initial response distribution of the node fails. For the failed nodes, the missing or overdue data fragments are re-sent to them through a backup channel or re-established connection, and the distribution state is updated. This re-sending process can be repeated until all nodes meet the delivery strategy or the maximum number of retries is reached. Only when the distribution verification confirms that the initial response distribution of all nodes is successful, the current distribution state is output to the subsequent consistency verification step as the initial response distribution. The preset threshold is set to 10 minutes, which is based on the average response time of the node (5 minutes) and the fault tolerance requirement. The 10-minute threshold is set to balance real-time performance and reliability.

[0055] In step S16, consistency verification is performed according to the initial response distribution to obtain a consistent response set, including: Extracting the response content data of the initial response distribution and performing preliminary difference comparison on the response content data to obtain a preliminary conflict node list; Obtaining node priority information and performing priority sorting based on the node priority information and the preliminary conflict node list to obtain a sorted priority sequence; Performing response content data update according to the priority sequence to obtain a data update state, and periodically comparing the data update state to obtain the final consistent response set.

[0056] It should be noted that the purpose of this step is to solve the data inconsistency or conflict in interpretation that may occur in different nodes in the system after data distribution, and finally form a unified and reliable response set. Among them, the node response content data is the data received by each node after data distribution. The preliminary conflict node list is a list recording which nodes have data differences. In the preliminary difference comparison, there will be many nodes in the multi-level architecture, and many of them will store the same data segment. By comparing the data segments stored by different nodes at the same time and the same monitoring point, if there is a data difference, it is determined as a preliminary conflict node. Illustratively, a node at a certain level is responsible for storing the PM2.5 concentration data of a specific monitoring point, and another node stores the PM2.5 concentration data of the same monitoring point. Check if the data of the two nodes in the same time period such as 14:00 to 15:00 is consistent. If it is found that the PM2.5 concentration data at the same time is displayed as 50 micrograms / cubic meter in a certain node, and 45 micrograms / cubic meter in another node, these two nodes are marked as a difference node pair, and all difference node pairs form a preliminary conflict node list.

[0057] Further, the node priority information refers to the value level set in advance according to the node role, reliability, data source credibility, etc. For example, the priority of the cloud fusion node is higher than that of the single edge sensor node, and the priority of the node in the core index data layer is higher than that of the node in the auxiliary data layer. The sorted priority sequence is a list obtained by sorting the importance or priority of the conflict nodes according to the node priority information, which determines the order of the nodes for subsequent response content update to solve the conflict. Illustratively, the first layer node is responsible for the core index data, and the priority is high; while the second layer node is responsible for the auxiliary data, and the priority is low. According to this, the conflict nodes are sorted, and the adjusted priority sequence is determined as updating the first layer node data first. This process ensures the priority processing of core data, especially when resources are limited, it can reasonably allocate the update order.

[0058] In the embodiment, the data update status records whether each conflict node has successfully received and adopted the updated data. The consistent response set is a completely unified data set distributed on all nodes after conflict resolution and synchronization. When updating the data, the multi-level distribution of step S15 is performed again for the conflict nodes with high priority in the sequence of the sorted priority, so as to obtain the update of the response content of the nodes. The multi-level distribution is only for the conflict nodes, and the normal nodes do not need to receive the content. Exemplarily, when updating the response content of the conflict nodes, the system pushes the latest PM2.5 data of 48 micrograms per cubic meter from 14:00 to 15:00 to the conflict nodes, and the system obtains the content receiving state of each node in real time and performs periodic comparison. If the first layer node shows that it has been received, and another node has also been received, it is judged that the data update is completed, and the data update state is obtained.

[0059] Specifically, the periodic comparison refers to re-comparing the response content data of each node according to a certain period, so as to ensure the consistency of the data, and finally obtain the consistent set. Exemplarily, the data of each node is compared every 30 minutes, and if it is found that a certain node has not been updated to the latest PM2.5 concentration data of 48 micrograms per cubic meter, the process of step S16 is repeated until the final consistent response set is formed. This process belongs to a loop mechanism, which effectively avoids data omission and can continuously maintain data integrity. The comparison period can be dynamically adjusted according to the network condition, for example, dynamically adjusted to 30-60 minutes.

[0060] In step S17, the storage configuration optimization is performed according to the consistent response set, and the final data storage scheme is obtained, including: obtaining the storage state and the scene adaptation state in the consistent response set, and performing storage structure division according to the storage state and the scene adaptation state to obtain a storage configuration scheme; performing data synchronization verification according to the storage configuration scheme to determine the final data storage scheme.

[0061] It should be noted that the purpose of this step is to dynamically generate an optimal and specific-scene-adapted data long-term storage strategy based on the verified final data result and system running state. Among them, the storage state is the real-time snapshot of the current storage resources of each node of the system, such as the remaining capacity of the node. The scene adaptation state is determined by the pre-defined scene configuration rules, which convert business requirements into storage strategy constraints. For example, in the air quality monitoring scene, the rule can be defined as "the data type of 'PM2.5 concentration' data, the access priority is 'high', and the storage redundancy level is 'double copy'"; in the real-time monitoring scene, the rule can be defined as "the data type of 'video stream' data, the write delay requirement is 'less than 50 milliseconds'". These specific rules together constitute the scene adaptation state that can be recognized and executed by the system.

[0062] The storage structure division refers to the planning and allocation of the storage logic and storage space according to the storage state and the scene adaptation state. The storage configuration scheme is the storage planning obtained after the storage structure division according to the current storage state and the scene adaptation state. Exemplarily, a node storing pollution data has a storage capacity of 500GB, but the current data volume is close to 450GB, and under the scene demand, the pollution data is the core index data, then in order to avoid insufficient space, more space is reserved for the high-frequency updated pollution data, such as an additional 300GB for the node for the storage of core index data. Further, the monitoring scene requires high write speed for video image data, so the video image data needs to be assigned a high-priority storage logic.

[0063] Further, the data storage scheme is a trusted storage solution that has been comprehensively optimized and finally verified, fully meeting the intelligent requirements of modern Internet of Things systems for data storage. Data synchronization verification refers to the consistency and integrity of the storage configuration scheme after the storage structure division is verified again. Exemplarily, according to the optimized storage configuration scheme, data migration and storage operations are performed. The operations include migrating data from the original location to the new storage area, applying new data compression or encoding formats, and creating data copies according to the strategy. After the operation is performed, a storage operation verification is performed to determine the final data storage scheme. The storage operation verification aims to verify whether the storage operation itself is successfully executed, for example, verifying whether the key data has been successfully written to the target storage medium according to its storage strategy, whether the data copy has been established according to the set number, and whether the allocation of the storage space meets the expected scheme. If the verification finds that the storage operation fails, a retry mechanism or system alarm is triggered. This verification does not involve the consistency of the data content, which has been guaranteed by step S16.

[0064] In summary, the application discloses an AI glasses data low-power storage management method, which comprises the following steps: acquiring multi-source environment signal data collected by edge devices on AI glasses, and performing data blocking to obtain multi-source signal data units; performing transmission processing on the multi-source signal data units to obtain a block sending sequence, and performing data transmission according to the block sending sequence to obtain cloud data; performing data reorganization according to the cloud data to obtain reorganized intermediate data flow; performing integrity verification according to the intermediate data flow to obtain complete environment data flow; performing multi-level distribution of target data according to the complete environment data flow to obtain a multi-level distribution state, and performing distribution verification on the multi-level distribution state to obtain an initial response distribution; performing consistency verification according to the initial response distribution to obtain a consistent response set; and performing storage configuration optimization according to the consistent response set to obtain a final data storage scheme. The application preliminarily blocks the original signal flow and adds identification codes, sends in batches after detecting transmission restrictions and records time stamps, reorganizes and verifies in the cloud according to the identification and sequence markers, distributes by using a multi-level node synchronization mechanism, and eliminates conflicts by combining consistency verification and priority adjustment. This integrated processing significantly realizes efficient cooperation between multi-level architectures, thereby improving the reliability of data transmission and the efficiency of storage management. The application provides clear identity identification for the circulation of data in multi-level architectures by the block transmission processing with additional coding identification, reduces the bandwidth consumption and power consumption of large file transmission, avoids repeated transmission caused by incomplete data, solves the problem of insufficient data circulation cooperation in the prior art, and has the technical effect of improving the data circulation cooperation of multi-level architectures. The application performs multi-level distribution, verifies the distribution state, unifies the priority of data processing, ensures that data is accurately delivered to the target level, avoids waste of computing resources caused by data misdelivery, and solves the problem of inconsistent data processing priority and insufficient cooperation in the prior art.

[0065] Reference Figure 2 The second embodiment of the application provides an AI glasses data low-power storage management system, which comprises: A multi-source data acquisition module is configured to acquire multi-source environment signal data collected by edge devices on AI glasses, and perform data blocking to obtain multi-source signal data units. A cloud data transmission module is configured to perform transmission processing on the multi-source signal data units to obtain a block sending sequence, and perform data transmission according to the block sending sequence to obtain cloud data. An intermediate data reorganization module is configured to perform data reorganization according to the cloud data to obtain reorganized intermediate data flow. A data integrity verification module is configured to perform integrity verification according to the intermediate data flow to obtain complete environment data flow. The multi-layer distribution response module is configured to perform multi-layer distribution of target data according to the complete environment data stream, obtain a multi-layer distribution state, and perform distribution verification on the multi-layer distribution state to obtain an initial response distribution. The response consistency verification module is configured to perform consistency verification according to the initial response distribution to obtain a consistent response set. The final storage optimization module is configured to perform storage configuration optimization according to the consistent response set to obtain a final data storage scheme.

[0066] It should be noted that the AI glasses data low-power storage management system provided by the embodiments of the present application is used to perform all process steps of the AI glasses data low-power storage management method provided by the above embodiments, and the working principles and beneficial effects of the two are one-to-one correspondence, so they will not be repeated here.

[0067] The embodiments of the present application also provide an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a multi-source data acquisition program. The processor implements the steps in the above various AI glasses data low-power storage management method embodiments when executing the computer program, such as Figure 1 The step S11 shown. Alternatively, the processor implements the functions of each module / unit in the above various device embodiments when executing the computer program, such as a multi-source data acquisition module.

[0068] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application. The one or more modules / units can be a series of computer program instruction segments that can complete a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0069] The electronic device can be a desktop computer, a notebook, a palm computer, and a smart tablet, etc. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above components are only examples of the electronic device and do not constitute a limitation on the electronic device, and can include more or fewer components than the above, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0070] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The processor is a control center of the electronic device, and connects various parts of the electronic device through various interfaces and lines.

[0071] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc.; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0072] The modules / units integrated in the electronic device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0073] It should be noted that the above-described device embodiments are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, the connection relationship between the modules in the device embodiment provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0074] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are only for the specific embodiments of the present application and are not used to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An AI glasses data low-power storage management method, characterized in that, The application comprises the following steps: acquiring multi-source environmental signal data collected by edge devices on AI glasses, and performing data blocking to obtain multi-source signal data units; performing transmission processing on the multi-source signal data units to obtain a block sending sequence, and performing data transmission according to the block sending sequence to obtain cloud data; performing data reorganization according to the cloud data to obtain reorganized intermediate data streams; performing integrity verification according to the intermediate data streams to obtain complete environmental data streams; performing multi-level distribution of target data according to the complete environmental data streams to obtain a multi-level distribution state, and performing distribution verification on the multi-level distribution state to obtain an initial response distribution; performing consistency verification according to the initial response distribution to obtain a consistent response set; performing storage configuration optimization according to the consistent response set to obtain a final data storage scheme. 2.The AI glasses data low-power storage management method of claim 1, wherein, The application comprises the following steps: acquiring multi-source environmental signal data collected by edge devices on AI glasses, and performing data blocking to obtain multi-source signal data units; acquiring multi-source environmental signal data collected by edge devices on AI glasses; 3.The AI glasses data low-power storage management method of claim 1, wherein, performing data blocking on the multi-source environmental signal data to obtain data block units, and adding block identification codes to the data block units to obtain the multi-source signal data units with block identification codes. The application comprises the following steps: acquiring transmission channel capacity states; performing preliminary grouping on the multi-source signal data units in combination with the transmission channel capacity states to obtain a preliminary grouping sequence; performing time sequence marking on the preliminary grouping sequence to obtain a data set with time sequence identification; 4.The AI glasses data low-power storage management method of claim 1, wherein, performing priority sorting on the data set to obtain the block sending sequence after sorting. The application comprises the following steps: extracting block identification codes in the cloud data, and performing preliminary sorting in combination with the block identification codes to obtain a preliminary arrangement order; if the receiving time interval of the preliminary arrangement order exceeds a preset window threshold, temporarily buffering data of the preliminary arrangement order to obtain a to-be-processed data set; performing time stamp priority selection on the to-be-processed data set to obtain a candidate data group meeting a condition; 5.The AI glasses data low-power storage management method of claim 1, wherein, performing data combination operations on the candidate data group to obtain the intermediate data stream. The application comprises the following steps: extracting field identification of the intermediate data stream, and performing integrity verification on the field identification; if the integrity of the field identification is insufficient, performing loss detection on the intermediate data stream to determine lost data; performing data retransmission according to the lost data to obtain supplementary data; performing sequence calibration in combination with the supplementary data and the intermediate data stream to obtain a data group meeting a condition; 6.The AI glasses data low-power storage management method of claim 1, wherein, performing batch integration processing on the data group to obtain the complete environmental data stream. The application comprises the following steps: Based on the complete environment data stream, target response sequence extraction and classification are performed to obtain a response fragment set; The response fragment set is hierarchically divided to determine the content distribution range of each node; According to the content distribution range, data pushing is performed to obtain the multi-level distribution state. 7.The AI glasses data low-power storage management method of claim 1, wherein, According to the initial response distribution, consistency verification is performed to obtain a consistent response set, including: Extract the response content data of the initial response distribution, and perform preliminary difference comparison on the response content data to obtain a preliminary conflict node list; Obtain each node priority information, and combine the each node priority information and the preliminary conflict node list to perform priority sorting to obtain a sorted priority sequence; According to the priority sequence, response content data updating is performed to obtain a data updating state, and the data updating state is periodically compared to obtain the final consistent response set. 8.The AI glasses data low-power storage management method of claim 1, wherein, According to the consistent response set, storage configuration optimization is performed to obtain a final data storage scheme, including: Obtain the storage state and scene adaptation state in the consistent response set, and perform storage structure division according to the storage state and the scene adaptation state to obtain a storage configuration scheme; According to the storage configuration scheme, data synchronization verification is performed to determine the final data storage scheme.

9. An AI glasses data low-power storage management system, characterized by, Including: A multi-source data acquisition module is configured to acquire multi-source environment signal data collected by edge devices on AI glasses and perform data blocking to obtain multi-source signal data units; A cloud data transmission module is configured to perform transmission processing on the multi-source signal data units to obtain a block sending sequence, and perform data transmission according to the block sending sequence to obtain cloud data; An intermediate data reorganization module is configured to perform data reorganization according to the cloud data to obtain reorganized intermediate data streams; A data integrity verification module is configured to perform integrity verification according to the intermediate data streams to obtain complete environment data streams; A multi-layer distribution response module is configured to perform multi-level distribution of target data according to the complete environment data streams to obtain a multi-level distribution state, and perform distribution verification on the multi-level distribution state to obtain an initial response distribution; A response consistency verification module is configured to perform consistency verification according to the initial response distribution to obtain a consistent response set; A final storage optimization module is configured to perform storage configuration optimization according to the consistent response set to obtain a final data storage scheme.

Images