A data management system for intelligent inspection robots

Abstract

This invention discloses a data management system for intelligent inspection robots. The invention relates to the field of data management technology and includes a work order generation module, a service provider dispatch engine, and a full-process tracking and performance evaluation module. The advantages of this invention are: the work order generation module enables automatic triggering and structured generation of maintenance work orders, eliminating manual judgment and shortening the fault response cycle; the work order merging and deduplication sub-unit effectively reduces redundant work orders caused by duplicate alarms, avoiding waste of maintenance resources; it ensures automation efficiency while retaining the ability for manual intervention, balancing efficiency and accuracy; and the service provider dispatch engine introduces a normalization process before weighted scoring, eliminating the impact of inconsistent dimensions on the weighted results, ensuring objective and accurate scoring results, preventing new service providers from being unreasonably excluded due to data gaps, and also avoiding scoring distortion caused by data gaps.

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically a data management system for intelligent inspection robots. Background Technology

[0002] With the deep integration of artificial intelligence, sensor technology, and automation control technology, intelligent inspection robots have gradually replaced traditional manual inspection modes and are widely used in many key fields such as industrial manufacturing, energy operation and maintenance, medical and health care, urban infrastructure, and 3C electronics. They have become the core equipment for realizing automated inspection, improving inspection efficiency and accuracy, and reducing the intensity of manual labor and operational risks. These robots can autonomously or semi-autonomously complete the inspection, analysis, and judgment tasks of target objects. Their efficient operation relies on the collaborative support of multiple key technologies. Among them, data, as the core element that runs through the entire inspection process, directly determines the inspection capability, decision-making accuracy, and application value of intelligent inspection robots. Common data management systems often rely on manual judgment to trigger maintenance work orders, resulting in long fault response cycles, low efficiency, and scattered alarm information, which easily leads to a large number of duplicate and redundant work orders and waste of maintenance resources. At the same time, automated processes are incompatible with manual intervention. Pure automation is prone to errors and omissions in work orders, while manual processing significantly reduces the efficiency of operation and maintenance. It is difficult to balance operation and maintenance efficiency and processing accuracy. Therefore, we propose a data management system for intelligent detection robots. Summary of the Invention

[0003] The purpose of this invention is to provide a data management system for intelligent inspection robots.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a data management system for intelligent inspection robots, comprising: The work order generation module is used to receive fault warning signals, safety anomaly level signals, and parts replacement records, and generate preventive maintenance work orders, emergency maintenance work orders, and regular maintenance work orders according to the type of signal received. The work order generation module has a built-in work order merging and deduplication subunit, which is used to compare multiple trigger signals generated by the same intelligent detection robot within a preset time window, merge multiple signals with the same fault type description and overlapping components into a single work order, and record the number of all merged source signals in the work order attributes. The work order generation module also has a built-in manual review bypass interface, which is used to display detailed information of the signals to be merged to the operator. The operator can confirm the automatic merging result or perform a split operation. The operation result is written to the work order queue in real time. The service provider dispatch engine is used to normalize the original scores of four dimensions—geographic location score, technical capability certification score, historical response time score, and customer comprehensive score—before executing the weighted scoring. It linearly maps the original scores of each dimension to the numerical range of zero to one. Then, it calculates the comprehensive matching score of each candidate service provider for the current work order through the weighted scoring model, and assigns the service provider with the highest comprehensive matching score among the candidate service providers to the corresponding work order. The service provider dispatch engine also includes a cold start rule unit, which is used to calculate the comprehensive matching score of a service provider by using only the normalized geographical location score and the normalized technical capability certification score with equal weight when the number of completed work orders in the service provider's history is lower than the preset work order number threshold. The service provider's dispatch engine also schedules maintenance time slots for work orders based on the interruptible window duration of the intelligent detection robot's production. The full-process tracking and performance evaluation module is used to write the parts replacement information back to the full life cycle database in real time during the work order execution process, and calculate the service provider's comprehensive performance score based on the service provider's response timeliness score, maintenance quality score and parts usage compliance score after the work order is completed, and dynamically adjust the service provider's order dispatch priority in the service provider dispatch engine. The full-process tracking and performance evaluation module has a minimum sample size judgment mechanism. The order dispatch priority will only be adjusted when the number of historical completed work orders of the service provider reaches the preset sample size threshold. If the number of historical completed work orders does not reach the preset sample size threshold, the order dispatch priority of the service provider will remain unchanged.

[0005] As a further aspect of the present invention: when the work order generation module receives a fault warning signal, it extracts the fault type, the components involved, and the estimated maintenance duration, and generates a preventive maintenance work order; Upon receiving a security anomaly level signal, the service response priority is determined based on the preset mapping relationship between the anomaly level and the service response priority, and an emergency maintenance work order with the corresponding response priority attribute is generated. Upon receiving a parts replacement record, the system reads the parts' warranty information and standard service life data, calculates the remaining estimated service life based on the parts' current cumulative usage time, and generates a periodic maintenance work order in advance when the remaining estimated service life is lower than the preset service life threshold.

[0006] As a further aspect of the present invention: the normalization processing method of the service provider dispatch engine for the same scoring dimension is as follows: Using the lower bound of the original score values ​​of all candidate service providers in the current candidate service provider pool in this scoring dimension as the lower bound of normalization and the upper bound of the original score values ​​as the upper bound of normalization, a linear mapping is performed on the original score of each candidate service provider in this dimension to obtain the normalized score of the candidate service provider in this dimension. When all candidate service providers in a certain dimension have the same original score, the normalized score of all candidate service providers in that dimension is set to one.

[0007] As a further aspect of the present invention: when the cold start rule unit determines that the number of work orders completed by the service provider in the past is lower than the preset work order number threshold, the sum of the normalized geographical location score and the normalized technical capability certification score, each accounting for half of the weight, is used as the comprehensive matching score of the service provider. Once the number of completed work orders by the service provider reaches the preset work order threshold, it will automatically switch to a standard weighted scoring model that includes four dimensions: geographic location score, technical capability certification score, historical response time score, and customer comprehensive score, to calculate the comprehensive matching score.

[0008] As a further aspect of the present invention: when calculating the comprehensive performance score of the service provider, the full-process tracking and performance evaluation module determines the response time score by the ratio of the actual response time to the time required by the priority of the work order; determines the maintenance quality score by the improvement of the deviation of the intelligent detection robot's operating parameters and the reduction of the product defect rate before and after maintenance; and determines the compliance score of the parts used by comparing the actual parts used in the execution of the work order with the compliance ratio of the certified parts catalog item by item. The overall performance of service providers is the result of the weighted sum of the scores of the above three items according to their respective weight coefficients. Each weight coefficient is greater than zero and the sum of the three is equal to one. The minimum sample size judgment mechanism dynamically updates the basic weight coefficient of the service provider in the weighted scoring model of the service provider dispatch engine based on the average comprehensive performance score of the service provider's recently completed work orders after the number of historical completed work orders of the service provider reaches the preset sample size threshold.

[0009] As a further aspect of the present invention: the system further includes a data access layer, which has a built-in protocol conversion engine that supports communication with the intelligent detection robot via MQTT, HTTP and WebSocket protocols, converts heterogeneous data uploaded by each robot into a unified standard data format, and performs timestamp calibration and coordinate system normalization on the data. The data access layer also maintains the robot's online status information, heartbeat monitoring information, and disconnection reconnection mechanism.

[0010] As a further aspect of the present invention, it also includes a data processing and governance center, comprising a streaming engine and a batch processing scheduler; The streaming engine cleans the real-time uploaded detection data, including removing duplicate data, filling in missing timestamps, and filtering noisy data, and marks the data as abnormal according to preset rules. The batch processing scheduler performs periodic in-depth governance on the stored historical data, including data quality scoring, format normalization, label completion, and data compression and archiving.

[0011] As a further aspect of the present invention, it also includes a supplier quality traceability module, which is used to query the component material labels to identify the batch and supplier code of the components involved when the detection data exceeds the preset safety standard threshold. The module calculates the corresponding direct quality cost by multiplying the number of defective parts in this abnormal event by the value of a single product, adding the downtime loss cost per unit time by multiplying the duration of this abnormality, and adding the rework cost of this abnormal event. The calculation result is then linked to the quality record of the corresponding supplier.

[0012] As a further aspect of the present invention: the supplier quality traceability module also includes a supplier dynamic rating unit, which uses the supplier code as the primary key, aggregates the abnormal trigger rate and batch weight data of all historical supply batches of the supplier, and calculates the supplier's comprehensive quality credit score with a value range of zero to one hundred by dividing the sum of the product of the difference between the abnormal trigger rate of each batch and the batch weight of that batch by the sum of the sum of the batch weights of all batches. The higher the credit score, the more stable the historical supply quality. The system also includes a procurement decision support module, which automatically generates procurement risk warning information, a list of recommended alternative suppliers, and a suggested percentage for contract quality deposit based on the supplier's comprehensive quality credit score.

[0013] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows: 1. This invention realizes the automatic triggering and structured generation of maintenance work orders through the work order generation module, eliminating the manual judgment link, shortening the fault response cycle, and the work order merging and deduplication sub-unit effectively reduces redundant work orders caused by duplicate alarms, avoiding the waste of maintenance resources. The manual review bypass interface retains the ability of manual intervention while ensuring automation efficiency, thus balancing efficiency and accuracy. 2. This invention introduces a normalization process before weighted scoring through the service provider dispatch engine, eliminating the impact of inconsistent dimensions of each scoring dimension on the weighted result, ensuring the objectivity and accuracy of the scoring result. The cold start rule unit is specifically designed to handle new service providers with insufficient historical data, and uses two directly obtainable dimensions, geographical location and technical capabilities, for scoring transition, avoiding unreasonable exclusion of new service providers due to data loss, and also avoiding scoring distortion caused by data loss. 3. This invention quantifies maintenance quality into a comprehensive performance score through a full-process tracking and performance evaluation module, forming a feedback loop with the order dispatch priority. The minimum sample size judgment mechanism ensures that the order dispatch priority is not adjusted before the sample size is sufficient, avoiding random fluctuations in performance scores under small sample size from affecting the stability of order dispatch decisions, thereby continuously optimizing the service provider structure. 4. This invention associates abnormal detection data with upstream component batches and supplier codes through a supplier quality traceability module, enabling precise attribution of quality costs. It also provides quantitative basis for procurement decisions through the supplier's comprehensive quality credit score, reducing recurring anomalies caused by component quality issues from the source. This achieves a complete business closed loop from data perception, work order management, service provider scheduling to supply chain quality control, transforming the technical data generated by the intelligent inspection robot into assets that can directly drive business management decisions. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the system flow in an embodiment of the present invention. Detailed Implementation

[0015] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0016] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0017] This invention relates to a data management system for intelligent inspection robots.

[0018] Therefore, in order to effectively solve the above problems, this application proposes a data management system for intelligent inspection robots, as shown in the accompanying drawings. Figure 1 As shown, it includes a work order generation module, a service provider dispatch engine, and a full-process tracking and performance evaluation module. It also includes a data access layer, a data processing and governance center, a supplier quality traceability module, and a procurement decision support module. Work order generation module: The work order generation module receives fault warning signals, safety anomaly level signals, and parts replacement records, and generates preventive maintenance work orders, emergency maintenance work orders, and periodic maintenance work orders according to the type of signal received. When a fault warning signal is received, the work order generation module automatically extracts the fault type, involved components, and estimated maintenance time, and generates a preventive maintenance work order. When a safety anomaly level signal is received, the service response priority is determined according to the preset mapping relationship between the anomaly level and the service response priority, and an emergency maintenance work order with the corresponding response priority attribute is generated. When a component replacement record is received, the component warranty information and standard service life data are read, and the remaining estimated service life is calculated based on the current cumulative usage time of the component. When the remaining estimated service life is lower than the preset service life threshold, a regular maintenance work order is generated in advance. The work order generation module has a built-in work order merging and deduplication subunit. Within a preset time window, if the same intelligent detection robot generates multiple trigger signals, the work order merging and deduplication subunit compares these signals. When the fault type descriptions of multiple signals are the same and the parts lists involved overlap, they are merged into a single work order, and the numbers of all merged source signals are recorded in the work order attributes to avoid wasting maintenance resources due to the generation of redundant work orders caused by duplicate alarms. For signals that do not meet the merging conditions, work orders are generated independently for each signal. The work order generation module also has a built-in manual review bypass interface. After the automatic merging operation is completed, the manual review bypass interface pushes a notification to the operator on duty, showing the signal details involved in the merging and the content of the merged work order. The operator can confirm the automatic merging result or perform a splitting operation (that is, split the merged work order into multiple independent work orders). The manual operation results are written to the work order queue in real time and take effect, ensuring that the ability to intervene manually is retained while the process is automated. Service Provider Order Dispatch Engine: The service provider dispatch engine matches each work order with a service provider using a weighted scoring model and schedules maintenance time slots for the work order based on the interruptible window duration of the intelligent inspection robot's production. Before applying the weighted scoring, the service provider dispatch engine normalizes the raw scores across four dimensions: geographic location score, technical capability certification score, historical response time score, and overall customer score. The normalization process uses the following formula: ; in, For the first The candidate service providers in the first The normalized scores for each rating dimension range from zero to one. For the first The candidate service providers in the first The original scores for each rating dimension, For all candidate service providers in the current candidate service provider pool, in the [number]th [year]... The lower bound of the original score for each rating dimension. For all candidate service providers in the current candidate service provider pool, in the [number]th [year]... The upper limit of the original score for each rating dimension. These are the scoring dimension numbers, corresponding in order to the geographical location scoring dimension, technical capability certification scoring dimension, historical response time scoring dimension, and overall customer scoring dimension. For candidate service provider serial numbers, when At that time, the first Normalized score of all candidate service providers in the dimension All values ​​are set to one. The above normalization process ensures that the dimensions of each scoring dimension are consistent and eliminates the impact of differences in the original scoring dimensions of different dimensions on the weighted results. After normalization, the number of historical completed work orders reaches the preset work order number threshold. For candidate service providers, a comprehensive matching score is calculated using a standard weighted scoring model: ; in, For the first The overall matching score of each candidate service provider for the current work order. For the first The normalized value of the candidate service provider's geographic location score is determined based on the distance between the service provider's location and the location of the equipment to be maintained; the closer the distance, the higher the original score. For the first The normalized value of the technical capability certification score for each candidate service provider is determined based on the level of technical capability certification label held by the service provider. For the first The normalized value of the candidate service provider's historical response time rating is determined based on the statistical mean of the difference between the actual arrival time and the agreed arrival time in the service provider's historical work orders. The smaller the mean, the higher the original rating. For the first The normalized value of the comprehensive customer rating for each candidate service provider is determined based on the statistical analysis of customer feedback data submitted after the completion of historical work orders. These are the weighting coefficients for the geographical location scoring dimension, technical capability certification scoring dimension, historical response time scoring dimension, and overall customer scoring dimension, respectively. All weighting coefficients are greater than zero and satisfy the following conditions: , The preset work order quantity threshold is a positive integer pre-configured by the system. For work orders where the number of completed work orders in the past is lower than the preset work order number threshold For candidate service providers, the service provider dispatch engine uses a cold start rule, and calculates the comprehensive matching score using the following cold start scoring formula: ; in, For the cold start phase The overall matching score of the candidate service providers Score the candidate service provider's normalized geographic location. This is a normalized technical capability certification score for the candidate service provider. Both scores have the same meaning as in the standard weighted scoring model. The cold start rule only uses two dimensions—geographical location and technical capability certification—that can be determined when the service provider joins the network. This avoids distortion of the overall matching score due to missing historical response time and customer comprehensive rating data. The cold start rule waits until the service provider's historical completed work orders reach a preset work order threshold. Then, it automatically switches to the standard weighted scoring model to calculate the comprehensive matching score; The service provider dispatch engine assigns the service provider with the highest comprehensive matching score among the candidate service providers to the corresponding work order. After the service provider matching is completed, the service provider dispatch engine reads the production task scheduling data of the intelligent inspection robot to be maintained, extracts its production interruption window duration information, and arranges the maintenance time slot within the production interruption window on the premise of meeting the work order response priority requirements, so as to avoid the maintenance operation from interfering with normal production. For emergency maintenance work orders whose response priority does not allow for delay, the maintenance time slot is arranged immediately and a notification is sent to the production scheduling system. Full-process tracking and performance evaluation module: The end-to-end tracking and performance evaluation module writes parts replacement information back to the full lifecycle database in real time during work order execution, ensuring the integrity of equipment lifecycle data. After the work order is completed, this module calculates the service provider's comprehensive performance score based on the following formula: ; in, As part of the service provider's overall performance score, The score for the timeliness of this work order response is determined by the ratio of the actual response time to the time required by the work order's priority; the shorter the actual response time, the higher the score. To maintain the quality score, it is jointly determined by the improvement in the deviation of the intelligent inspection robot's operating parameters before and after maintenance, and the reduction in the product defect rate. The greater the improvement in deviation and the greater the reduction in the defect rate, the higher the score. The compliance score for parts usage is determined by the compliance ratio of the actual parts used during work order execution to the certified parts catalog, item by item. These are the weighting coefficients for the response timeliness dimension, maintenance quality dimension, and parts usage compliance dimension, respectively. All weighting coefficients are greater than zero and satisfy the following conditions: ; The end-to-end tracking and performance evaluation module includes a minimum sample size determination mechanism: after each work order is completed, it first determines whether the service provider's historical number of completed work orders has reached a preset sample size threshold. ( (A positive integer pre-configured by the system) When the number of historically completed work orders reaches a preset sample size threshold. At that time, the system dynamically updates the basic weight coefficient of the service provider in the weighted scoring model of the service provider dispatch engine based on the average comprehensive performance score of the recently completed work orders. This gives service providers with consistently high comprehensive performance scores a higher priority for subsequent work orders, while service providers with consistently low comprehensive performance scores have a correspondingly lower priority for subsequent work orders. When the number of historically completed work orders has not reached the preset sample size threshold... In this case, the service provider's order priority will remain unchanged until the sample size is sufficient before adjusting the priority, in order to avoid the random fluctuation of performance scores leading to unstable order allocation decisions in the case of a small sample size. Data access layer: The data access layer is responsible for establishing data connections with various intelligent detection robots deployed on site. Before each robot joins the network, it must complete device registration with the system and submit metadata such as the device's unique identifier, device type, sensor configuration parameters, and detection capability description. The system will then create a corresponding device file for each device. The data access layer has a built-in protocol conversion engine that supports MQTT, HTTP and WebSocket protocols. The protocol conversion engine converts all heterogeneous raw data uploaded by all access devices into the system's internal standard data format, and performs timestamp calibration (eliminating clock deviations between different devices) and coordinate system normalization (ensuring the comparability of spatial data generated by different robots). The data access layer also maintains the online status information of each robot and monitors the device connection status through a timed heartbeat mechanism. When the heartbeat times out and is determined to be disconnected, the system automatically triggers the disconnection reconnection mechanism and marks the device offline status in the device status panel so that operators can know it in a timely manner. Data Processing and Governance Center: The data processing and governance center comprises two core components: a streaming engine and a batch processing scheduler. The streaming engine performs online cleaning of the real-time uploaded detection data: it identifies and removes duplicate data records, interpolates and fills in missing timestamps based on context, filters out noisy data that deviates significantly from normal working conditions based on preset normal value ranges, and marks suspected abnormal data according to preset anomaly judgment rules. The marking results, along with the original data, are stored in the historical database for subsequent analysis and retrieval. The batch processing scheduler performs in-depth governance of historical data according to a preset scheduling cycle: it calculates a data quality score for each historical data record, standardizes the data format to make the historical data conform to a unified format standard, completes the labels for data records with missing labels, and compresses and archives historical data that meets the archiving conditions to free up online storage space. Supplier quality traceability module: The supplier quality traceability module maintains a real-time data interface with the data processing and governance center. When the streaming engine marks a piece of detection data as a safety anomaly exceeding the safety standard threshold, and the product defect rate data exceeds the preset upper limit threshold, the supplier quality traceability module is triggered to start. The module first queries the material tags of the components used by the intelligent inspection robot that triggered the anomaly. Combined with the robot's bill of materials data, it identifies the batch number and corresponding supplier code of the components used during the anomaly period. Then, it calculates the direct quality cost corresponding to this anomaly event using the following formula: ; in, For direct quality costs, This represents the number of defective parts in this incident, in units of pieces. Value per unit, in yuan. The cost of downtime is expressed in yuan per hour. This represents the duration of the anomaly, in hours. The rework cost for this anomaly is expressed in yuan. The sum of these three items is the direct quality cost corresponding to this anomaly. The calculation results are linked to the corresponding supplier's quality records; The supplier quality traceability module also includes a supplier dynamic rating unit, which uses the supplier code as the primary key and aggregates the anomaly trigger rate and batch weight data of all historical supply batches of that supplier. The supplier's comprehensive quality credit score is calculated using the following formula: ; in, This is a comprehensive quality credit score for suppliers, ranging from zero to one hundred. A higher score indicates more stable historical supply quality. For the first The anomaly trigger rate for each historical supply batch, ranging from zero to one, is equal to the ratio of the number of parts that triggered an anomaly event in that batch to the total number of parts in that batch. For the first The batch weight of each historical supply batch is positively correlated with the quantity supplied in that batch. The total number of historical supply batches used in the calculation, which is a positive integer. This is the batch number, with values ​​ranging from 1 to 2. integers, The value is recalculated after each new abnormal event to maintain real-time correspondence with the supply quality status and to update the latest value. Value updated to supplier credit database; Procurement decision support module: The procurement decision support module periodically reads the supplier credit database and... Suppliers whose values ​​are below a preset warning threshold will generate procurement risk warning information, based on the candidate suppliers. Ranking generates a list of recommended alternative suppliers, based on... The system maps values ​​to preset mapping rules and outputs suggested contract quality guarantee deposit ratios for corresponding suppliers, providing quantitative decision support for the purchasing department and forming a complete closed loop from testing data to purchasing management decisions.

[0019] Example 1, System Overall Architecture: The system described in this invention consists of a data access layer, a data processing and governance center, a work order generation module, a service provider dispatch engine, a full-process tracking and performance evaluation module, a supplier quality traceability module, and a procurement decision support module. Each module interacts with data through standardized interfaces, together forming a closed loop for intelligent maintenance work order management for intelligent inspection robots. Example 2: Implementation of the data access layer: The data access layer is responsible for establishing data connections with various intelligent detection robots deployed on site. In practice, each robot needs to register with the system before joining the network, submitting metadata such as the unique identifier of the device, device type, sensor configuration parameters and detection capability description. The system will then create a corresponding device file for each device. The data access layer has a built-in protocol conversion engine that supports MQTT, HTTP and WebSocket protocols. Different manufacturers' robots may use different communication protocols and data formats. The protocol conversion engine converts all heterogeneous raw data uploaded by all access devices into the system's internal standard data format. At the same time, the protocol conversion engine timestamps the data to eliminate clock deviations between different devices and normalizes the coordinate system to ensure that the spatial data generated by different robots are comparable. The data access layer maintains the online status information of each robot and monitors the device connection status through a timed heartbeat mechanism. When the heartbeat times out and is determined to be disconnected, the system automatically triggers the disconnection reconnection mechanism and marks the device offline status in the device status panel so that operators can know it in time. Example 3: Implementation of the Data Processing and Governance Center: The data processing and governance center comprises two core components: a streaming engine and a batch processing scheduler. The streaming engine performs online cleaning of the real-time uploaded detection data. The cleaning process includes: identifying and removing duplicate data records, interpolating and filling data records with missing timestamps based on context, filtering noisy data that deviates significantly from the normal working state based on preset normal value ranges, marking suspected abnormal data based on preset anomaly judgment rules, and storing the marking results together with the original data in the historical database for subsequent analysis and retrieval. The batch processing scheduler performs in-depth governance of historical data according to a preset scheduling cycle. Specifically, this includes: calculating a data quality score for each historical data record, standardizing the data format to make the historical data conform to a unified format standard, completing the labels for data records with missing labels, and compressing and archiving historical data that meets the archiving conditions to free up online storage space. Example 4: Implementation of the work order generation module: The work order generation module connects to three types of trigger sources to generate three types of maintenance work orders, and has a built-in work order merging and deduplication sub-unit and a manual review bypass interface; The first type of trigger source is the fault warning signal. When the upstream fault warning subsystem outputs the fault type warning status of a specific device, the work order generation module automatically reads the fault type description, the list of involved parts and the estimated maintenance time of the system, encapsulates it into a preventive maintenance work order and writes it into the deduplication judgment process to be merged. The second type of trigger source is the security anomaly level signal. When the streaming engine in the data processing and governance center marks the detected data as a security anomaly, the system outputs the corresponding security anomaly level. The work order generation module converts the anomaly level into the corresponding response priority based on the mapping table between the anomaly level and the service response priority, generates an emergency maintenance work order, and writes the response priority into the work order attribute for the service provider's dispatch engine to use when sorting. The third type of trigger source is the parts replacement record. The full life cycle database records the time nodes and model information of the historical parts replacement for each robot. The system reads the warranty information and standard service life data of each part, calculates its remaining expected service life based on the current cumulative usage time of the parts, and generates a regular maintenance work order in advance when the remaining expected service life is lower than the preset threshold to prevent unplanned downtime caused by the overuse of parts. The work order merging and deduplication subunit compares multiple trigger signals generated by the same intelligent inspection robot within a preset time window. When the comparison results show that multiple signals have the same fault type description and the parts list involved overlaps, the work order merging and deduplication subunit merges these signals into a single work order and records the number of all merged signals in the source signal field of the work order for subsequent traceability. For signals that do not meet the merging conditions, work orders are generated independently for each signal. After the automatic merging is completed, the manual review bypass interface pushes a notification to the operator on duty, showing the signal details involved in this merging and the content of the merged work order. The operator can choose to confirm the automatic merging result or choose to perform a split operation to split the merged work order into multiple independent work orders. The operation result is written to the work order queue in real time and takes effect. Example 5: Implementation of the service provider's order dispatch engine: When each work order enters the dispatch process, the service provider dispatch engine first retrieves a list of service providers that meet the technical capability requirements of the work order from the candidate service provider database, then normalizes the original scores of each dimension, and then calculates the comprehensive matching score of each service provider based on the standard weighted scoring model or the cold start scoring model. Finally, the service provider with the highest comprehensive matching score is assigned to the work order. During the normalization phase, each candidate service provider in the pool was normalized according to four dimensions: geographical location, technical capability certification, historical response time, and overall customer rating, using a normalization formula. The original scores are mapped to a numerical range of zero to one to eliminate dimensional differences; During the scoring calculation phase, it is determined whether the number of completed work orders in the history of each candidate service provider has reached the preset threshold. Those who have reached the threshold will be scored using the standard weighted scoring formula. Calculate the overall matching score; those who do not reach the threshold will be subject to the cold start scoring formula. Calculate the overall matching score to avoid scoring distortion due to missing historical data; Weight coefficients of each dimension It can be configured by the system administrator according to business needs and can be dynamically adjusted as historical performance data of the service provider accumulates; After the service provider matching is completed, the service provider dispatch engine reads the production task scheduling data of the intelligent inspection robot to be maintained, extracts its production interruption window duration information, and arranges the maintenance time slot within the production interruption window on the premise of meeting the work order response priority requirements, so as to avoid maintenance work from interfering with normal production. Example 6: Implementation of the end-to-end tracking and performance evaluation module: After the work order is dispatched, the full-process tracking and performance evaluation module tracks the entire execution process of the work order; During the work order execution process, after the service provider's technicians complete each part replacement operation, they submit information such as the model, quantity, serial number, and replacement time of the replaced parts to the system in real time. The system then writes the above information back to the maintenance record entry of the corresponding equipment in the full lifecycle database to ensure the continuous integrity of the full lifecycle data. After a work order is completed, the system collects key performance indicator data for this work order execution and calculates the response time score accordingly. Maintenance quality score Compliance score for the use of components Substitute into the comprehensive performance score formula Computing service provider comprehensive performance score ; The calculation relies on comparison data of equipment operating parameters before and after maintenance: the system reads the operating parameter values ​​from the most recent inspection data before maintenance and compares them with the corresponding operating parameter values ​​from the first inspection data after maintenance to calculate the deviation improvement rate. Simultaneously, it reads product defect rate data for the corresponding time period before and after maintenance to calculate the defect rate change rate. Both the deviation improvement rate and the defect rate reduction rate are included in the calculation. The more significant the improvement in either indicator, the better. The higher the score; The calculation relies on certified parts catalog data: the system checks the actual parts models used in the work order against the certified parts catalog item by item, using the ratio of compliant parts usage items to all parts usage items as the basis for calculation. Score; Comprehensive performance score After the calculation is completed, the system writes the performance data to the corresponding service provider's performance history and performs a minimum sample size judgment: if the number of completed work orders in the service provider's history reaches the preset sample size threshold... The system dynamically updates the service provider's base weight coefficient in the service provider dispatch engine's weighted scoring model based on the average comprehensive performance score of the recently completed work orders. This results in service providers with consistently high comprehensive performance scores receiving higher priority in subsequent work order dispatches, while service providers with consistently low comprehensive performance scores receive lower priority. If the number of historically completed work orders does not reach the preset sample size threshold... If so, the service provider's order priority will remain unchanged until the sample size is sufficient before implementing priority adjustment. Example 7: Implementation of the Supplier Quality Traceability Module: The supplier quality traceability module maintains a real-time data interface with the data processing and governance center. When the streaming engine marks a piece of detection data as a safety anomaly exceeding the safety standard threshold, and the product defect rate data exceeds the preset upper limit threshold, the supplier quality traceability module is triggered to start. The module first queries the material tags of the components used by the intelligent inspection robot that triggered the anomaly. Combined with the robot's bill of materials data, it identifies the batch number and corresponding supplier code of the components used during the anomaly period. Then, it applies the direct quality cost formula... Calculate the direct quality cost corresponding to this abnormal event. The calculation results are then linked to the corresponding supplier's quality records. The supplier dynamic rating unit uses the supplier code as the primary key, aggregates the anomaly trigger rate and batch weight data of all historical supply batches of that supplier, and applies them according to the comprehensive quality credit score formula. calculate Value, and the latest The value is updated to the supplier credit database, and each new abnormal event triggers a corresponding supplier. Recalculation of the value; The procurement decision support module periodically reads the supplier credit database and... Suppliers whose values ​​are below a preset warning threshold will trigger a procurement risk warning. Simultaneously, based on the candidate suppliers... The ranking generates a list of recommended alternative suppliers, and based on... The system maps values ​​to preset mapping rules and outputs suggested contract quality guarantee deposit ratios for corresponding suppliers, providing quantitative decision support for the procurement department. While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Any variations and modifications can be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the invention, fall within the protection scope defined by the claims of the present invention.

Claims

1. A data management system for intelligent inspection robots, characterized in that, include: The work order generation module is used to receive fault warning signals, safety anomaly level signals, and parts replacement records, and generate preventive maintenance work orders, emergency maintenance work orders, and regular maintenance work orders according to the type of signal received. The work order generation module has a built-in work order merging and deduplication subunit, which is used to compare multiple trigger signals generated by the same intelligent detection robot within a preset time window, merge multiple signals with the same fault type description and overlapping components into a single work order, and record the number of all merged source signals in the work order attributes. The work order generation module also has a built-in manual review bypass interface, which is used to display detailed information of the signals to be merged to the operator. The operator can confirm the automatic merging result or perform a split operation. The operation result is written to the work order queue in real time. The service provider dispatch engine is used to normalize the original scores of four dimensions—geographic location score, technical capability certification score, historical response time score, and customer comprehensive score—before executing the weighted scoring. It linearly maps the original scores of each dimension to the numerical range of zero to one. Then, it calculates the comprehensive matching score of each candidate service provider for the current work order through the weighted scoring model, and assigns the service provider with the highest comprehensive matching score among the candidate service providers to the corresponding work order. The service provider dispatch engine also includes a cold start rule unit, which is used to calculate the comprehensive matching score of a service provider by using only the normalized geographical location score and the normalized technical capability certification score with equal weight when the number of completed work orders in the service provider's history is lower than the preset work order number threshold. The service provider's dispatch engine also schedules maintenance time slots for work orders based on the interruptible window duration of the intelligent detection robot's production. The full-process tracking and performance evaluation module is used to write the parts replacement information back to the full life cycle database in real time during the work order execution process, and calculate the service provider's comprehensive performance score based on the service provider's response timeliness score, maintenance quality score and parts usage compliance score after the work order is completed, and dynamically adjust the service provider's order dispatch priority in the service provider dispatch engine. The full-process tracking and performance evaluation module has a minimum sample size judgment mechanism. The order dispatch priority will only be adjusted when the number of historical completed work orders of the service provider reaches the preset sample size threshold. If the number of historical completed work orders does not reach the preset sample size threshold, the order dispatch priority of the service provider will remain unchanged.

2. The data management system for intelligent inspection robots according to claim 1, characterized in that: When the work order generation module receives a fault warning signal, it extracts the fault type, the components involved, and the estimated maintenance time, and generates a preventive maintenance work order. Upon receiving a security anomaly level signal, the service response priority is determined based on the preset mapping relationship between the anomaly level and the service response priority, and an emergency maintenance work order with the corresponding response priority attribute is generated. Upon receiving a parts replacement record, the system reads the parts' warranty information and standard service life data, calculates the remaining estimated service life based on the parts' current cumulative usage time, and generates a periodic maintenance work order in advance when the remaining estimated service life is lower than the preset service life threshold.

3. A data management system for intelligent inspection robots according to claim 1, characterized in that, The normalization method of the service provider's order dispatch engine for the same scoring dimension is as follows: Using the lower bound of the original score values ​​of all candidate service providers in the current candidate service provider pool in this scoring dimension as the lower bound of normalization and the upper bound of the original score values ​​as the upper bound of normalization, a linear mapping is performed on the original score of each candidate service provider in this dimension to obtain the normalized score of the candidate service provider in this dimension. When all candidate service providers in a certain dimension have the same original score, the normalized score of all candidate service providers in that dimension is set to one.

4. A data management system for intelligent inspection robots according to claim 1, characterized in that: When the cold start rule unit determines that the number of work orders completed by a service provider in the past is lower than the preset work order number threshold, it uses the sum of the normalized geographical location score and the normalized technical capability certification score, each with a weight of half, as the comprehensive matching score of the service provider. Once the number of completed work orders by the service provider reaches the preset work order threshold, it will automatically switch to a standard weighted scoring model that includes four dimensions: geographic location score, technical capability certification score, historical response time score, and customer comprehensive score, to calculate the comprehensive matching score.

5. A data management system for intelligent inspection robots according to claim 1, characterized in that: When calculating the comprehensive performance score of the service provider, the full-process tracking and performance evaluation module determines the response time score by the ratio of the actual response time to the time required by the priority of the work order; determines the maintenance quality score by the improvement of the deviation of the intelligent detection robot's operating parameters and the reduction of the product defect rate before and after maintenance; and determines the compliance score of parts usage by the compliance ratio of the actual parts used in the execution of the work order to the certified parts catalog. The overall performance of service providers is the result of the weighted sum of the scores of the above three items according to their respective weight coefficients. Each weight coefficient is greater than zero and the sum of the three is equal to one. The minimum sample size judgment mechanism dynamically updates the basic weight coefficient of the service provider in the weighted scoring model of the service provider dispatch engine based on the average comprehensive performance score of the service provider's recently completed work orders after the number of historical completed work orders of the service provider reaches the preset sample size threshold.

6. A data management system for intelligent inspection robots according to claim 1, characterized in that: The system also includes a data access layer, which has a built-in protocol conversion engine that supports communication with intelligent detection robots via MQTT, HTTP and WebSocket protocols. This converts heterogeneous data uploaded by each robot into a unified standard data format and performs timestamp calibration and coordinate system normalization on the data. The data access layer also maintains the robot's online status information, heartbeat monitoring information, and disconnection reconnection mechanism.

7. A data management system for intelligent inspection robots according to claim 6, characterized in that: It also includes a data processing and governance center, which contains a streaming engine and a batch processing scheduler; The streaming engine cleans the real-time uploaded detection data, including removing duplicate data, filling in missing timestamps, and filtering noisy data, and marks the data as abnormal according to preset rules. The batch processing scheduler performs periodic in-depth governance on the stored historical data, including data quality scoring, format normalization, label completion, and data compression and archiving.

8. A data management system for intelligent inspection robots according to claim 1, characterized in that: It also includes a supplier quality traceability module, which is used to query the component material labels to identify the batch and supplier code of the components involved when the test data exceeds the preset safety standard threshold. The module calculates the corresponding direct quality cost by multiplying the number of defective parts in this abnormal event by the value of a single product, adding the downtime loss cost per unit time by multiplying the duration of this abnormality, and adding the rework cost of this abnormal event. The calculation result is then linked to the quality record of the corresponding supplier.

9. A data management system for intelligent inspection robots according to claim 8, characterized in that: The supplier quality traceability module also includes a supplier dynamic rating unit. Using the supplier code as the primary key, it aggregates the abnormal trigger rate and batch weight data of all historical supply batches of the supplier. The cumulative sum of the product of the difference between the abnormal trigger rate of each batch and the batch weight of that batch is divided by the cumulative sum of the batch weights of all batches, and then multiplied by one hundred, is used to calculate the supplier's comprehensive quality credit score, which ranges from zero to one hundred. The higher the credit score, the more stable the historical supply quality. The system also includes a procurement decision support module, which automatically generates procurement risk warning information, a list of recommended alternative suppliers, and a suggested percentage for contract quality deposit based on the supplier's comprehensive quality credit score.

Images