Three-break short-distance communication breakpoint resume method based on multi-source data fusion algorithm

Abstract

The invention provides a three-break short-distance communication breakpoint resume method based on a multi-source data fusion algorithm, and the method comprises the steps: collecting multi-source data, carrying out the normalization processing, and calculating the interruption confidence of a physical layer, a protocol session and an application transaction; calculating a three-break judgment index according to the confidence coefficients, and adaptively determining a preset threshold value in combination with the link reliability and the break frequency; identifying an interruption type through a three-interruption judgment result; selecting a breakpoint based on three-break judgment; recording a physical layer frame sequence number, a protocol packet sequence number and an application transaction identifier as a multi-layer breakpoint position index; recording a reserved breakpoint position in advance in the early warning interval; selecting a target breakpoint position according to the actual interruption moment; sending a continuous transmission request control message and receiving a continuous transmission response control message; and the interruption judgment is optimized by adaptively updating the weight coefficient. According to the method, the interruption type can be accurately identified, breakpoint selection is optimized, the retransmission data volume is reduced, the recovery time is shortened, and the performance and reliability of a short-distance communication system are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method for resuming interrupted short-range communication based on a multi-source data fusion algorithm. Background Technology

[0002] In scenarios such as the Internet of Things (IoT), smart homes, in-vehicle terminals, and warehousing and logistics, short-range communication technologies (such as Bluetooth, Wi-Fi, and ZigBee) are widely used for data interaction between terminals. Due to factors such as limited device power consumption, susceptibility to wireless channel obstruction and interference, and constantly changing mobile terminal locations, short-range communication links frequently experience momentary interruptions, dropped sessions, or application-layer transaction interruptions during operation. To ensure business continuity, existing systems generally employ breakpoint resumption mechanisms, retransmitting unfinished data from a certain location after the link is restored.

[0003] Existing breakpoint resumption solutions typically share the following characteristics: First, interruption detection is primarily based on single-layer information, such as the received signal strength and bit error rate at the physical layer, or the number of retransmissions and timeouts at the link / transport layer to determine if the link is interrupted. Interruption criteria generally use fixed thresholds, lacking adaptability to historical operating states, environmental changes, and different service types. Second, in selecting breakpoint locations, most solutions only use coarse-grained sequence numbers or offsets at the transport or application layer, failing to fully utilize multiple location indices at different layers such as physical layer frames, protocol packets, and application transactions. They also rarely reserve breakpoints in advance based on risk trends before an interruption occurs, often requiring the retransmission of large amounts of data that have already been successfully transmitted from earlier positions after an interruption. Third, information at each layer is independent, lacking effective integration of multi-source data such as physical layer link quality, link layer buffer status, application layer session status, device power supply status, and operating environment parameters. This makes it difficult to distinguish between physical link interruptions, protocol session interruptions, and application transaction interruptions in a timely and accurate manner, hindering the selection of appropriate breakpoint levels and resumption strategies for different interruption types.

[0004] In complex and dynamic short-range communication environments, the existing technologies mentioned above are prone to misjudgment or omission in interruption detection, have overly coarse granularity in breakpoint location, involve large amounts of retransmitted data, have long application-layer service session recovery times, and result in low overall system reliability and resource utilization efficiency. Therefore, a new short-range communication breakpoint resumption method is needed. This method should be able to accurately identify the interruption type based on multi-source data fusion when an anomaly occurs in the short-range communication link, and determine the breakpoint location in advance and with precision by combining cross-layer breakpoint location indexes, thereby reducing retransmission redundancy and shortening service recovery time. Summary of the Invention

[0005] To overcome the aforementioned technical deficiencies, the present invention aims to provide a three-segment short-range communication breakpoint resumption method based on a multi-source data fusion algorithm. The present invention constructs physical layer interruption confidence, protocol session interruption confidence, and application transaction interruption confidence based on multi-source data fusion, adaptively determines thresholds by combining link reliability and interruption frequency, and uses a multi-layer breakpoint location index composed of physical layer frame sequence number, protocol packet sequence number, and application transaction identifier for advance reservation and selection, thereby accurately identifying the interruption type and precisely determining the breakpoint location to reduce the amount of retransmitted data and shorten the service recovery time.

[0006] This invention discloses a method for resuming interrupted short-range communication based on a multi-source data fusion algorithm, comprising the following steps:

[0007] S1. After the short-range communication link is established, a multi-source communication status data set is obtained for each short-range communication link. The multi-source communication status data set includes at least physical layer link quality data, link layer buffer status data, application layer service session status data, short-range communication equipment power supply status data, and working environment parameter data.

[0008] S2, within a preset sliding time window, sample the multi-source communication state data set, and perform time alignment and numerical normalization processing on the sampled data within each preset sliding time window, mapping the multi-source communication state data set into normalized index variables. , , , , , , And calculate the physical link interruption confidence level according to formulas (1) to (3). Protocol session interruption confidence level and application transaction interruption confidence :

[0009] (1) / (2) / (3);

[0010] in, This is a normalized signal quality index obtained based on physical layer link quality data. This is the normalized bit error rate metric obtained from physical layer link quality data. This is a normalized link retransmission count metric obtained based on link layer cache state data. This is a normalized session hold time metric obtained based on application layer business session state data. This is a normalized connection rebuild count metric obtained from application layer business session state data. This is a normalized application data discard ratio metric obtained based on the application layer data cache status. This is a normalized application request timeout ratio metric obtained based on the application layer request processing status. , , , , , , These are positive weighting coefficients;

[0011] S3, calculate link reliability indicators based on historical multi-source communication status data. and interruption frequency index And calculate the preset threshold according to formula (4). : (4);

[0012] in, This is a normalized link reliability index obtained based on historical multi-source communication status data. This is a normalized interruption frequency index obtained based on the number of historical interruptions. , The positive weighting coefficients are... The preset threshold;

[0013] S4, Calculate the three-section judgment index Three-part judgment indicators for , , The maximum value in the three-part judgment index Greater than or equal to the preset threshold At that time, the current interruption type is determined to be one of three types of interruption: physical link interruption, protocol session interruption, or application transaction interruption, based on the confidence level of the corresponding maximum value. These three interruption types are collectively referred to as the three interruptions, and the corresponding breakpoint location identifier is obtained.

[0014] S5 maintains physical layer frame sequence number, protocol packet sequence number, and application transaction identifier for each short-range communication link. The physical layer frame sequence number is used as the breakpoint location index for physical link-level breakpoints, the protocol packet sequence number is used as the breakpoint location index for protocol session-level breakpoints, and the application transaction identifier is used as the breakpoint location index for application transaction-level breakpoints. During normal transmission of the short-range communication link, when the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier reach a preset interval, the corresponding breakpoint location index is automatically recorded.

[0015] S6, Set warning threshold Warning threshold Less than the preset threshold When the three-part judgment index Less than the preset threshold And greater than or equal to the warning threshold The physical layer frame number, protocol packet number, and application transaction identifier of the next moment are recorded as reserved breakpoint position indexes.

[0016] S7. When an interruption of the short-range communication link is detected, the corresponding breakpoint location index is selected as the target breakpoint location index according to the interruption type determined in step S4. If there is a reserved breakpoint location index that is later than the most recent normal transmission confirmation location index, the corresponding reserved breakpoint location index is used as the target breakpoint location index. Otherwise, the most recent normal transmission confirmation location index is used as the target breakpoint location index. The starting boundary of the data unit to be resumed is determined according to the target breakpoint location index. The data unit to be resumed includes the physical layer frame unit corresponding to the physical link interruption, the protocol packet unit corresponding to the protocol session interruption, and the application transaction unit corresponding to the application transaction interruption.

[0017] S8, after the short-range communication link is restored, a resume request control message carrying the interruption type and the target breakpoint location index is sent. The resume request control message includes an integrity verification field for verifying the integrity of the data unit to be resumed and a session identifier field for identifying the service session. At the receiving end, a resume response control message is generated based on the integrity verification field and the session identifier field. At the sending end, after receiving the resume response control message, the breakpoint is resumed according to the starting boundary of the data unit to be resumed, and the corresponding application layer service session is restored after all the data units to be resumed have been confirmed for transmission.

[0018] S9, after completing a breakpoint resume process, determines the actual interrupt level based on the interrupt type carried in the resume request control message and the diagnostic results collected after the short-range communication link is restored. When the actual interrupt level is inconsistent with the corresponding interrupt type, the positive weighting coefficient is adjusted according to the preset update rules. , , , , , , At least one of the values ​​is selected, and the adjusted positive weighting coefficient is used to recalculate the value in the subsequent breakpoint resume process. , , .

[0019] Compared with existing technologies, the above technical solution has the following advantages:

[0020] 1. The technical solution adopted in this invention: In step S3, a normalized link reliability index is calculated based on long-term historical data. and interruption frequency index , and according to Such a linear combination relationship yields a preset threshold. By adjusting the coefficient , It enables scenario-based configuration of reliability and interruption frequency weights, with preset thresholds adaptively changing with overall link stability and interruption frequency. This allows interruption decisions to balance sensitivity and robustness in different environments, significantly reducing false alarm and missed alarm rates in test scenarios and improving the reliability of interruption detection.

[0021] 2. The technical solution adopted in this invention: In step S6, a warning threshold is introduced. When the three-part judgment index Falling into the warning zone Instead of immediately determining an interruption, the system records the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier of the next moment as an index for the reserved breakpoint location. This can be combined with cooling time and trend determination to avoid frequent recording. In the stage where the risk of interruption has just increased but has not yet actually occurred, the system reserves a breakpoint location close to the moment of interruption in advance, so that the retransmission starting point after the actual interruption is closer to the real disconnection location. This significantly reduces the amount of retransmitted data and recovery latency while ensuring reliability. Attached Figure Description

[0022] Figure 1 This is a schematic curve diagram illustrating the status data of multi-source communication.

[0023] Figure 2 This is a schematic diagram illustrating how the confidence levels of three types of interruptions change with the sliding time window;

[0024] Figure 3 For preset threshold With link reliability Schematic diagram of the changes;

[0025] Figure 4 Three types of interruption confidence levels and preset thresholds. Schematic diagram showing changes over time;

[0026] Figure 5 A schematic diagram showing the breakpoint locations at the physical layer, protocol layer, and application layer.

[0027] Figure 6 Three-part judgment index With preset threshold and warning thresholds Schematic diagram;

[0028] Figure 7 A diagram illustrating the selection of the target breakpoint location index;

[0029] Figure 8 This is a schematic diagram illustrating the adaptive update of weight coefficients.

[0030] Figure 9 This is a schematic diagram illustrating the steps of the three-segment short-range communication breakpoint resumption method based on the multi-source data fusion algorithm of the present invention. Detailed Implementation

[0031] See Figure 9 As shown in this embodiment, a method for resuming interrupted short-range communication based on a multi-source data fusion algorithm is provided. The method includes: collecting multi-source data and performing normalization processing; calculating the interruption confidence levels of the physical layer, protocol session, and application transaction; calculating the three-interruption judgment index based on these confidence levels and adaptively determining a preset threshold by combining link reliability and interruption frequency; identifying the interruption type through the three-interruption judgment results; selecting the interruption point based on the three-interruption judgment; recording the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier as multi-layer interruption location indexes; pre-reserving interruption locations within the warning interval; selecting the target interruption location based on the actual interruption time; sending a resumption request control message and receiving a resumption response control message; and optimizing the interruption judgment by adaptively updating the weight coefficients. This method can accurately identify the interruption type, optimize interruption point selection, reduce the amount of retransmitted data, shorten recovery time, and significantly improve the performance and reliability of the short-range communication system.

[0032] In this embodiment, step S1 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on multi-source data fusion algorithm, step S1 is used to continuously acquire the set of multi-source communication status data related to the short-range communication link after the short-range communication link is established, and to uniformly manage the acquisition process to ensure that various types of data have alignability and comparability in the time dimension and link dimension during subsequent multi-source data fusion processing.

[0033] In one embodiment, after the short-range communication terminal establishes a short-range communication link with the peer device via Bluetooth short-range communication protocol, Wi-Fi Direct short-range communication protocol, or ultra-wideband short-range communication protocol, the short-range communication terminal first assigns a unique link identifier to the current short-range communication link. The link identifier can be in unsigned integer form and is managed uniformly through a link identifier allocation table. When a new short-range communication link is established, an unused link identifier is taken from the set of idle link identifiers and assigned to that short-range communication link. When the short-range communication link is released, the corresponding link identifier is reclaimed, thus ensuring that no link identifier conflicts occur between different short-range communication links during the system's operating cycle. Afterward, the short-range communication terminal activates the multi-source communication status data acquisition module, which periodically collects multi-source communication status data related to the short-range communication link or collects data when specific trigger conditions are met. Each acquisition operation is appended with a unified format timestamp and the corresponding link identifier through a unified time synchronization module and written to the multi-source data buffer, thereby forming a multi-source communication status data set for the current short-range communication link.

[0034] The multi-source communication status data set includes at least physical layer link quality data, link layer buffer status data, application layer service session status data, short-range communication device power supply status data, and operating environment parameter data. Physical layer link quality data reflects the channel status of the short-range communication link at the physical layer. In a preferred embodiment, the multi-source communication status data acquisition module reads the statistical register inside the short-range communication chip or RF front-end through the physical layer driver module in each physical layer sampling period to obtain the average received signal strength indication value, average signal-to-noise ratio estimate, and cumulative bit error rate within that sampling period. This is combined with the total number of bits received within that sampling period to calculate the bit error rate. These parameters are then combined to constitute the physical layer link quality data for the current moment. Under conditions where the distance between the short-range communication terminal and the peer device is approximately 5 meters and environmental interference is minimal, the average received signal strength indication value can be stabilized at around -60dBm in actual testing, and the bit error rate is within a certain range. The magnitude is significant; when the distance increases to approximately 15 meters and multiple co-channel interference sources are present in the vicinity, the average received signal strength indication may drop below -80 dBm, and the bit error rate may rise to [a certain level]. Scale. By continuously collecting physical layer link quality data in this way, the deterioration trend of the physical layer link caused by factors such as channel fading, obstruction, or increased interference can be identified during subsequent multi-source data fusion processing, providing basic data for judging the risk of physical link outage.

[0035] Link layer buffer status data is used to reflect the buffer load and retransmission status of short-range communication links at the link layer. In one embodiment, the multi-source communication status data acquisition module reads the number of data units to be sent in the link layer transmit buffer queue and the number of data units to be processed in the link layer receive buffer queue through the link layer protocol stack module in each link layer sampling period, and counts the number of link layer retransmission events and the number of link layer drop events that occur in that sampling period. The above parameters are combined to form the link layer buffer status data at the current moment. When the short-range communication link is in a stable working state and the service load is relatively stable, the length of the link layer transmit buffer queue and the length of the link layer receive buffer queue are generally kept within a small range, and the number of link layer retransmission events is basically 0. When the upper layer application suddenly generates a large amount of data or the short-range communication link temporarily decreases in capacity, the length of the link layer transmit buffer queue will increase significantly, and the number of link layer retransmission events and drop events will also increase significantly. By continuously collecting link layer buffer status data in step S1, anomalies at the link layer level can be detected in time before the physical layer link quality data shows significant deterioration, providing a basis for distinguishing between protocol session interruption and physical link interruption.

[0036] Application layer service session status data is used to reflect the operational status of short-range communication links at the application layer service session dimension. In a specific embodiment, when a short-range communication terminal establishes an application layer service session with the peer device, the application layer session management module remembers the service session start time and assigns a service session identifier to the service session. Subsequently, in each sampling process, the application layer session management module calculates the service session hold time based on the difference between the current time and the service session start time, and simultaneously counts the number of application layer service session reconstructions, the number of application layer request timeouts, and the number of application layer requests still waiting for a response within the current sampling period. These parameters are combined to constitute the application layer service session status data at the current moment. When both the physical layer and the link layer of the short-range communication link are in normal condition, if the server-side processing capacity decreases, the application layer service malfunctions, or there are logical errors in the application layer protocol interaction, the number of application layer request timeouts will increase significantly, and the number of application layer service session reconstructions will also increase. By obtaining the application layer service session status data in step S1, application service anomalies can be detected before the short-range communication link is completely disconnected, providing direct input for subsequent identification of application transaction interruptions.

[0037] The power supply status data of short-range communication devices is used to reflect the operating status of the power system of the short-range communication terminal. In a preferred embodiment, the short-range communication terminal integrates a power management module, which periodically acquires the battery percentage, battery voltage, power output current, and power output stability indication value obtained through power voltage fluctuation analysis. When the battery level drops by more than 5% or the battery voltage drops by more than a preset threshold, the power management module can immediately trigger the acquisition of short-range communication device power supply status data and write it to the multi-source data buffer.

[0038] Environmental parameters are used to reflect the temperature, humidity, and other environmental conditions of the short-range communication terminal's environment. In one embodiment, the short-range communication terminal periodically collects ambient temperature and humidity data through an environmental sensor module, and optionally collects device attitude change information through an accelerometer to determine whether the terminal is in motion. Test results in different indoor and outdoor environments show that when the ambient temperature increases from 25°C to 45°C and the ambient humidity increases from 40% to 80%, under the premise that other conditions remain unchanged, the average bit error rate of the short-range communication link decreases from... Rise to The number of instantaneous physical layer interruptions increases by approximately 40%. By collecting working environment parameter data in step S1 and then correlating and analyzing it with physical layer link quality data, link layer buffer status data, and short-range communication equipment power supply status data in subsequent multi-source data fusion, it is possible to more accurately distinguish between recoverable performance fluctuations caused by environmental changes and continuous performance degradation caused by hardware failures, thereby improving the reliability of interruption type determination.

[0039] During the acquisition of multi-source communication status data sets, the multi-source communication status data acquisition module preferably adopts an adaptive sampling scheduling strategy. The unified sampling scheduling module dynamically adjusts the sampling period of different types of data based on the current service load of the short-range communication link, historical interruption frequency, and statistical results of application layer service session status data. For example, when the statistical results show that the number of short-range communication link interruptions is low in the past 24 hours and the number of application layer request timeouts in the application layer service session status data remains at a low level, the sampling period of physical layer link quality data and link layer cache status data can be adjusted from 100 milliseconds to 500 milliseconds to reduce the system overhead of data acquisition and processing. When the number of short-range communication link interruptions increases significantly in the past hour, or the number of application layer request timeouts in the application layer service session status data suddenly increases, the unified sampling scheduling module can shorten the sampling period of physical layer link quality data, link layer cache status data, and application layer service session status data to 50 milliseconds or even 20 milliseconds, and increase the acquisition frequency of short-range communication equipment power supply status data and operating environment parameter data, so as to capture signs of impending interruptions at a finer-grained time scale. In addition, trigger rules can be set to automatically increase the sampling frequency of the corresponding short-range communication link when the bit error rate continues to be higher than the empirical threshold or the number of link layer retransmission events continues to rise within several consecutive sampling windows, so as to achieve key monitoring of potentially faulty links.

[0040] In scenarios where multiple short-range communication links operate concurrently, the multi-source data buffer can be divided into multiple logical sub-buffers according to the link identifier. Each logical sub-buffer corresponds to a set of multi-source communication status data for one short-range communication link. A unified time synchronization module provides a unified time reference for all acquisition operations, ensuring that the multi-source communication status data sets of different short-range communication links all have timestamps in a unified format. Thus, in subsequent processing based on multi-source data fusion algorithms, for any short-range communication link, physical layer link quality data, link layer buffer status data, application layer service session status data, short-range communication device power supply status data, and operating environment parameter data within the same time window can be extracted from the logical sub-buffer corresponding to that short-range communication link for fusion calculation. It can also perform horizontal comparisons of similar data between different short-range communication links to identify local environmental interference and systemic anomalies.

[0041] like Figure 1 As shown, Figure 1The horizontal axis represents time, and the vertical axis represents the normalized values ​​of the state variables. In the legend, the curves for "Physical Layer Link Quality Data" and "Link Layer Buffer State Data," "Application Layer Service Session State Data," "Short-Range Communication Equipment Power Supply State Data," and "Operating Environment Parameter Data" represent the normalized values ​​of the operating environment parameters over time.

[0042] In this embodiment, step S2 will be described in detail. Step S2 is used to sample the multi-source communication state data set within a preset sliding time window, and perform time alignment and numerical normalization processing on the sampled data within each preset sliding time window, thereby mapping the multi-source communication state data set obtained in step S1 into a set of normalized index variables, including the normalized signal quality index corresponding to the physical layer link quality data. Normalized bit error rate (NBER) metric corresponding to physical layer link quality data Normalized link retransmission count metric corresponding to link layer buffer state data Normalized session persistence time metrics corresponding to application layer business session state data Normalized connection reconstruction count metric corresponding to application layer business session state data Normalized application data discard ratio metric corresponding to application layer business session state data And the normalized application request timeout ratio metric corresponding to application layer business session state data. Based on this, the confidence level of physical link interruption is calculated. Protocol session interruption confidence level Application transaction interruption confidence .

[0043] In one embodiment, for the multi-source communication status data set that has been written to the multi-source data buffer in a unified format timestamp in step S1, the multi-source communication status data acquisition module first determines a time range based on the length of a preset sliding time window, which serves as the time interval corresponding to the current preset sliding time window. The length of the preset sliding time window can be set according to the current transmission rate and target sampling delay of the short-range communication link, so that a sufficient number of physical layer link quality data, link layer buffer status data, application layer service session status data, short-range communication device power supply status data, and working environment parameter data samples can be covered within the time window. Subsequently, the multi-source communication status data acquisition module uses the time reference provided by the unified time synchronization module to select multi-source communication status data samples whose timestamps fall within the range of the current preset sliding time window from the multi-source data buffer, and maps the sampling points from different data sources to a unified time grid through a time alignment strategy.

[0044] During time alignment, timestamp-based interpolation or nearest-neighbor-based resampling methods can be employed. For example, a preset sliding time window can be divided into several equally spaced time periods, and the sampled values ​​of each type of multi-source communication state data within these time periods can be interpolated or averaged to obtain a sampling sequence corresponding to a unified time grid. For cases where no sampled values ​​exist within a certain time period, forward hold, backward hold, or linear interpolation can be used to fill the missing values, ensuring that each type of multi-source communication state data has a sampling sequence of consistent length within the preset sliding time window, facilitating subsequent numerical normalization and statistical feature extraction.

[0045] After time alignment is completed, numerical normalization processing needs to be performed on the sampled data within the preset sliding time window. The purpose of numerical normalization is to map physical layer link quality data, link layer buffer state data, and application layer service session state data with different dimensions and value ranges to the same or similar numerical ranges, thereby preventing one type of data from having an unreasonable dominant effect on other data during multi-source data fusion due to excessively large dimensions. In a preferred embodiment, a linear normalization method can be used to map the received signal strength indicator and bit error rate corresponding to the physical layer link quality data, the link layer retransmission count corresponding to the link layer buffer state data, and the session hold time, connection reconstruction count, application data drop ratio, and application request timeout ratio corresponding to the application layer service session state data to... Range. Taking the received signal strength indicator (RSSI) value corresponding to physical layer link quality data as an example, the maximum and minimum values ​​of the RSI value can be determined during the system deployment phase by statistically analyzing a large number of RSI values ​​under normal working scenarios. Then, the average RSI value within the current window can be mapped to a normalized signal quality index according to a linear mapping relationship. Similarly, the bit error rate corresponding to the physical layer link quality data can be mapped to a normalized bit error rate metric according to the set upper and lower limits of the bit error rate. The link layer retransmission count corresponding to the link layer buffer state data is mapped to a normalized link retransmission count index according to the preset maximum retransmission count. The session persistence time, connection rebuild count, application data discard ratio, and application request timeout ratio corresponding to application layer business session state data are mapped to normalized session persistence time metrics. Normalized connection reconstruction frequency index Normalized application data discard ratio index and normalized application request timeout ratio metric .

[0046] In another embodiment, to enhance the robustness of numerical normalization to outliers, a quantile-based normalization strategy can be adopted. This involves calculating specific quantiles from historical data to determine a normal range. Linear normalization is performed when the sampled data within the current preset sliding time window falls within this normal range. When the sampled data exceeds the normal range, clipping or stretching methods are used to limit the impact of extreme outliers on the normalization result. For example, the 95th quantile of historical bit error rate data can be used as an effective upper limit for the bit error rate. Bit error rate values ​​exceeding this upper limit are directly mapped to the normalized value corresponding to the upper limit during normalization, thereby avoiding the impact of extreme interference occurring in a short period on the normalized bit error rate index. The change was too sudden and drastic.

[0047] After completing the numerical normalization process, the multi-source data fusion module uses the normalized index variables... Normalized index variables Normalized index variables Normalized index variables Normalized index variables Normalized index variables and normalized index variables Calculate the physical layer interruption confidence level Protocol session interruption confidence level and application transaction interruption confidence In a preferred embodiment, each confidence level can be calculated using the following weighted formula: , , , where the coefficient ,coefficient ,coefficient ,coefficient ,coefficient ,coefficient ,coefficient ,coefficient The positive weighting coefficient can be configured based on different short-range communication service scenarios or optimized through training with historical data. (Physical layer interruption confidence level) Protocol session interruption confidence level is used to characterize the risk of physical link interruption occurring within the current preset sliding time window. Used to characterize the risk level of protocol session interruption, application transaction interruption confidence level. This is used to characterize the degree of risk of application transaction interruption. The weighted formula above linearly weights multiple normalized indicator variables according to their importance, which helps to highlight indicators that are more sensitive to interruption determination in different business scenarios.

[0048] In an improved embodiment, to enhance the physical layer interruption confidence... Protocol session interruption confidence level and application transaction interruption confidence To improve discriminative power, a non-linear mapping step can be added after the linear combination of normalized indicator variables and positive weight coefficients. For example, a sigmoid function or piecewise linear function can be used to compress the weighted result, resulting in slower confidence changes when the normalized indicator variables are within the normal range, and faster changes when the normalized indicator variables are close to or exceed the threshold range. This design allows for more sensitive detection of impending short-range communication link interruption risks in subsequent steps, while avoiding excessive false alarms within normal fluctuation ranges.

[0049] like Figure 2 As shown, Figure 2 The horizontal axis represents the index of the preset sliding time window, and the vertical axis represents the confidence level value. The legend shows the "Physical Link Disruption Confidence Level". The curve represents the trend of physical layer interruption confidence as a function of a preset sliding time window index. The legend shows "Protocol Session Interruption Confidence". The curve represents the trend of protocol session interruption confidence as a function of the preset sliding time window index. The legend shows "Application Transaction Interruption Confidence". The curve represents the trend of application transaction interruption confidence as a function of the preset sliding time window index.

[0050] In this embodiment, step S3 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on a multi-source data fusion algorithm, step S3 is used to quantitatively evaluate the overall stability of the short-range communication link over a long time scale using historical multi-source communication state data, and obtain a normalized link reliability index based on this. and interruption frequency index Then, the preset threshold is calculated according to the preset linear combination relationship. Preset threshold This will be used as the physical layer interruption confidence level in the subsequent three-fragment determination process. Protocol session interruption confidence level Application transaction interruption confidence The comparison benchmark is used to achieve adaptive adjustment of the interruption judgment threshold for different link environments.

[0051] In one embodiment, for the multi-source communication state data set and its corresponding three confidence levels obtained in steps S1 and S2, the system maintains a historical record for each short-range communication link. This historical record can be divided into multiple historical sliding time windows in chronological order, with each historical sliding time window corresponding to a set of pre-calculated normalized index variables. Normalized index variables Normalized index variables Normalized index variables Normalized index variables Normalized index variables Normalized index variables and the corresponding physical layer interruption confidence level Protocol session interruption confidence level Application transaction interruption confidence level It also records whether a short-range communication link interruption event actually occurred within the historical sliding time window and the type of the interruption event. During system operation, only the historical records within the most recent period can be retained, such as retaining historical sliding time window information from the most recent hours or days, in order to calculate link reliability indicators. and interruption frequency index It can reflect the current link environment without being dominated by historical situations that are too far back.

[0052] Normalized link reliability index Used to comprehensively characterize the statistical stability of short-range communication links. In a preferred embodiment, the normalized link reliability index... The calculation can be based on the proportion of time during which short-range communication links are in normal working condition within a historical sliding time window. Specifically, a binary marker can be recorded within each historical sliding time window to indicate whether the short-range communication link remains continuously available within that time window. If no physical link interruption, protocol session interruption, or application transaction interruption occurs within that time window, then that time window is marked as a normal link window; otherwise, it is marked as a link abnormal window. Assuming that in the most recent... The number of normal links within each historical sliding time window is The number of link anomaly windows is Then the original link reliability value can be calculated according to the following relationship: Based on this, in order to maintain consistency with other normalized index variables, the original link reliability value can be... Directly used as a normalized link reliability indicator Alternatively, weighting can be applied by incorporating the weighting factor of historical sliding time windows when necessary. For example, more weight can be assigned to more recent historical sliding time windows, and less weight to earlier ones, to construct a weighted link reliability index with a "forgetting factor." In a specific test, when a simple averaging method is used to calculate the normalized link reliability index... At that time, for a continuous and stable indoor short-range communication link, It remains above 0.9 for a long period; for environments with periodic interference, It typically fluctuates between 0.6 and 0.8; however, for short-range communication links that are frequently blocked or move frequently, It may remain below 0.5 for an extended period.

[0053] Interrupt frequency index This is used to characterize the frequency of outages in short-range communication links over a certain time scale. In one embodiment, an outage statistics counter can be maintained in the system to record the total number of outages detected within a specified historical time window. Assuming that in the recent... Within a historical time unit (e.g., the most recent) Minutes, recent Hours or recent (days), the total number of interrupt events detected was Then the original interrupt frequency can be defined as To facilitate integration with normalized link reliability metrics. Together they are used in the calculation of preset thresholds, and the interruption frequency index. By adjusting the original interrupt frequency This is obtained by performing a normalized mapping. For example, it can be based on the maximum acceptable interrupt frequency determined during system design. Original interrupt frequency Mapped to according to linear relationship Interval: Among them, when the original interrupt frequency Less than or equal to the maximum acceptable interrupt frequency Interruption frequency index It varies linearly within the range of 0 to 1, when the original interrupt frequency... Exceeding the maximum acceptable interrupt frequency Interruption frequency index It is limited to 1 to avoid excessively large values ​​in extreme cases.

[0054] In another embodiment, to more fully reflect the temporal distribution of interruption events, an interruption frequency index can be calculated. This involves introducing statistical information about interrupt intervals. For example, the time interval between two adjacent interrupt events can be calculated, and the interrupt frequency index can be obtained by inversely normalizing the time interval. This results in shorter time intervals, i.e., more frequent interruptions, increasing the interrupt frequency index. The larger the value, the higher the interruption frequency index when the time interval is long. Relatively small. This method can distinguish between "concentrated burst interruptions" and "sparsely distributed interruptions" when the total number of interruptions is the same, thus setting a preset threshold. The settings provide more detailed information.

[0055] After obtaining the normalized link reliability index and interruption frequency index Next, the preset threshold needs to be calculated according to the linear combination relationship. In a preferred embodiment, a preset threshold is used. The calculation can be performed using the following formula: , where the coefficient Sum of coefficients These are positive weighting coefficients used to represent the normalized link reliability index. and interruption frequency index In forming a preset threshold The relative importance of time. Due to the normalized link reliability index A higher value indicates a more stable short-range communication link, and the frequency of interruptions is also a key indicator. A larger value indicates more frequent interruptions, therefore the coefficient can be adjusted. Sum of coefficients The value of this parameter is used to define the threshold adjustment strategy for different business scenarios. In scenarios that are sensitive to real-time performance but allow for a certain degree of misjudgment, the coefficient can be increased. The value of makes the interrupt frequency index For preset threshold The impact is more pronounced, so that the preset threshold can be lowered when the interrupt frequency increases. This improves sensitivity to future interruptions; in scenarios with low tolerance for false alarms and a need to maintain high decision reliability, the coefficient can be increased. The value of makes the normalized link reliability index It plays a dominant role in the calculation of the preset threshold and maintains a high preset threshold when the overall reliability of the link is high. This reduces false alarms.

[0056] like Figure 3 As shown, Figure 3 The horizontal axis represents the normalized link reliability index. The vertical axis represents the preset threshold. The legend shows the "interruption frequency index". The curve represents the preset threshold when the interruption frequency index is 0.2. With normalized link reliability index The changing trend of the "interruption frequency index" The curve represents the preset threshold when the interruption frequency index is 0.5. The changing trend of the "interruption frequency index" The curve represents the preset threshold when the interruption frequency index is 0.8. The changing trend.

[0057] In summary, step S3 constructs a normalized link reliability index by statistically analyzing and normalizing historical multi-source communication status data. and interruption frequency index and using coefficients Sum of coefficients The linear combination relationship yields the preset threshold. This mechanism enables the preset threshold to adapt to the long-term operating status of the short-range communication link, providing a threshold basis that balances robustness and sensitivity for subsequent three-break judgment. This is beneficial for improving the accuracy of interruption identification and the overall performance of breakpoint resume control in complex and variable short-range communication environments.

[0058] In this embodiment, step S4 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on the multi-source data fusion algorithm, step S4 is used to calculate the physical layer interruption confidence level obtained in step S2. Protocol session interruption confidence level Application transaction interruption confidence Construct a three-stage judgment index And the three-part judgment index Compared with the preset threshold obtained in step S3 Comparison was made, and the three-part judgment indicator was detected. Exceeding the preset threshold At that time, it is determined that the most likely interruption type is one of physical link interruption, protocol session interruption, or application transaction interruption. At the same time, the breakpoint location identifier corresponding to the interruption type is obtained, providing basic information for subsequent breakpoint location index selection and breakpoint resume control.

[0059] In one embodiment, for each preset sliding time window, the system has already utilized the normalized index variable in step S2. Normalized index variables Normalized index variables Normalized index variables Normalized index variables Normalized index variables Normalized index variables Calculate the physical layer interruption confidence level Protocol session interruption confidence level Application transaction interruption confidence Meanwhile, in step S3, the normalized link reliability index is used... and interruption frequency index Calculate the preset threshold In step S4, the multi-source data fusion module first interrupts the confidence level at the physical layer. Protocol session interruption confidence level Application transaction interruption confidence The largest value is selected as the three-section judgment index for the current preset sliding time window. ,Right now: After obtaining the three-part judgment index Next, the multi-source data fusion module will use the three-stage judgment indicators. With preset threshold Compare the three criteria. Less than the preset threshold At this point, it is considered that the current short-range communication link has not yet reached the level of interruption risk that requires the activation of breakpoint resume control. Therefore, only the physical layer interruption confidence level is set. Protocol session interruption confidence level Application transaction interruption confidence level and the three-part judgment index Record it in the historical status and keep the short-range communication link operating in normal mode. When the three-disconnection judgment index... Greater than or equal to the preset threshold If this occurs, it indicates that the multi-source communication status data set within the current preset sliding time window has reflected a significant risk of interruption. It is necessary to choose among the three interruption types: physical link interruption, protocol session interruption, and application transaction interruption, and prepare corresponding level of breakpoint location identifiers for the subsequent breakpoint resume process.

[0060] In a preferred embodiment, when the three-part judgment index Greater than or equal to the preset threshold At that time, the multi-source data fusion module compares the confidence levels of physical layer interruptions. Protocol session interruption confidence level Application transaction interruption confidence The size determines the corresponding interrupt type. If the physical layer interrupt confidence level... The maximum value among the three and the physical layer interruption confidence level Greater than or equal to the preset threshold If the current interruption type is determined to be a physical link interruption, the physical layer frame sequence number will be used as the breakpoint location index when selecting the breakpoint location index subsequently; if the protocol session interruption confidence level is low... The maximum of the three and the confidence level of protocol session interruption. Greater than or equal to the preset threshold If the current interruption type is determined to be a protocol session interruption, the breakpoint location index established by the protocol packet sequence number will be used first when selecting subsequent breakpoint locations; if the application transaction interruption confidence level is not specified... The maximum of the three and the application of transaction interruption confidence. Greater than or equal to the preset threshold If the current interrupt type is determined to be an application transaction interrupt, the breakpoint location index established by the application transaction identifier will be given priority when selecting subsequent breakpoint locations. These three interrupt types are collectively referred to as the three interrupts.

[0061] Two special cases may occur during the above determination process. One case is that the physical layer interrupts the confidence level. Protocol session interruption confidence level Application transaction interruption confidence There are two or three cases where the confidence scores are similar, and another case is where all confidence scores are slightly higher than the preset threshold. However, the differences between them are minor. To address these scenarios, an improved embodiment can introduce interrupt type priority rules and confidence difference judgment rules. When the physical layer interrupt confidence... Protocol session interruption confidence level Application transaction interruption confidence There are multiple values ​​in the data that are close to the three-section judgment index. When the confidence level is close, the difference between the maximum confidence level and the second-highest confidence level can be calculated first. If the difference exceeds the preset difference threshold, the interrupt type corresponding to the maximum confidence level is taken as the current interrupt type. If the difference is less than the preset difference threshold, the interrupt type can be selected according to the preset priority rules. For example, the priority order among physical layer interruption, protocol session interruption and application transaction interruption can be set as physical layer interruption takes precedence over protocol session interruption, and protocol session interruption takes precedence over application transaction interruption. This allows the more robust interrupt type to be selected when the confidence levels are close.

[0062] In another improved implementation, to reduce the occurrence of frequent interruption type switching within a short period of time, a three-interrupt determination index can be used. Based on this, a time continuity constraint is introduced. Specifically, a memory variable can be maintained in the system to record the most recent interruption type determination result and the corresponding three-interrupt determination index. The value is determined by the fact that, over several subsequent preset sliding time windows, the confidence level corresponding to a certain interruption type remains at a preset threshold. When the interrupt type is nearby and changes gradually, a bias can be applied to the most recent interrupt type during interrupt type determination to avoid frequent jumps between physical layer interruptions, protocol session interruptions, and application transaction interruptions. This design helps maintain stability in subsequent breakpoint location index selection and breakpoint resumption execution.

[0063] After determining the current interruption type, step S4 also needs to obtain the breakpoint location identifier corresponding to that interruption type. The breakpoint location identifier is used to indicate the logical disconnection location of the short-range communication link under the current interruption type, providing a reference for subsequently selecting a specific target breakpoint location index among the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier. In one embodiment, the system can maintain three types of breakpoint location identifiers for each short-range communication link, corresponding to the breakpoint location identifiers of physical link-level breakpoints, protocol session-level breakpoints, and application transaction-level breakpoints, respectively. The physical link-level breakpoint location identifier can be composed of the most recently successfully confirmed physical layer frame sequence number or a reserved physical layer frame breakpoint location index; the protocol session-level breakpoint location identifier can be composed of the most recently successfully confirmed protocol packet sequence number or a reserved protocol packet breakpoint location index; and the application transaction-level breakpoint location identifier can be composed of the most recently successfully confirmed application transaction identifier or a reserved application transaction breakpoint location index.

[0064] In a preferred embodiment, when the three-part judgment index Exceeding the preset threshold Furthermore, when the current interruption type is determined to be a physical link interruption, step S4 uses the physical link-level breakpoint location identifier corresponding to the current short-range communication link as a candidate breakpoint location identifier in the subsequent breakpoint resumption process; when the current interruption type is determined to be a protocol session interruption, the protocol session-level breakpoint location identifier is used as a candidate breakpoint location identifier; when the current interruption type is determined to be an application transaction interruption, the application transaction-level breakpoint location identifier is used as a candidate breakpoint location identifier. In subsequent steps, when the short-range communication link is actually interrupted and enters the breakpoint resumption stage, the final target breakpoint location index will be further determined based on the candidate breakpoint location identifiers and the reserved breakpoint location index.

[0065] like Figure 4 As shown, Figure 4 The horizontal axis represents time, and the vertical axis represents the confidence level. The legend shows "Physical Layer Interruption Confidence Level". The curve represents the trend of physical layer interruption confidence over time. The legend shows "Protocol Session Interruption Confidence". The curve represents the trend of protocol session interruption confidence over time, while the legend shows "application transaction interruption confidence". The curve represents the trend of application transaction interruption confidence over time, with the "preset threshold" in the legend. "The dashed line represents the numerical level of the preset threshold."

[0066] In this embodiment, step S5 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on a multi-source data fusion algorithm, step S5 is used to maintain the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier for each short-range communication link during normal transmission. The physical layer frame sequence number is used as the breakpoint location index for the physical link-level breakpoint, the protocol packet sequence number as the breakpoint location index for the protocol session-level breakpoint, and the application transaction identifier as the breakpoint location index for the application transaction-level breakpoint. Simultaneously, when the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier reach a preset interval, the corresponding breakpoint location index is automatically recorded, thereby providing a basis for accurately locating the starting position of the data unit to be resumed under different interruption types.

[0067] In one embodiment, for each established short-range communication link, the sending end maintains three types of incremental sequence numbers or identifiers at the physical layer, link layer, and application layer: physical layer frame sequence number, protocol packet sequence number, and application transaction identifier. The physical layer frame sequence number identifies each physical layer data unit transmitted at the physical layer. Typically, a frame sequence number field already exists in the physical layer or link layer protocol, and the system can directly use this field as the physical layer frame sequence number or extend it. The protocol packet sequence number identifies protocol packet units transmitted in the link layer or transport layer protocol. It can correspond to the link layer data frame number, transport layer sequence number, or a logical packet sequence number encapsulated at a higher layer. The application transaction identifier identifies a business transaction unit on the application layer side, such as a complete file transfer, a batch of data packets, or a specific transaction within a higher-level business session. In actual implementation, it can use a globally incremental transaction number, a combined identifier consisting of a session identifier and a transaction sequence number, or other unique identifiers.

[0068] After the short-range communication link is established, the transmitting end sets initial values ​​for these three types of sequence numbers and increments the corresponding sequence numbers each time a corresponding layer data unit is sent. For example, when the physical layer sends each frame of physical layer data unit, the physical layer frame sequence number is incremented by 1; when the protocol layer sends each new protocol packet unit, the protocol packet sequence number is incremented by 1; and when the application layer completes the packaging of business data for each application transaction and begins transmission, the application transaction identifier is incremented by 1. In this way, the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier each form three monotonically increasing sequences over time, which together constitute the logical position indexing system of the short-range communication link at different layers.

[0069] To select the appropriate breakpoint location index when physical link interruption, protocol session interruption, or application transaction interruption occurs in short-range communication links, and to avoid temporarily backtracking to find the sequence number position after interruption detection, step S5 introduces a preset interval breakpoint recording mechanism during normal transmission of the short-range communication link. In a preferred embodiment, the system sets corresponding preset interval parameters for physical layer frame sequence numbers, protocol packet sequence numbers, and application transaction identifiers, respectively denoted as physical layer breakpoint recording interval, protocol layer breakpoint recording interval, and application layer breakpoint recording interval. When the physical layer frame sequence number increases by the number of times relative to the previous physical layer breakpoint location index, reaching the physical layer breakpoint recording interval, the system automatically records the current physical layer frame sequence number as the new physical link-level breakpoint location index; when the protocol packet sequence number increases by the number of times relative to the previous protocol session-level breakpoint location index, reaching the protocol layer breakpoint recording interval, the system automatically records the current protocol packet sequence number as the new protocol session-level breakpoint location index; when the application transaction identifier increases from the previous application transaction-level breakpoint location index to the next preset transaction count position, the system automatically records the current application transaction identifier as the new application transaction-level breakpoint location index.

[0070] In one specific embodiment, the physical layer breakpoint recording interval can be set to record a physical link-level breakpoint location index once every 50 physical layer data units sent; the protocol layer breakpoint recording interval can be set to record a protocol session-level breakpoint location index once every 10 protocol packet units sent; and the application layer breakpoint recording interval can be set to record an application transaction-level breakpoint location index once after the completion of one application transaction. For continuous transmission services with large data volumes, the granularity of breakpoint recording can also be dynamically adjusted according to the size of the application layer service. For example, for large application transactions, the corresponding sub-breakpoint location indexes within the application transaction can be recorded every time it passes through several key data blocks, so that the range of data units to be resumed can be more precisely trimmed when the application layer is interrupted.

[0071] In scenarios where multiple short-range communication links operate concurrently, to ensure that breakpoint location indexes on different short-range communication links do not interfere with each other, the system can divide the breakpoint location index cache according to the short-range communication link identifier. For each short-range communication link, a separate physical link-level breakpoint location index list, a protocol session-level breakpoint location index list, and an application transaction-level breakpoint location index list are maintained. In this way, when a short-range communication link is interrupted, only the breakpoint location index at the corresponding level needs to be read from the breakpoint location index cache corresponding to that short-range communication link, without affecting the breakpoint resumption process of other short-range communication links.

[0072] like Figure 5 As shown, Figure 5The horizontal axis represents the timeline, and the vertical axis is divided into three different hierarchical lines, representing the physical layer frame sequence number layer, the protocol packet sequence number layer, and the application transaction identifier layer, respectively. On the physical layer frame sequence number layer, a horizontal line is drawn to indicate the increment of the physical layer frame sequence number, and several physical link-level breakpoint location indices are marked with dots at the positions where the physical layer breakpoint recording interval is met. On the protocol packet sequence number layer, a horizontal line is drawn to indicate the increment of the protocol packet sequence number, and several protocol session-level breakpoint location indices are marked with squares at the positions where the protocol layer breakpoint recording interval is met. On the application transaction identifier layer, a horizontal line is drawn to indicate the increment of the application transaction identifier, and several application transaction-level breakpoint location indices are marked with triangles at the positions where each application transaction is completed or the application layer breakpoint recording interval is met.

[0073] In this embodiment, step S6 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on multi-source data fusion algorithm, step S6 is used to set a preset threshold... Set warning threshold below and in the three-stage judgment indicators The preset threshold has not yet been reached. However, it is already close to the preset threshold. The system records the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier in advance, corresponding to the reserved breakpoint location index. This allows the system to select a breakpoint location index closer to the interruption time when an actual interruption occurs in a short-distance communication link, thereby reducing the amount of retransmitted data during the breakpoint resumption process.

[0074] In one embodiment, the warning threshold The value is less than the preset threshold And by using a preset threshold The proportional relationship between them is determined. Specifically, a warning factor between 0 and 1 can be set. The warning threshold is calculated according to the following relationship. : Among them, early warning factors The closer the value is to 1, the higher the warning threshold. The closer to the preset threshold The narrower the warning interval, the stronger the warning factor. The smaller the value, the lower the warning threshold. With preset threshold The greater the difference between the two, the wider the warning interval. In scenarios that are sensitive to interruptions and require earlier pre-marking of breakpoints, a smaller warning factor can be selected. For example, 0.6 or 0.7; in scenarios where it is desirable to reduce the number of index records at reserved breakpoint positions and reduce system overhead, a larger warning factor can be selected. For example, 0.8 or 0.9.

[0075] At the warning threshold After the calculation is completed, for each preset sliding time window, the system has already calculated the physical layer interruption confidence level in step S2. Protocol session interruption confidence level Application transaction interruption confidence level In step S4, a three-part judgment index is constructed. In step S6, the system needs to determine the three-part criteria. Preset threshold and warning threshold The relationship between the three segments determines whether the warning zone has been entered. This is based on the current preset sliding time window's three-part judgment criteria. Less than the warning threshold If the current short-range communication link is still within the normal fluctuation range, no record of the reserved breakpoint location index will be made; only the current three-break judgment index will be recorded. And related information are included in historical data; if the current preset sliding time window's three-part judgment index Greater than or equal to the preset threshold If the interruption type is determined directly according to step S4, preparations are made for the subsequent breakpoint resume process; when the three interruption determination indicators of the current preset sliding time window are... Meets the warning threshold And less than the preset threshold When the conditions are met, it indicates that the short-range communication link has entered a high-risk state but has not yet reached the level to trigger an interruption judgment. At this time, it enters the warning interval and needs to record the reserved breakpoint location index according to the current physical layer frame sequence number, protocol packet sequence number and application transaction identifier.

[0076] In a preferred embodiment, when a three-part judgment index is detected... When a transmission falls within the warning range, the system marks the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier for the next moment, and uses these numbers as the physical link-level reserved breakpoint location index, the protocol session-level reserved breakpoint location index, and the application transaction-level reserved breakpoint location index, respectively. This is because the data unit corresponding to the current physical layer frame sequence number, protocol packet sequence number, and application transaction identifier may already be in the transmission path, while the data unit corresponding to the next moment is more likely to be completely cached in the buffer queues of the sending and receiving ends, making it more suitable as the starting boundary for subsequent breakpoint resumption. By recording the "next moment's physical layer frame sequence number, protocol packet sequence number, and application transaction identifier" as reserved breakpoint location indices, a set of breakpoint location indices closer to the time of the interruption can be prepared in advance for potential interruptions without interrupting the current transmission process.

[0077] To avoid excessive records of reserved breakpoint location indexes due to frequent entry and exit from the warning interval, an improved embodiment can add a time interval limit to the warning recording operation. For example, a cooldown timer can be started after each recording of the reserved breakpoint location index. Before the cooldown timer expires, even if the three-breakpoint judgment index is within a subsequent set of preset sliding time windows... If the system remains in the warning zone, it will not repeatedly record new reserved breakpoint location indices; only after the cooldown timer expires will the system allow the recording of reserved breakpoint location indices again within a new warning zone. The length of the cooldown timer can be set to the sum of the lengths of several preset sliding time windows, based on the characteristics of short-range communication services.

[0078] In another implementation, the system can determine the three-part criteria. The trend of change within the warning interval is analyzed to determine whether to record the reserved breakpoint location index. Specifically, the three-break judgment index can be analyzed within several consecutive preset sliding time windows. Monitor the values ​​of the three-section judgment indicators. If the short-range communication link shows a continuous upward trend within the warning interval, such as monotonically increasing or significantly increasing in average value within several consecutive preset sliding time windows, it is considered that the short-range communication link is developing towards interruption, and it is suitable to record the index of the reserved breakpoint location; if the three-break judgment index If the price only briefly enters the warning zone and then quickly falls back, the reserved breakpoint index can be left unrecorded, thus avoiding excessive sensitivity to short-term fluctuations.

[0079] When multiple short-range communication links are operating concurrently, the warning threshold... The determination of warning intervals also needs to be performed separately for each short-range communication link. The system can maintain its current preset threshold for each short-range communication link individually. Warning threshold Three-part judgment indicators And the set of reserved breakpoint location indices corresponding to this short-range communication link. When the three-break judgment index of a certain short-range communication link... When a signal falls within the warning range, the reserved breakpoint location index is recorded only in the breakpoint location index cache of that short-range communication link, without affecting the breakpoint index recording strategy of other short-range communication links, thereby ensuring the isolation and scalability of the system in multi-link concurrent scenarios.

[0080] like Figure 6 As shown, Figure 6 The horizontal axis represents time, and the vertical axis represents the confidence level. The legend shows the "three-part judgment index". The curve represents the trend of the three-section judgment index over time. The "preset threshold" in the legend... "The dashed line represents the preset threshold value level, and the 'warning threshold' in the legend..." "The dotted line represents the numerical level of the warning threshold, and the 'Reserved Breakpoint Location Index Record Point' marker in the legend indicates the three-breakpoint judgment index." When the system falls into the warning zone, it records the time and position of the reserved breakpoint location index.

[0081] In this embodiment, step S7 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on multi-source data fusion algorithm, step S7 is used to select among the multi-level breakpoint location indexes when an actual interruption of the short-range communication link is detected, in combination with the interruption type determined in step S4 and the breakpoint location index and reserved breakpoint location index maintained in steps S5 and S6, to determine the final target breakpoint location index, and to determine the starting boundary of the data unit to be resumed based on the target breakpoint location index.

[0082] In one embodiment, when the system detects an actual interruption event on a short-range communication link, it first obtains the current interruption type based on the three-stage interruption determination result in step S4. The current interruption type can be one of physical link interruption, protocol session interruption, or application transaction interruption. Based on the current interruption type, the system determines the corresponding breakpoint level: when the current interruption type is physical link interruption, the breakpoint location index corresponding to the physical layer is selected as the candidate breakpoint location index; when the current interruption type is protocol session interruption, the breakpoint location index corresponding to the protocol packet layer is selected as the candidate breakpoint location index; when the current interruption type is application transaction interruption, the breakpoint location index corresponding to the application layer is selected as the candidate breakpoint location index.

[0083] When a short-range communication link is in normal transmission mode, the sending end maintains a recent normal transmission acknowledgment location index for each short-range communication link. This index indicates the logical position of the last data unit successfully received and acknowledged by the peer at the corresponding layer, such as the sequence number of the most recently acknowledged physical layer frame, the sequence number of the most recently acknowledged protocol packet, or the identifier of the most recently acknowledged application transaction. In a preferred embodiment, the system updates the recent normal transmission acknowledgment location index by receiving acknowledgment messages from the peer (e.g., link layer acknowledgment frames, transport layer acknowledgment messages, or application layer acknowledgment messages) and stores it in the corresponding layer's acknowledgment location index cache.

[0084] When a short-range communication link interruption is detected, the system needs to compare the most recent normal transmission confirmation location index with each of the reserved breakpoint location indices to select the most suitable target breakpoint location index. In one embodiment, for the level corresponding to the current interruption type, the system first obtains the value of the most recent normal transmission confirmation location index and searches the reserved breakpoint location index list for all reserved breakpoint location indices that are later than the most recent normal transmission confirmation location index in time. If there are no reserved breakpoint location indices later than the most recent normal transmission confirmation location index, it means that the warning interval has not been entered or the warning recording operation has not been triggered in the recent period. In this case, the system directly uses the most recent normal transmission confirmation location index as the target breakpoint location index. If there are one or more reserved breakpoint location indices later than the most recent normal transmission confirmation location index, the reserved breakpoint location index that is closest to the interruption time in time and does not exceed the interruption time is selected as the target breakpoint location index to ensure that the target breakpoint location index is as close as possible to the actual interruption location.

[0085] In a preferred embodiment, to maintain consistency across breakpoints at different layers, the system can establish a cross-layer mapping relationship for physical layer frame sequence numbers, protocol packet sequence numbers, and application transaction identifiers. For example, when sending each protocol packet unit, the physical layer frame sequence number range covered by that protocol packet unit can be remembered, and when sending each application transaction unit, the set of protocol packet sequence numbers contained in that application transaction unit can be remembered. Through this cross-layer mapping relationship, when the current interruption type is a protocol session interruption or an application transaction interruption, the corresponding physical layer frame sequence number range can be located while determining the target breakpoint location index at the protocol layer or application layer.

[0086] After determining the target breakpoint location index, the system needs to determine the starting boundary of the data unit to be transmitted based on the target breakpoint location index. For physical link interruption scenarios, when the current interruption type is physical link interruption and the target breakpoint location index is a physical layer frame number, the physical layer frame unit corresponding to the target breakpoint location index can be used as the starting physical layer frame unit of the data unit to be resumed, and this physical layer frame unit and subsequent unconfirmed physical layer frame units are included in the scope of the data unit to be resumed. For protocol session interruption scenarios, when the current interruption type is protocol session interruption and the target breakpoint location index is a protocol packet number, the protocol packet unit corresponding to this protocol packet number can be used as the starting protocol packet unit of the data unit to be resumed, and this protocol packet unit and subsequent unconfirmed protocol packet units are included in the scope of the data unit to be resumed. For application transaction interruption scenarios, when the current interruption type is application transaction interruption and the target breakpoint location index is an application transaction identifier, the application transaction unit corresponding to this application transaction identifier can be used as the starting application transaction unit of the data unit to be resumed, and this application transaction unit and subsequent uncompleted application transaction units are included in the scope of the data unit to be resumed.

[0087] In one specific embodiment, to avoid excessive redundant retransmissions due to selecting the target breakpoint location index too early, or losing data that has not yet been buffered by the peer due to selecting the target breakpoint location index too late, the system can calibrate by combining the peer's status feedback information when determining the starting boundary of the data unit to be resumed. For example, if the peer has already sent some acknowledgment messages to confirm the reception status of certain data units before the interruption occurs, the system can compare the acknowledgment positions contained in these acknowledgment messages with the target breakpoint location index: when the acknowledgment position is ahead of the target breakpoint location index, the starting boundary of the data unit to be resumed can be slightly moved back to reduce unnecessary retransmissions; when the acknowledgment position is behind the target breakpoint location index, the starting boundary can be appropriately moved forward to ensure coverage of data intervals that may not have been fully and reliably received. Simultaneously, by checking the status of the physical layer, protocol layer, and application layer buffers, data units that have been clearly discarded or failed can be further eliminated, reducing invalid retransmissions.

[0088] In scenarios where multiple short-range communication links coexist, the target breakpoint location index selection process in step S7 can be performed independently for each short-range communication link. The system can maintain a separate set of the most recent normal transmission confirmation location indexes, a set of reserved breakpoint location indexes, and a current interruption type identifier for each short-range communication link. When a short-range communication link is interrupted, the target breakpoint location index selection and the determination of the starting boundary of the data unit to be resumed are only performed on that short-range communication link, and the breakpoint resumption control of other short-range communication links is not affected.

[0089] like Figure 7 As shown, Figure 7 The horizontal axis represents the time axis, and the vertical axis represents the logical position. On the time axis, the most recent normal transmission confirmation position index is marked by a square mark, several reserved breakpoint position indices are marked by a circle mark, the interruption time is indicated by a dashed vertical line, and the final selected target breakpoint position index is indicated by a star mark.

[0090] In this embodiment, step S8 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on the multi-source data fusion algorithm, step S8 is used to initiate a resumption request control message by the sending end after the short-range communication link is restored. The receiving end generates a resumption response control message based on the integrity verification field and session identifier field in the resumption request control message. Then, the sending end performs breakpoint resumption according to the starting boundary of the data unit to be resumed determined in step S7. After all data units to be resumed have completed transmission confirmation, the corresponding application layer service session is restored, thereby realizing a unified resumption process under the three interruption types.

[0091] In one embodiment, when the short-range communication link recovers from an interrupted state to an available state, the sending end first organizes and generates a resume request control message based on the target breakpoint location index determined in step S7 and the starting boundary of the data unit to be resumed. The resume request control message includes at least the following information: an interruption type field for identifying the interruption type, a target breakpoint location index field for identifying the breakpoint location, an integrity verification field for verifying the integrity of the data unit to be resumed, and a session identifier field for identifying the service session. The interruption type field indicates whether the current resume is for a physical link interruption, protocol session interruption, or application transaction interruption, allowing the receiving end to select the corresponding buffer level and session management strategy based on the interruption type. The target breakpoint location index field indicates the starting position of the resumed transmission. In a physical link interruption scenario, it can be the physical layer frame sequence number; in a protocol session interruption scenario, it can be the protocol packet sequence number; and in an application transaction interruption scenario, it can be the application transaction identifier. The integrity verification field is used to verify the integrity of the data unit to be resumed or the resume control process, and can take the form of a verification sequence, hash value, or message verification code. The session identifier field is used to identify the application layer business session to which the data unit to be resumed belongs, ensuring that the resumed transmission process of each business session does not interfere with each other when there are multiple concurrent business sessions.

[0092] In a preferred embodiment, the integrity verification field can be calculated by the sender based on the overall characteristics or summary information of the data unit to be resumed when generating the resume request control message. For example, the sender can calculate the checksum, cyclic redundancy check, or hash value of the expected data sequence in the data unit to be resumed according to predetermined rules, and write the calculation result into the integrity verification field of the resume request control message. Alternatively, a partitioned integrity verification method can be used, in which the data unit to be resumed is divided into several data segments according to a predetermined size, a corresponding segment check value is generated for each data segment, and all segment check values ​​are encapsulated into the integrity verification field in sequence. In this way, after receiving the resume request control message, the receiver can perform integrity verification on the subsequently received data units to be resumed based on the verification information contained in the integrity verification field, thereby detecting possible mistransmissions, retransmission errors, or malicious tampering while resuming interrupted transmissions.

[0093] The session identifier field is used to distinguish different application-layer service sessions in multi-service scenarios. When an application-layer service session is established, the sending end assigns a unique session identifier to each session and appends this identifier to all application-layer data units associated with that session. After a short-range communication link is interrupted, the sending end selects the target application-layer service session requiring resume transmission based on the list of active application-layer service sessions before the interruption, and writes the corresponding session identifier field into the resume request control message. Upon receiving the resume request control message, the receiving end can look up the locally maintained application-layer service session context based on the session identifier field, restore the state management of that application-layer service session, and associate subsequently received data units to be resumed with the correct service session. In this way, when multiple application-layer service sessions are running concurrently, the resume transmission process for different service sessions can be executed orderly on the same short-range communication link, avoiding data misdelivery or session confusion.

[0094] When the resume request control message successfully reaches the receiving end after the short-range communication link is restored, the receiving end first parses the various fields in the resume request control message. The receiving end reads the interruption type field to determine whether subsequent processing should be applied to the physical layer buffer, protocol layer buffer, or application layer buffer; reads the target breakpoint location index field to determine where the data receiver should begin receiving the data unit to be resumed; reads the session identifier field to bind this resume operation to the corresponding application layer service session; and reads the integrity verification field to obtain a reference value for subsequent integrity verification. After parsing, the receiving end generates a resume response control message locally based on its current buffer state and session state. The resume response control message may include fields used to confirm resume parameters, such as the interruption type field, the target breakpoint location index field, the session identifier field, and an optional receiver capability description field. The receiving end can specify in the resume response control message whether to accept the target breakpoint location index proposed by the sending end, or when it finds a difference between the receiving position recorded locally by the receiving end and the target breakpoint location index, it can propose an adjustment suggestion for the target breakpoint location index, such as adjusting the breakpoint position forward or backward by a few units, to ensure the consistency of the breakpoint position between the sending end and the receiving end.

[0095] In a preferred embodiment, when generating a resume response control message, the receiving end checks the current session context by combining the integrity verification field and the session identifier field. For example, if the application layer service session indicated by the session identifier field has been actively closed or released due to abnormal reasons at the receiving end, the receiving end can return an error status in the resume response control message and suggest that the sending end re-establish the application layer service session. If the application layer service session indicated by the session identifier field is still active, but the receiving end's locally recorded receiving position is significantly behind the target breakpoint position index provided by the sending end, the receiving end can request the sending end to resume transmission from the earlier breakpoint position in the resume response control message to ensure that no important data segments are missed. Through this bidirectional confirmation process between the resume request control message and the resume response control message, a consistent resume starting point can be quickly reached after the short-range communication link is restored, laying the foundation for the integrity verification of subsequent data units to be resumed.

[0096] After the resume response control message returns from the receiving end and reaches the sending end, the sending end parses the resume response control message and checks the interruption type confirmation field, the target breakpoint location index confirmation field, the session identifier field, and the possible adjustment suggestion field. If the receiving end does not suggest adjusting the target breakpoint location index, the sending end directly performs breakpoint resume transmission according to the starting boundary of the data unit to be resumed determined in step S7. That is, starting from the physical layer frame sequence number, protocol packet sequence number, or application transaction identifier corresponding to the target breakpoint location index, the sending end sequentially sends the data units to be resumed until all data units to be resumed have been confirmed for transmission. If the receiving end suggests adjusting the target breakpoint location index, the sending end can redetermine the starting boundary of the data unit to be resumed based on the adjusted breakpoint location index provided in the resume response control message, and resume transmission from the adjusted position to ensure consistency between the sending and receiving ends on the data boundary.

[0097] During the breakpoint resumption process, the sending end can append sequence number information to each data unit to be resumed, and, if necessary, append segment check values ​​corresponding to the integrity check field. After receiving the data units to be resumed in sequence, the receiving end can perform integrity verification on the received data based on the integrity check field carried in the resume request control message and the check value corresponding to each data segment. If the check result of a certain data segment does not meet expectations, the receiving end can request the sending end to retransmit the data segment through a partial retransmission mechanism, thereby avoiding repeated transmission of the entire data unit to be resumed. After all the data units to be resumed have passed integrity verification and obtained the necessary hierarchical confirmations (such as link layer confirmation, transport layer confirmation, or application layer confirmation), the receiving end can release the temporary buffer and switch the status of the corresponding application layer service session from "interruption recovery" to "normal operation," while simultaneously reporting the service session recovery event upwards at the application layer.

[0098] In one specific embodiment, to verify the impact of the resume request control message and resume response control message mechanism involved in step S8 on the overall system performance, a test scenario containing multiple random interruptions can be constructed. In this test scenario, two breakpoint resume schemes are compared: one scheme does not use dedicated resume request control messages and resume response control messages, but simply retransmits data from the most recent normal transmission confirmation position after the short-range communication link is restored; the other scheme adopts the resume request control message and resume response control message mechanism in this embodiment, selects the starting boundary of the data unit to be resumed according to the target breakpoint position index and the reserved breakpoint position index, and ensures data integrity and session continuity through integrity verification fields and session identifier fields.

[0099] In scenarios with multiple concurrent short-range communication links, resume request control messages and resume response control messages can be associated with their respective short-range communication link identifiers and session identifiers. When processing resume requests, the sending and receiving ends isolate the resume process of each link and each service session using the link identifier and session identifier, ensuring that the resume processes of different links and different sessions do not interfere with each other. For multiple application layer service sessions on the same short-range communication link, multiple resume request control messages can be queued and scheduled according to the different session identifier fields, realizing the control of different service priorities.

[0100] In this embodiment, step S9 will be described in detail. In the implementation of the three-segment short-range communication breakpoint resumption method based on a multi-source data fusion algorithm, step S9 is used to, after completing a full breakpoint resumption process, calculate the physical layer interruption confidence level based on the actual interruption situation of this breakpoint resumption process. Protocol session interruption confidence level Application transaction interruption confidence The positive weight coefficients are adaptively updated, thereby gradually correcting the importance of various normalized index variables under different interruption scenarios during long-term operation, and improving the consistency between the three interruption judgment results and the actual interruption type.

[0101] In one embodiment, after the breakpoint resumption process in step S8 is completed, the system has obtained the actual interruption level information for this interruption. The actual interruption level can be determined by comprehensively judging the diagnostic information during the link recovery and service recovery phases. For example, if after the short-range communication link is restored, it is found that the physical layer bit error rate is abnormally high for a long time, the link layer retransmission events are frequent, and after the application layer service session is restored, there are no more large number of request timeouts, and the application layer server is running normally, the actual interruption level can be judged as a physical link interruption; if the physical layer link quality is kept within an acceptable range, but the number of protocol session reconstructions increases significantly before and after the interruption, there are abnormal drops in the link layer cache, and the application layer requests can be completed normally after the session reconstruction, the actual interruption level can be judged as a protocol session interruption; if the physical layer link quality and protocol session are basically normal, but there are a large number of application layer request timeouts, application data drops, or application service abnormalities during the interruption and after recovery, the actual interruption level can be judged as an application transaction interruption. In this embodiment, the actual interruption level can be represented by the same set of types as the three interruption judgment results, namely, physical link interruption, protocol session interruption, and application transaction interruption.

[0102] In step S4, the system determines the interruption confidence level based on the physical layer. Protocol session interruption confidence level Application transaction interruption confidence Size relationship and preset threshold Due to limitations, the interrupt type for this interruption is determined, i.e., the predicted interrupt type. Step S9 requires comparing the predicted interrupt type with the actual interrupt level obtained through diagnosis. When they match, it indicates that the current positive weighting coefficient setting has good discriminative ability for this type of scenario, and no adjustment or only minor adjustments can be made. When they do not match, it indicates that the current positive weighting coefficient setting has a deviation in the three-interrupt judgment result under this type of scenario, and it is necessary to correct the positive weighting coefficient based on the misjudgment of this interruption.

[0103] Used to calculate physical layer interruption confidence. Protocol session interruption confidence level Application transaction interruption confidence The positive weighting coefficients include those appearing in the physical layer interruption confidence calculation. ,coefficient ,coefficient The coefficients appearing in the calculation of confidence level for protocol session interruption ,coefficient And the coefficients that appear in the calculation of application transaction interruption confidence. ,coefficient These positive weighting coefficients correspond to the normalized signal quality index. Normalized bit error rate (NBER) Normalized link retransmission count metric Normalized session hold time index Normalized connection reconstruction frequency index Normalized application data discard ratio index and normalized application request timeout ratio metric The importance of the positive weight coefficients is determined during the initial deployment phase. These coefficients can be set based on experience or offline sample training results. However, in actual operation, step S9 iteratively adjusts these positive weight coefficients using an online update strategy, gradually adapting them to changes in the current short-range communication environment and business characteristics.

[0104] In a preferred embodiment, a corresponding weight coefficient vector can be introduced for each set of positive weight coefficients, and the corresponding weight coefficient vector can be updated based on the judgment error of the interruption after each breakpoint resume process. For example, for the physical layer interruption confidence... It can make This represents a set of weighted coefficients associated with the confidence level of a physical layer interruption. This set of weighted coefficients may include coefficients... ,coefficient Sum of coefficients Or its normalized form. In the first When the resume transmission process ends, if the predicted interrupt type and the actual interrupt level are inconsistent, an error indicator can be defined. This indicates the direction and magnitude of the deviation between the current judgment result and the actual situation. Error indication quantity. The specific form can be determined according to the system design. For example, discrete values ​​of +1, -1 or 0 can be used to indicate that the current judgment is biased towards a certain interrupt type, while the actual interrupt level belongs to another interrupt type. Alternatively, a real number error metric can be used to characterize the degree of deviation between the prediction confidence and the actual interruption.

[0105] After determining the error indication E, the current weighting coefficients can be corrected using the following update formula: ,in, This indicates the weight coefficient value prior to the current breakpoint resume process. This indicates the updated weight coefficient values. The learning rate coefficient. The value of controls the magnitude of the impact of each update on the weights, and is generally a positive number less than 1. (Learning rate coefficient) A larger learning rate coefficient results in faster weight updates and a stronger adaptability to the latest interrupted samples, but may also lead to greater weight fluctuations in noisy environments; The smaller the value, the smoother the weight update, which is more conducive to maintaining the stability of the system's judgment, but more breakpoint resume samples are needed to significantly correct the weight coefficient.

[0106] Similarly, regarding the confidence level of protocol session interruption. Application transaction interruption confidence Alternatively, corresponding weight coefficient vectors can be defined separately, and the error indicator can be calculated based on the deviation between the predicted interruption type and the actual interruption level after each breakpoint resume process. The corresponding weight coefficient vectors are adjusted according to the same update formula. In one specific embodiment, three sets of weight coefficient vectors and corresponding error indicators can be maintained for the three interruption types respectively. After each breakpoint resume process ends, only the set of weight coefficients that participated in the determination of the maximum confidence in this case is updated to reduce mutual interference between weights. Alternatively, a joint update strategy can be adopted to adjust all weight coefficients that participated in the confidence calculation in this case according to different proportions.

[0107] To ensure that the updated weight coefficients still satisfy the constraint of positive weights, normalization or projection can be performed on the weight coefficients after each update. For example, when a component in the weight coefficient vector becomes negative due to an update operation, this component can be truncated to a small positive value close to 0, while other components are appropriately scaled to ensure that the sum of all weights remains within the expected range. In some scenarios, upper and lower bound constraints can also be added to the weight coefficients to prevent a certain weight component from being amplified too much or shrunk too much during iteration, thereby maintaining the balance among multi-source indicators.

[0108] In scenarios where multiple short-range communication links coexist, the adaptive weight update mechanism in step S9 can choose between two modes: global update or link-by-link update, depending on business needs. In the global update mode, all short-range communication links share a set of positive weight coefficients, and this set of weight coefficients is updated after each breakpoint resume process. This is suitable for scenarios where multiple short-range communication links have similar interference environments and business characteristics. In the link-by-link update mode, each short-range communication link maintains an independent set of weight coefficients, and weight updates are performed only based on the breakpoint resume history of that short-range communication link itself. This is suitable for scenarios where different short-range communication links have significantly different operating environments. In actual deployment, a suitable mode can be selected based on the actual situation of computing and storage resources, or a group management approach can be adopted to group short-range communication links with similar characteristics into the same weight update group.

[0109] In summary, step S9 introduces an error indicator-based approach after each breakpoint resume process. The weight coefficient update mechanism gradually adjusts the physical layer interruption confidence level. Protocol session interruption confidence level and application transaction interruption confidence The positive weighting coefficients in the algorithm enable the multi-source data fusion algorithm to adaptively adapt to the actual interruption characteristics of short-range communication links during long-term operation, improve the accuracy and robustness of the three-break judgment, and provide support for the long-term stable operation of the three-break short-range communication breakpoint resumption method based on the multi-source data fusion algorithm in complex and dynamic environments.

[0110] like Figure 8 As shown, Figure 8 The horizontal axis represents the breakpoint resume stream program number, and the vertical axis represents the weighting coefficient value. The "weighting coefficient" in the legend... The curve represents the confidence level used to calculate physical layer outages. Weighting coefficients As the breakpoint resume streaming procedure number changes, the "weighting coefficient" in the legend... The curve represents the weighting coefficient, indicating the trend of the weighting coefficient. The "weighting coefficient" in the legend... The curve represents the weighting coefficient. The changing trend.

[0111] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for resuming interrupted short-range communication in three-segment communication based on a multi-source data fusion algorithm, characterized in that, Includes the following steps: S1. After the short-range communication link is established, a multi-source communication status data set is obtained for each short-range communication link. The multi-source communication status data set includes at least physical layer link quality data, link layer cache status data, application layer service session status data, short-range communication equipment power supply status data, and working environment parameter data. S2, the multi-source communication state data set is sampled within a preset sliding time window. Time alignment and numerical normalization are performed on the sampled data within each preset sliding time window, mapping the multi-source communication state data set into normalized index variables. , , , , , , And calculate the physical link interruption confidence level according to formulas (1) to (3). Protocol session interruption confidence level and application transaction interruption confidence : （1） （2） （3） in, This is a normalized signal quality index obtained based on the physical layer link quality data. This is the normalized bit error rate metric obtained based on the physical layer link quality data. This is a normalized link retransmission count metric obtained based on the link layer cache state data. This is a normalized session hold time metric obtained based on the application layer service session state data. This is a normalized connection reconstruction count metric obtained based on the application layer service session state data. This is a normalized application data discard ratio metric obtained based on the application layer data cache status. This is a normalized application request timeout ratio metric obtained based on the application layer request processing status. , , , , , , These are positive weighting coefficients; S3, calculate link reliability indicators based on historical multi-source communication status data. and interruption frequency index And calculate the preset threshold according to formula (4). : （4） in, This is a normalized link reliability index obtained based on the historical multi-source communication status data. This is a normalized interruption frequency index obtained based on the number of historical interruptions. , The positive weighting coefficients are... The preset threshold; S4, Calculate the three-section judgment index The three-part judgment index For the The above The above The maximum value in the three-part judgment index Greater than or equal to the preset threshold At that time, the current interruption type is determined to be one of three types of interruption: physical link interruption, protocol session interruption, or application transaction interruption, based on the confidence level of the corresponding maximum value. These three interruption types are collectively referred to as the three interruptions, and the corresponding breakpoint location identifier is obtained. S5. For each short-range communication link, maintain the physical layer frame sequence number, protocol packet sequence number, and application transaction identifier respectively. Use the physical layer frame sequence number as the breakpoint location index of the physical link level breakpoint, the protocol packet sequence number as the breakpoint location index of the protocol session level breakpoint, and the application transaction identifier as the breakpoint location index of the application transaction level breakpoint. During the normal transmission of the short-range communication link, when the physical layer frame sequence number, the protocol packet sequence number, and the application transaction identifier reach a preset interval, automatically record the corresponding breakpoint location index. S6, Set warning threshold The warning threshold Less than the preset threshold When the three-part judgment index Less than the preset threshold And greater than or equal to the warning threshold The physical layer frame sequence number, the protocol packet sequence number, and the application transaction identifier of the next moment are recorded as reserved breakpoint position indexes; S7. When an interruption of the short-range communication link is detected, the breakpoint location index of the corresponding layer is selected as the target breakpoint location index according to the interruption type determined in step S4. If there is a reserved breakpoint location index that is later than the location index of the most recent normal transmission confirmation, the corresponding reserved breakpoint location index is used as the target breakpoint location index. Otherwise, the location index of the most recent normal transmission confirmation is used as the target breakpoint location index. The starting boundary of the data unit to be resumed is determined according to the target breakpoint location index. The data unit to be resumed includes a physical layer frame unit corresponding to the physical link interruption, a protocol packet unit corresponding to the protocol session interruption, and an application transaction unit corresponding to the application transaction interruption. S8, after the short-range communication link is restored, a resume request control message carrying the interruption type and the target breakpoint location index is sent. The resume request control message includes an integrity verification field for verifying the integrity of the data unit to be resumed and a session identifier field for identifying the service session. The receiving end generates a resume response control message based on the integrity verification field and the session identifier field. After receiving the resume response control message, the sending end resumes the data unit to be resumed according to the starting boundary of the data unit to be resumed, and restores the corresponding application layer service session after all the data units to be resumed have completed the transmission confirmation. S9. After completing one breakpoint resume transmission process, the actual interruption level is determined based on the interruption type carried in the resume request control message and the diagnostic results collected after the short-range communication link is restored. When the actual interruption level is inconsistent with the corresponding interruption type, the positive weight coefficient is adjusted according to the preset update rule. , , , , , , The value of at least one of them is selected, and the adjusted positive weighting coefficient is used to recalculate the value in the subsequent breakpoint resume process. The above The above .

2. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, The warning threshold Calculate according to formula (5): （5） in, The preset threshold is a coefficient located between 0 and 1. The preset threshold is calculated in step S3 of claim 1.

3. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, The length of the preset sliding time window is based on the current transmission rate of the short-range communication link. and target sampling delay The length of the preset sliding time window is dynamically adjusted. Satisfying formula (6): （6） in, To preset the length of the sliding time window, For the current transmission rate, The target sampling delay.

4. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, The operating environment parameter data includes temperature and humidity parameters, and the power supply status data includes battery level and power stability parameters. During the numerical normalization process in step S2, the temperature, humidity, battery level, and power stability parameters are mapped as additional factors affecting link reliability and participate in the... The above The above The calculation.

5. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, When determining the starting boundary of the data unit to be resumed based on the target breakpoint location index, for physical link interruption, the data unit to be resumed is limited to a physical layer frame unit with a fixed maximum frame length not exceeding a preset physical frame length threshold; for protocol session interruption, the data unit to be resumed is limited to a complete protocol packet unit; and for application transaction interruption, the data unit to be resumed is limited to a complete application transaction unit.

6. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, The short-range communication link is a short-range wireless communication link, and the short-range wireless communication link uses at least one of the following short-range communication protocols: Bluetooth protocol, Wi-Fi Direct protocol, and ultra-wideband communication protocol.

7. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, When multiple concurrent short-range communication links exist, a set of multi-source communication state data and the normalized index variable are maintained for each short-range communication link. , , , , , , The breakpoint location index and the reserved breakpoint location index, and when any short-range communication link is detected to be interrupted, only the set of data corresponding to the short-range communication link is processed according to steps S2 to S9 of claim 1.

8. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, The integrity verification field is a verification sequence calculated from the data unit to be resumed, and the session identifier field is an application layer business session identifier. At the receiving end, the corresponding data unit to be resumed is written into the receiving buffer only when the verification sequence passes the verification and the application layer business session identifier matches the current application layer business session.

9. The method for resuming interrupted short-range communication based on a multi-source data fusion algorithm according to claim 1, characterized in that, The preset update rules include: When the actual interrupt level is inconsistent with the interrupt type, calculate the error indication. The error indication quantity It is an integer that takes the value 1 or -1; For the weighting coefficient corresponding to the incorrect judgment Update according to formula (7): （7） in, The weighting coefficients before the update. The updated weighting coefficients, To learn the step size parameter, This is the error indication quantity.

10. The method for resuming interrupted short-range communication based on multi-source data fusion algorithm according to claim 1, characterized in that, After completing multiple breakpoint resume processes, the link reliability index is recalculated based on the statistical results of each breakpoint resume process. and the interruption frequency index And update the preset threshold according to formula (4). The statistical results include the number of occurrences for each type of interruption and the number of retransmissions of the data unit to be resumed in each interruption resume process.

Images