Cold-chain logistics monitoring method and system

Abstract

The invention relates to the technical field of logistics monitoring, in particular to a cold-chain logistics monitoring method and system. The method comprises the following steps: calculating a data integrity index value at each moment in a historical time period, namely an accumulated value of a product of an effective data proportion and an importance weight of each sensor; obtaining threshold exceeding degree index values of all moments, and preprocessing data and corresponding set numerical value intervals based on the sensor; obtaining a multi-source consistency index value at each moment according to a redundant sensor data standard deviation, a semantic tag and a data variable quantity at adjacent moments; carrying out weighted summation on the three indexes to obtain a confidence coefficient weight at each moment; constructing an enhanced input vector according to the confidence coefficient weight, the semantic tag and the data of each sensor; based on the enhanced input vector at each moment in the historical time period, the data and the semantic tag preprocessed by each sensor in the future preset time step length are predicted, the cold-chain logistics can be accurately monitored, and the monitoring reliability is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics monitoring, in particular to a cold-chain logistics monitoring method and system. BACKGROUND

[0002] In the cold-chain logistics system, goods such as medicines, vaccines and fresh foods that are highly sensitive to the environment require the key parameters such as temperature and humidity to be stable and controllable throughout the transportation process. The common practice is to install a vehicle terminal on the vehicle to collect real-time vehicle compartment environment data and report it to the cloud platform, and the supervision end (or mobile end) monitors and checks compliance accordingly. With the industry moving from "seeing data" to "predicting risks and leaving traces for auditing", relying solely on uploading and displaying is no longer sufficient to meet the needs of compliance and risk control: the system needs to achieve data consistency between the vehicle end, the cloud platform and the supervision / mobile end (hereinafter referred to as "the three ends"), process traceability and abnormal chain.

[0003] Existing cold-chain logistics future state prediction mostly takes Markov chain, graph model and statistical single trajectory prediction algorithm as the core. Markov chain relies on the strong assumption that the state is determined only by the previous time, making it difficult to capture the cumulative impact of cross-period operations such as cold machine start-stop and door opening to take goods, and the state discretization process will lose key information such as temperature fluctuations. The fixed topology structure of the graph model is difficult to adapt to dynamic scenarios such as temporary sensor failures and environmental mutations, and the conditional probability calculation is limited by the completeness of prior knowledge. The statistical single trajectory algorithm has a sudden drop in prediction accuracy when facing non-stationary disturbances such as drastic temperature changes during loading and unloading and extreme weather due to the stationarity assumption. With the development of deep learning models, by fusing multi-sensor time series data to construct an input vector, using models such as LSTM can capture long-range dependencies and non-linear features, effectively solving the problems of poor adaptability to dynamic scenarios and weak multi-source information fusion capability of traditional methods, and further improving the timeliness and robustness of prediction.

[0004] However, the way of simply concatenating multi-source data to construct an input vector lacks deep modeling of semantic associations of data, such as the causal relationship between cold machine power changes and temperature fluctuations, and the differences in data features at different periods (such as during transportation and at stops for unloading), which are difficult to fully exploit these prior knowledge through model self-learning, resulting in insufficient information density of the input vector, which may still cause prediction bias under complex working conditions. SUMMARY

[0005] Therefore, the technical problem to be solved by the present application is to overcome the defects of the existing cold-chain logistics monitoring method based on cold-chain logistics future state prediction, which simply concatenates multi-source data to construct an input vector, lacks deep modeling of semantic associations of data, and results in poor prediction accuracy of monitoring parameters at future time.

[0006] To address the aforementioned technical problems, this invention provides a cold chain logistics monitoring method, comprising: The data integrity index value for each moment in the historical time period is obtained by multiplying the percentage of valid data from each sensor in the cold chain compartment with the importance weight of that sensor, and then summing the product results of all sensors. Based on the preprocessed data of each sensor at each moment within a historical time period and the corresponding set value range of that data, the threshold exceedance index value at each moment within the historical time period is obtained. Based on the standard deviation of the pre-processed data of redundant sensors that measure the same physical quantity as each sensor in the historical time period, the semantic label of the cold chain compartment at each time in the historical time period, and the change in the pre-processed data of each sensor at adjacent times in the historical time period, the multi-source consistency index value at each time in the historical time period is obtained. The confidence weight of each moment in the historical time period is obtained by weighting and summing the data integrity index value, threshold exceedance index value, and multi-source consistency index value at each moment. Using the confidence weights at each moment in the historical time period, the semantic labels of the cold chain vehicle compartment, and the preprocessed data from each sensor, an enhanced input vector is constructed for each moment in the historical time period. Based on the enhanced input vectors at various points in the historical time period, the data from each sensor and the semantic labels of the cold chain vehicle compartment are obtained within a preset time step in the future.

[0007] Preferably, the step of obtaining the threshold exceedance index value for each moment in the historical time period based on the preprocessed data of each sensor at each moment within the historical time period and the corresponding set numerical range of that data includes: Based on the preprocessed data of each sensor at each moment within a historical time period and the corresponding set numerical range, the exceedance range of each sensor at each moment within the historical time period is calculated using the following formula: , The exceedance magnitude of each sensor at each moment within the historical time period is normalized to obtain the normalized exceedance magnitude of each sensor at each moment within the historical time period. Based on the importance weight of each sensor and the normalized exceedance magnitude of each sensor at each moment within the historical time period, the threshold exceedance index value at each moment within the historical time period is calculated using the following formula: , in, Within a historical period The threshold at any given time exceeds the degree indicator value. For the first time in the historical period At the [time]th moment The exceedance of each sensor, For the first Each sensor corresponds to the lower limit of a set numerical range. For the first Each sensor corresponds to the upper limit of a set numerical range. For the first time in the historical period At the [time]th moment Data from one sensor, For the first time in the historical period At the [time]th moment Excess amplitude after normalization of each sensor For the first The importance weight of each sensor This represents a time index within a historical period. Indicates the sensor index. This is a collection of sensors for use inside the cold chain vehicle compartment.

[0008] Preferably, the method for obtaining the multi-source consistency index value at each moment in the historical time period based on the standard deviation of the preprocessed data of redundant sensors that measure the same physical quantity as each sensor within the historical time period, the semantic tags of the cold chain compartment at each moment in the historical time period, and the change in the preprocessed data of each sensor at adjacent moments in the historical time period includes: The standard deviation of the preprocessed data of redundant sensors that measure the same physical quantity as each sensor within the historical time period is normalized at each moment to obtain the redundant sensor consistency index value of each sensor within the historical time period. The importance weight of each sensor is weighted and summed with the consistency index value of each sensor in the historical time period to obtain the consistency index value of redundant sensors at each moment in the historical time period. Based on the semantic labels of each moment in the historical time period and its previous moment, determine whether there is an anomaly in the amount of change of the data of each sensor after preprocessing at each moment in the historical time period. If there is, set the cross-modal consistency index value of that moment in the historical time period to 0; if there is no, set the cross-modal consistency index value of that moment in the historical time period to 1. The absolute value of the difference between the preprocessed data of each sensor at adjacent moments within the historical time period is normalized to obtain the temporal smoothing consistency index value of each sensor at each moment within the historical time period. The importance weight of each sensor is weighted and summed with the temporal smoothing consistency index value of each sensor in the historical time period to obtain the temporal smoothing consistency index value at each moment in the historical time period. The weighted summation of the redundant sensor consistency index value, cross-modal consistency index value, and time smoothing consistency index value at each moment within the historical time period yields the multi-source consistency index value at each moment within the historical time period.

[0009] Preferably, the step of obtaining the data from each sensor and the semantic tags of the cold chain vehicle compartment based on the enhanced input vectors at various moments within a historical time period includes: Will arrive The confidence weights at each time point within the time period are scaled using a normalization function to obtain... arrive The scaled confidence weights for each time point within the time period; where... This marks the beginning of a historical period. The total number of moments in the historical time period; Will arrive The scaled confidence weights at each time step within the time period are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... arrive Forget gate confidence correction term and input gate confidence correction term at each time point within the time period; based on arrive The augmented input vector, forget gate confidence correction term, and previous hidden state at each time step within the time period are passed through the forget gate to obtain... arrive The output of the forget gate at each moment within the time period; based on arrive The augmented input vector, input gate confidence correction term, and previous hidden state at each time step within the time period are passed through the input gate to obtain... arrive The input gate output at each time point within the time period; Will arrive The forget gate output, input gate output, candidate cell state, and previous cell state at each time step within the time period are obtained through cell state updates. arrive Cell state at each moment within the time period; Will arrive The augmented input vector, cell state, and previous hidden state at each time step within the time period are passed through the output gate to obtain... arrive The output gate outputs at each time point within the time period; based on arrive The output gate output and cell state at each time point within the time period are calculated. arrive The hidden state at each moment within the time period; where the hidden state at each moment represents the data of each sensor at that moment; by The hidden states at time t and their corresponding semantic labels and confidence weights are used to construct... The boosted input vector at time t; where, , Predetermine the time step for the future; Will The scaled confidence weights corresponding to the hidden states at time t are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... Forget gate confidence correction term and input gate confidence correction term at different times; based on Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at a given time is obtained through the forget gate. The forget gate outputs at any given moment; based on Time-based boosting input vector, input gate confidence correction term, The hidden state at time t is obtained through the input gate. Input gate output at any given time; Will Forget gate output, input gate output, candidate cell state at any given time The cell state at any given time is obtained through cell state updates. Cellular state at any given moment; Will Time-dependent augmented input vector, cell state, The hidden state at time t is obtained through the output gate. The output gate outputs at any given time; based on The output gate output and cell state at each time point are calculated. The hidden state at any given moment; The hidden state at a given moment is the data from each sensor within a future preset time step; Based on the data from each sensor at each moment within a future preset time step, semantic tags are generated for the cold chain compartment at each moment within the future preset time step.

[0010] Preferably, the based Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at a given time is obtained through the forget gate. The forget gate output at time step is given by the following formula: , in, for The output of the forget gate at any moment, It is the Sigmoid activation function. This is the first weight matrix of the forget gate. This is the second weight matrix of the forget gate. For the bias term of the forget gate, for The boosted input vector at time step, for The hidden state at any given moment. for The forgetting gate confidence correction term at any given time. The learnable confidence correction coefficient corresponds to the forget gate. for Confidence weights after time scaling.

[0011] Preferably, the based Time-based boosting input vector, input gate confidence correction term, The hidden state at time 1 is obtained through the input gate. The input gate output at time t is given by the formula: , in, for Input gate output at any time, It is the Sigmoid activation function. This is the first weight matrix of the input gate. This is the second weight matrix of the input gate. For the bias term of the input gate, for The boosted input vector at time step, for The hidden state at any given moment. for Input gate confidence correction term at time step. This is the learnable confidence correction coefficient corresponding to the input gate. for Confidence weights after time scaling.

[0012] Preferably, after calculation arrive After the hidden state at each time point within the time period, The hidden states at time t and their corresponding semantic labels and confidence weights are used to construct... time The enhanced input vector of parallel branches; where, The number of parallel branches, The weight matrices of the forget gate, input gate, and output gate corresponding to each parallel branch are the same; Will The scaled confidence weights corresponding to the hidden states at time t are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... Forget gate confidence correction term and input gate confidence correction term at different times; based on Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at any given moment, through The forget gate with parallel branches yields... time The forget gate outputs a parallel branch; based on Time-based boosting input vector, input gate confidence correction term, The hidden state at any given moment, through The input gate of the parallel branches yields... time Input gate output with parallel branches; Will time Forget gate output, input gate output, candidate cell state, and parallel branch. The cell state at any given time is obtained through cell state updates. time Cell states with parallel branches; Will Time-dependent augmented input vector, cell state, The hidden state at any given moment, through The output gate of the parallel branch obtains time The output gate has multiple parallel branches; based on At any given moment, the output gate of each parallel branch, the cell state, and the calculation are performed. The hidden state of each parallel branch at any given time; based on The hidden state of each parallel branch at any given time is determined, and the probability weight corresponding to each parallel branch is obtained. based on Given the hidden state of each parallel branch at each time step and the probability weight corresponding to that branch, filter each parallel branch and select the branch corresponding to the hidden state at each time step. The hidden state at a given moment serves as the data from each sensor within a future preset time step; Based on the data from each sensor at each moment within a future preset time step, semantic tags are generated for the cold chain compartment at each moment within the future preset time step.

[0013] Preferably, the based The hidden state of each parallel branch at each time step and the probability weight corresponding to that branch are used to filter each parallel branch, including: Each The average of the product of the scaled confidence weight and the probability weight corresponding to the hidden state of each parallel branch at time t is used as the confidence score of each parallel branch. Will The average number of times the data from each sensor exceeds the threshold range corresponding to the hidden state of each parallel branch at any given time is used as the risk event score for each parallel branch. The difference between 1 and the risk event score of each parallel branch is used as the risk assessment index value of each parallel branch. Based on the arrival time to the target region and the maximum acceptable duration for each parallel branch, the efficiency index value for each parallel branch is obtained, using the following formula: , in, For the first The efficiency index value of parallel branches, For the first The arrival time of each parallel branch to the target region. For the first The starting transport time corresponding to each parallel branch. The maximum acceptable duration; Will The ratio of the number of times the hidden state of each parallel branch satisfies regulatory requirements to the total number of regulatory requirements is used as the compliance indicator value for each parallel branch. The confidence score, risk assessment index value, efficiency index value, and compliance index value of each parallel branch are weighted and summed to obtain the comprehensive score of each parallel branch. Based on the comprehensive score of each parallel branch, obtain the pruned branch set; Using dynamic programming, the optimal branch in the pruned branch set is obtained, and the corresponding branch is... The hidden state at a given moment serves as the data from each sensor within a future preset time step; Based on the data from each sensor at each moment within a future preset time step, semantic tags are generated for the cold chain compartment at each moment within the future preset time step.

[0014] Preferably, after obtaining the data from each sensor and the semantic tags of the cold chain vehicle compartment within a future preset time step, the data within the future preset time step... The data from each sensor in each parallel branch, the semantic tags of the cold chain compartment, the comprehensive score of each parallel branch, the confidence score of each parallel branch, the risk assessment index value, the efficiency index value, the compliance index value and their corresponding weights, the pruning parameters, the dynamic programming algorithm parameters, and the optimal branch index are used as the original data. Within the preset time step in the future The data from each sensor in each parallel branch, along with the semantic tags of the cold chain vehicle compartment and the optimal branch index, are hashed. The generated hash value, along with the comprehensive score of each parallel branch within a preset future time step, the confidence score of each parallel branch, the risk assessment index value, the efficiency index value, the compliance index value and their corresponding weights, pruning parameters, and dynamic programming algorithm parameters, are used as leaf nodes. Through a Merkle tree structure, the Merkle root hash of the original data is generated. When the original data is accessed, the Merkle root hash of the original data is used to verify whether the original data has been tampered with.

[0015] The present invention also provides a cold chain logistics monitoring system, comprising: A memory for storing computer programs; a processor for executing the computer programs to implement the steps of the aforementioned cold chain logistics monitoring method.

[0016] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: The cold chain logistics monitoring method and system described in this invention, at the data integrity assessment level, constructs a data integrity index by weighted aggregation of the effective data ratio of each sensor. This quantifies data availability and highlights the contribution of key sensor signals by highlighting the importance weight of each sensor, thereby improving the sensitivity to data loss. At the abnormal risk perception level, a threshold exceedance index is used to achieve refined risk prediction. First, the deviation of each sensor's data from the preset normal range is calculated. After normalization to eliminate differences in the magnitude of different parameters, a weighted sum is performed based on the importance weight of the sensors to construct the threshold exceedance index. This index can dynamically reflect abnormal fluctuations of different parameters, enabling the model to detect potential exceedance trends that have a greater impact on cold chain safety in advance. At the level of multi-source data consistency verification, the standard deviation of redundant sensor data is calculated by measuring the dispersion of multiple sensor data measuring the same physical quantity, which intuitively reflects the data consistency in the spatial dimension. Cross-modal semantic label verification combines the real-time operating condition labels of cold chain vehicles (such as "loading and unloading" and "high-speed transportation") to verify whether the changes in sensor data match the current scenario. The data change at adjacent time moments is evaluated by calculating the temporal fluctuation amplitude of sensor data to assess the smoothness of data in the time dimension. The confidence weight formed by the weighting of the three factors, together with the data of each sensor and the semantic labels of the cold chain compartment, constitutes an enhanced input vector. This input design not only preserves the temporal characteristics of the monitoring data, but also incorporates the data quality assessment results through confidence weights, and supplements the background information of the operating conditions with semantic labels. This enables the prediction model to adaptively focus on high-quality, highly consistent effective signals and actively weaken the interference of abnormal or low-confidence data. Ultimately, it achieves more robust and accurate predictions of the data of each sensor and the semantic labels of the cold chain compartment in the future, which are more in line with the actual transportation conditions, effectively meeting the safety monitoring needs of complex and dynamic scenarios in cold chain logistics.

[0017] To address the problem that existing methods often only output a single path or a single future state in future state prediction, lacking the ability to model multi-trajectory distributions in the complex environment of cold chain transportation, leading to delayed or even ineffective risk warnings, and failing to meet the stringent requirements of cold chain logistics for uninterrupted supply in case of anomalies, this invention improves the LSTM model. It multiplies the scaled confidence weights by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, as confidence correction terms for the forget gate and input gate, and integrates them into the gating calculation process. This design allows the model to dynamically adjust its information processing strategy based on data quality, preserving the impact of information when confidence is high and weakening its contribution when confidence is low, thereby improving robustness to anomalies / missing data. Based on obtaining reliable hidden states, multiple future branches are expanded in parallel, each representing a possible future trajectory. Shared backbone parameters ensure consistency in temporal patterns and physical logic among the branches; for example, different trajectories all conform to the temperature regulation characteristics of the cold chain refrigeration system, avoiding discrete and disordered distribution of multiple trajectories. This also reduces parameter redundancy and lowers the computational cost of the inference stage. The differentiated design of the branch bias terms allows each branch to flexibly cover different future scenarios caused by complex factors such as sudden changes in road conditions, loading and unloading operations, and slight fluctuations in equipment. This breaks the limitations of traditional methods that rely on single-path prediction, enabling comprehensive modeling of multiple possibilities in cold chain transportation. By improving the prediction accuracy of a single branch through data quality perception, and by covering complex scenarios in parallel with multiple branches, it effectively solves the problem of delayed risk warnings and provides reliable technical support for the stringent monitoring requirements of uninterrupted cold chain logistics. Attached Figure Description

[0018] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a system architecture diagram of a cold chain logistics monitoring method according to the present invention.

[0019] Figure 2 This is a schematic diagram of the data acquisition and semantic tag generation unit.

[0020] Figure 3 This is a schematic diagram of a cloud-based future state overlay prediction unit.

[0021] Figure 4 This is a schematic diagram of the decision collapse and compliance certificate generation unit.

[0022] Figure 5 This is a schematic diagram of the endpoint-cloud-endpoint consistency and degradation verification unit.

[0023] Figure 6 This is a flowchart of the monitoring process for vaccine cold chain transportation.

[0024] Figure 7 This is a monitoring flowchart for a joint delivery scenario using cold chain vehicle fleets. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0026] Reference Figure 1 As shown, this embodiment provides a cold chain logistics monitoring method, including: This embodiment collects multi-source data related to the environment and transportation status in real time. After filtering, alignment and fusion processing, it automatically generates semantic labels and confidence weights, and finally outputs the enhanced input vector for each moment in the historical time period. This provides key inputs for subsequent prediction and decision-making models, ensuring that the system can perform environmental monitoring and risk warning in real time and reliably, and providing interpretable inputs for the prediction model.

[0027] The collected multi-source data includes, but is not limited to: temperature ,humidity ,Location ,speed hatch status Data such as temperature and humidity are provided by environmental sensors in the carriage, while position and speed are provided jointly by the GPS module and inertial navigation unit. Door status is collected by reed switches or Hall effect sensors, and mission phase information is issued by the transportation dispatch system (e.g., "loading," "in transit," "unloading," etc.). For the first time in the historical period The temperature of the cold chain compartment at that moment For the first time in the historical period The humidity in the cold chain vehicle compartment at any given moment For the first time in the historical period The location of the cold chain transport vehicle at any given time. For the first time in the historical period The x-coordinate of the cold chain transport vehicle at any given moment. For the first time in the historical period The longitudinal coordinate of the cold chain transport vehicle at any given moment. For the first time in the historical period The speed of cold chain transport vehicles at any given moment For the first time in the historical period The status of the cargo door of a cold chain transport vehicle at any given moment.

[0028] Preprocessing is performed on the data collected by each sensor at each moment within the historical time period to obtain the preprocessed data for each sensor at each moment within the historical time period, including: For analog sensor data such as temperature and humidity, which are easily affected by equipment vibration, Kalman filtering or moving average filtering algorithms are used for smoothing to eliminate sensor noise and instantaneous jitter; for example, short-term jumps in temperature sensor readings. The anomalies can be smoothed out to reveal the true trend of change.

[0029] Abnormal data that exceeds the physically reasonable range (such as temperatures below -50℃ or above 150℃) or missing data are treated as invalid data and filled in using redundant sensors or interpolation methods.

[0030] Based on the preprocessed data from each sensor at each moment within a historical time period, semantic labels are generated for each moment within the historical time period according to the set rules. Semantic labels are used to describe the current transportation state. These labels help subsequent models understand the specific transportation context. The rules for generating semantic labels are usually defined based on the vehicle's operating state and sensor output values. These labels help the system understand the vehicle's current state and provide contextual information for subsequent predictions and decisions. For example: Based on data from the door status sensor (door magnetic switch) and GPS positioning data, the semantic label is "loading" when the door is open (door magnetic switch outputs a high level) and the GPS positioning coordinates fall within the preset warehouse geofence area; based on data from the vehicle speed sensor and door status, the semantic label is "in transit" when the door is closed (door magnetic switch outputs a low level) and the speed is greater than 5 km / h; based on data from the temperature sensor and regulatory thresholds, the semantic label is "overheating" when the temperature exceeds the regulatory threshold (e.g., 8℃, duration is configurable, default 10 minutes); based on the spatial distance between GPS positioning data and the preset transportation route, the semantic label is "deviation" when the vehicle deviates from the planned route by more than 2 kilometers.

[0031] Based on the semantic labels corresponding to each moment within a historical time period and the preprocessed data, a confidence weight is generated for each moment within the current sliding event. The confidence score comprehensively considers data integrity, threshold exceedance, and multi-source consistency. The specific scheme is as follows: Step S1: Multiply the percentage of valid data (the ratio of valid data to the total data) of each sensor in the cold chain compartment within the historical time period by the importance weight of that sensor. Then, sum the products for all sensors to obtain the data integrity index value for each moment within the historical time period. The data integrity index measures the completeness and validity of data collection and is defined as the historical time period. ( In this embodiment, The percentage of valid data within 10 seconds is calculated using the following formula: , in, Within a historical period Data integrity index value at any given time. Indicates time, This is a collection of sensors for the cold chain vehicle compartment. For the first time in the historical period Total number of valid data from each sensor For the first time in the historical period Total number of data from each sensor For the first time in the historical period The percentage of effective data from each sensor For the first The importance weight of each sensor Indicates the sensor index. This refers to a historical period.

[0032] For example, if the effective data percentage of the temperature sensor in the cold chain truck compartment during a historical period is 0.95, and the importance weight of the temperature sensor is 0.5; the effective data percentage of the position sensor is 0.85, and the importance weight of the position sensor is 0.3; and the effective data percentage of the humidity sensor is 0.98, and the importance weight of the humidity sensor is 0.2, then... .

[0033] Step S2: Based on the preprocessed data of each sensor at each moment within the historical time period and the corresponding set numerical range, obtain the threshold exceedance index value for each moment within the historical time period. This index measures the degree to which the sensor data at the current moment exceeds the set numerical range. Unlike simply judging whether it exceeds the limit, this embodiment sets the numerical range. Set as a regulatory threshold range, and use quantitative calculations to reflect the compliance of data by exceeding the allowed range, including: Based on the preprocessed data of each sensor at each moment within a historical time period and the corresponding set numerical range, the exceedance range of each sensor at each moment within the historical time period is calculated using the following formula: , Indicates the first [number]th ... The extent to which the regulations are exceeded at any given moment.

[0034] The exceedance magnitude of each sensor at each moment within the historical time period is normalized to obtain the normalized exceedance magnitude of each sensor at each moment within the historical time period, as shown in the formula: , Based on the importance weight of each sensor and the normalized exceedance magnitude of each sensor at each moment within the historical time period, the threshold exceedance index value at each moment within the historical time period is calculated using the following formula: , Since the confidence level should increase with compliance, it will eventually enter... The calculated index is defined as the complementary form of the normalized excess amplitude, when hour, ,when hour, .

[0035] in, For the first time in the historical period The threshold at any given moment exceeds the degree indicator value. Within a historical period Time of the first The exceedance of each sensor, For the first Each sensor corresponds to the lower limit of a set numerical range. For the first Each sensor corresponds to the upper limit of a set numerical range. For the first time in the historical period At the [time]th moment Data from one sensor, This is a truncation function used to restrict the output to the range [0,1]. For the first The maximum allowable excess for each sensor For the first time in the historical period At the [time]th moment Excess amplitude after normalization of each sensor For the first The importance weight of each sensor Indicates an index representing a moment within a historical time period. Indicates the sensor index. This is a collection of sensors for use inside the cold chain vehicle compartment.

[0036] Step S3: Based on the standard deviation of the preprocessed data of redundant sensors that measure the same physical quantity with each sensor within the historical time period, the semantic tags of the cold chain compartment at each time point within the historical time period, and the changes in the preprocessed data of each sensor at adjacent times within the historical time period, obtain the multi-source consistency index value at each time point within the historical time period, including: The standard deviation of the preprocessed data from redundant sensors measuring the same physical quantity as each sensor within the historical time period is normalized at each moment to obtain the redundancy sensor consistency index value for each sensor within the historical time period. The formula is as follows: , The redundancy sensor consistency index value at each moment in the historical time period is obtained by weighting and summing the importance weight of each sensor with the consistency index value of each sensor in the historical time period. The formula is as follows: , Based on the semantic labels of each moment within a historical time period and its previous moment, it is determined whether there are any anomalies in the changes of the preprocessed data of each sensor at each moment within the historical time period. If so, the cross-modal consistency index value at that moment within the historical time period is set accordingly. If it is 0, then set the cross-modal consistency index value at that moment within the historical time period to 0. =1; By establishing a mapping rule base for semantic label transformation and expected data change, that is, based on the characteristics of cold chain transportation scenarios, pre-setting reasonable change trends of sensor data under different semantic label switching scenarios, and performing consistency verification of sensor data change at each moment based on the mapping rule base, if the change of all sensors conforms to the expected trend under the current semantic label switching scenario, then cross-modal consistency is determined to be normal; if the change of any key sensor deviates from the expected trend, then cross-modal consistency is determined to be abnormal. The value is 0; for example, after the hatch is opened, the temperature at the current moment should increase compared to the previous moment. If the data does not show this trend, it is recorded as 0; if it meets the expectation, it is recorded as 1. The absolute value of the difference between the preprocessed data of each sensor at adjacent moments within a historical time period is normalized to obtain the temporal smoothing consistency index value of each sensor at each moment within the historical time period. The formula is as follows: , The time smoothing consistency index value at each moment within the historical time period is obtained by weighting and summing the importance weight of each sensor with the time smoothing consistency index value of each sensor within the historical time period. The formula is as follows: , The multi-source consistency index value at each moment in the historical time period is obtained by weighted summing of the redundant sensor consistency index value, cross-modal consistency index value, and time-smooth consistency index value. The calculation formula is: , The weights for the redundancy sensor consistency index, As the weight of the cross-modal consistency index, The weights for the time smoothing consistency index.

[0037] Multi-source consistency metrics measure the consistency between data from different sensors or different time dimensions, and can reflect the overall reliability of the system's data.

[0038] in, For the first time in the historical period At the [time]th moment Redundancy sensor consistency index value of each sensor. In order to be with the first The first sensor measures the same physical quantity. The redundant sensor in the historical time period Preprocessed data at each time point This is a truncation function used to restrict the output to the range [0,1]. , In order to be with the first Two sensors measure the same physical quantity A redundant sensor within a historical time period The average value of the preprocessed data at each time step. For the first The maximum permissible variance of each sensor In order to be with the first The total number of redundant sensors that measure the same physical quantity. For the first time period in history The redundancy sensor consistency index value at each moment. For the first time in the historical period At the [time]th moment Data from one sensor, For the first time in the historical period At the [time]th moment Data from one sensor, For absolute values, For the first The maximum allowable range of change for each sensor For the first time in the historical period At the [time]th moment The time smoothing consistency index value of each sensor For the first time in the historical period The time smoothing consistency index value at each time point For the first The importance weight of each sensor Indicates the time index of a historical period. Indicates the sensor index. This is a collection of sensors for use inside the cold chain vehicle compartment.

[0039] Step S4: Weight the data integrity index value, threshold exceedance index value, and multi-source consistency index value for each moment within the historical time period to obtain the confidence weight for each moment within the historical time period. The formula is: , To ensure that the three indicators can be weighted in the calculation, all indicators must be normalized to [0,1] before being included in the confidence score calculation, where 1 indicates greater confidence. For the first time in the historical period The confidence weight at each time point.

[0040] Step S5: Using the data from each sensor at each moment within the historical time period, along with the semantic labels of the cold chain vehicle compartment, confidence weights, and the semantic labels of the cold chain vehicle compartment, construct the elements of the enhanced input vector for each moment within the historical time period; where, the elements of the enhanced input vector for each moment within the historical time period are... The augmented input vector at each time step , This vector represents the data from each sensor. It not only contains raw numerical information but also carries semantic labels and confidence weights. It can describe the objective state of the cold chain environment and reflect its reliability, thus providing higher interpretability and robustness for subsequent prediction and decision-making.

[0041] Step S6: Based on the enhanced input vectors at each moment in the historical time period, obtain the data of each sensor and the semantic labels of the cold chain compartment within the future preset time step.

[0042] In this embodiment, optionally, the enhanced input vectors at each moment in the historical time period are passed through an LSTM model to obtain the data of each sensor and the semantic labels of the cold chain compartment within a future preset time step.

[0043] In this embodiment, preferably, based on the enhanced input vectors at each moment within a historical time period, the data from each sensor and the semantic labels of the cold chain vehicle compartment within a future preset time step are obtained by improving the LSTM model (Long Short-Term Memory with Time-Superposition and Tagging, LSTM-TT), including: Historical time period ( arrive The confidence weight at each moment within a given time period is scaled using a normalization function to obtain the scaled confidence weight at each moment within the historical time period. The calculation formula is as follows: , in, For the first time in the historical period Confidence weights scaled up at each time point This marks the beginning of a historical period. The total number of moments in the historical time period; In specific applications, the Sigmoid function or power transformation can also be used. Calibration is performed to enhance sensitivity to boundary conditions.

[0044] Will arrive The scaled confidence weights at each time step within the time period are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... arrive Forget gate confidence correction term and input gate confidence correction term at each time point within the time period; based on arrive The augmented input vector, forget gate confidence correction term, and previous hidden state at each time step within the time period are passed through the forget gate to obtain... arrive The output of the forget gate at each moment within the time period; The LSTM-TT model possesses temporal memory properties; its hidden state and cell state at each time step are derived recursively from the state at the previous time step. The initial hidden state (e.g., at the start of the task) is... and cell state Typically initialized as a zero vector or trainable parameters. The hidden state at each recursive time step comes from the model's calculation of the augmented input vector from the previous time step. It is a dynamic encoding of historical information by the model, carrying the environmental state, semantic information, and confidence features from previous time steps, providing temporal context for the prediction at the current time step.

[0045] based on arrive The augmented input vector, input gate confidence correction term, and previous hidden state at each time step within the time period are passed through the input gate to obtain... arrive The input gate output at each time point within the time period; Will arrive The forget gate output, input gate output, candidate cell state, and previous cell state at each time step within the time period are obtained through cell state updates. arrive Cell state at each moment within the time period; Will arrive The augmented input vector, cell state, and previous hidden state at each time step within the time period are passed through the output gate to obtain... arrive The output gate outputs at each time point within the time period; based on arrive The output gate output and cell state at each time point within the time period are calculated. arrive The hidden state at each moment within the time period; where the hidden state at each moment represents the data of each sensor at that moment; Based on the data from each sensor at each moment within a future preset time step, semantic tags for the cold chain compartment at each moment within the future preset time step are generated. by The hidden states at time t and their corresponding confidence weights are used to construct... The boosted input vector at time t; where, , Predetermine the time step for the future; Will The scaled confidence weights corresponding to the hidden states at time t are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... Forget gate confidence correction term and input gate confidence correction term at different times; based on Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at a given time is obtained through the forget gate. The forget gate output at time step is given by the following formula: , in, for The output of the forget gate at any moment, It is the Sigmoid activation function. This is the first weight matrix of the forget gate. This is the second weight matrix of the forget gate. For the bias term of the forget gate, for The boosted input vector at time step, for The hidden state at any given moment. for The forgetting gate confidence correction term at any given time. The learnable confidence correction coefficient corresponds to the forget gate. for Confidence weights after time scaling.

[0046] based on Time-based boosting input vector, input gate confidence correction term, The hidden state at time t is obtained through the input gate. The input gate output at time t is given by the formula: , in, for Input gate output at any time, It is the Sigmoid activation function. This is the first weight matrix of the input gate. This is the second weight matrix of the input gate. For the bias term of the input gate, for The boosted input vector at time step, for The hidden state at any given moment. for Input gate confidence correction term at time step. This is the learnable confidence correction coefficient corresponding to the input gate. for Confidence weights after time scaling.

[0047] Will Forget gate output, input gate output, candidate cell state at any given time The cell state at any given time is obtained through cell state updates. The cell state at time t is given by the formula: , in, for Cellular state at any given moment For Hadamard element-wise multiplication, for The state of candidate cells at any given time. for The state of a cell at any given moment.

[0048] The process of updating the candidate cell state is as follows: Time-based boosting input vector The hidden state at time t is obtained by calculating the candidate cell states. The candidate cell state at time t is given by the formula: ,in, For the tanh function, This is the first weight matrix for the candidate cell states. This is the second weight matrix for the candidate cell states. This is a bias term for the candidate cell state; Will Time-dependent augmented input vector, cell state, The hidden state at time t is obtained through the output gate. The output of the gate at time is calculated using the following formula: ,in, for The output gate outputs at any time. This is the first weight matrix of the output gate. This is the second weight matrix of the output gate. This is the bias term for the output gate.

[0049] based on The output gate output and cell state at each time point are calculated. The hidden state at time t is given by the formula: , The hidden state at each moment is the data of each sensor within the future preset time step. Based on the data of each sensor at each moment within the future preset time step, semantic labels of the cold chain compartment at each moment within the future preset time step are generated by setting preset rules or by using a pre-trained deep learning model. By introducing a confidence correction term, the gating state no longer depends solely on the input and historical hidden state, but also dynamically considers the reliability of the data, making the prediction more robust.

[0050] Existing methods for future state prediction largely rely on Markov chains, graphical models, or statistically based single-trajectory prediction algorithms. These methods often only output a single path or a single future state, lacking the ability to model multi-trajectory distributions in the complex environment of cold chain transportation. When sudden changes occur in the transportation environment (such as sudden deviations or refrigeration unit failures), single-trajectory predictions often cannot cover multiple possibilities, leading to insufficient early warning. However, existing multi-trajectory prediction methods, such as MDN (Mixed Density Network) and MultiPath++, typically have significant shortcomings: First, these methods mostly generate multiple trajectories through probability distribution at the model output layer, and the training phase still focuses on optimizing the prediction error of a single trajectory, failing to achieve parallel learning of multiple trajectories at the underlying architecture level. This leads to the problem of discrete distribution of the generated multiple trajectories—some trajectories are disconnected from the physical laws of the cold chain (e.g., the temperature cannot drop suddenly after a refrigeration unit failure), resulting in low practicality. Second, the inference phase relies on a large amount of sampling (e.g., Gaussian mixture sampling in MDN) or beam search (e.g., path search in MultiPath++) to obtain effective trajectories, which not only increases computational latency (making it difficult to meet the millisecond-level response requirements of real-time cold chain monitoring) but may also miss key risk trajectories (e.g., extreme temperature fluctuation trajectories) due to sampling randomness. Third, the generated multiple trajectories lack strong interpretability related to the cold chain scenario, making it difficult to clarify the working context corresponding to the trajectories. This makes it difficult for operators to quickly determine the priority of trajectories, delaying risk handling.

[0051] Therefore, this application achieves underlying optimization of multi-trajectory prediction by reconstructing the LSTM unit architecture. A Time-Superposition Layer (TSL) is introduced inside the LSTM unit. After obtaining the hidden states within historical time periods, the model is expanded in parallel based on these hidden states. The model employs multiple parallel branches, each representing a possible future evolution path. Each branch shares the backbone parameters, with only the bias terms privatized to capture the multimodal distribution structure. Finally, each branch is normalized using softmax to obtain probability weights, thus forming the probability distribution of the future trajectory. This approach breaks the limitations of traditional models where the output layer generates multiple trajectories. During the training phase, the model forms parallel future branches during temporal feature learning—each parallel branch corresponds to a potential cold chain operating state (e.g., "normal refrigeration branch," "inefficient refrigeration branch," "sudden door opening branch"). Simultaneously, each branch shares gating parameters (weight matrices for the forget gate, input gate, and output gate), avoiding the trajectory logic confusion caused by independent branch training and ensuring consistency in physical characteristics such as "temperature change rate" and "parameter correlation patterns," conforming to the actual operating logic of cold chain transportation. Furthermore, real-time semantic tags for cold chain vehicles (e.g., "high-speed transportation," "stopping and unloading") and confidence weights at each time point are deeply embedded in the branch calculation process, enabling each parallel branch to clearly associate the "scenario-data quality-prediction result" relationship, significantly improving the interpretability of multiple trajectories. This design fundamentally avoids reliance on sampling or beam search during the inference phase, providing operators with more accurate and interpretable decision-making support. The specific solution is as follows: In this embodiment, preferably, after calculation... arrive After the hidden state at each time point within the time period, The hidden states at time t and their corresponding semantic labels and confidence weights are used to construct... time The enhanced input vectors of the parallel branches; the weight matrices of the forget gate, input gate, and output gate corresponding to each parallel branch are the same; in this embodiment, let... ; Will The scaled confidence weights corresponding to the hidden states at time t are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... Forget gate confidence correction term and input gate confidence correction term at different times; based on Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at any given moment, through The forget gate with parallel branches yields... time The forget gate outputs multiple parallel branches; where, The weight matrices of the forget gate, input gate, and output gate corresponding to each parallel branch are the same; based on Time-based boosting input vector, input gate confidence correction term, The hidden state at any given moment, through The input gate of the parallel branches yields... time Input gate output with parallel branches; Will time Forget gate output, input gate output, candidate cell state, and parallel branch. The cell state at any given time is obtained through cell state updates. time Cell states with parallel branches; Will Time-dependent augmented input vector, cell state, The hidden state at any given moment, through The output gate of the parallel branch obtains time The output gate has multiple parallel branches; based on At any given moment, the output gate of each parallel branch, the cell state, and the calculation are performed. The hidden state of each parallel branch at any given time; based on The hidden state of each parallel branch at any given time is determined, and the probability weight corresponding to each parallel branch is obtained. based on Given the hidden state of each parallel branch at each time step and the probability weight corresponding to that branch, filter each parallel branch and select the branch corresponding to the hidden state at each time step. The hidden state at a given moment serves as the data from each sensor within a future preset time step.

[0052] based on Hidden state of each parallel branch at time step Obtain the probability weight corresponding to each parallel branch; parallel branches at each time step. The hidden state itself represents the parallel branch at that moment. The predicted values ​​are the data from each sensor, but they are not directly involved in the softmax calculation; among them, for Time-of-flight parallel branch The hidden state; Will The hidden state of each parallel branch at any given time is determined by calculating the confidence score of each parallel branch using a linear layer or a small network placed at the end of each parallel branch. , For parallel branches The confidence score, , For parallel branches The corresponding trainable parameters; The confidence score of each parallel branch is normalized using softmax to obtain the probability weight corresponding to each parallel branch, as shown in the formula: , , For parallel branch indexes, For parallel branches The confidence score, For parallel branches The corresponding probability weights, meaning the first... The probability of the occurrence of the next future trajectory or the first The reliable weights of the trajectory patterns, the weights satisfying It reflects the confidence distribution of the model across all possible future evolutionary patterns.

[0053] The final multi-track prediction result can be expressed as , , express Time-of-flight parallel branch The vector composed of the corresponding sensor parameters and semantic labels reflects the model's judgment on different future evolution paths, rather than a normalization of the predicted values ​​themselves.

[0054] In this embodiment, specifically, to ensure the accuracy, coverage, diversity, and probabilistic calibration of multi-trajectory prediction, a composite loss function is used during the LSTM-TT model training phase. The composite loss function is as follows: , in, For composite loss function, To account for the error between the predicted trajectory (the predicted data from each sensor and the semantic labels of the cold chain compartment within a future preset time step) and the actual trajectory (the actual data from each sensor and the semantic labels of the cold chain compartment within a future preset time step), To improve trajectory coverage, constraints can be imposed on minimum average deviation (minADE) coverage and top-k coverage. Used to avoid trajectory overlap, such as DTW / Fréchet distance regularization. Used to improve the accuracy of probability distribution calibration, such as desired calibration error ECE or temperature scaling. These are the weights of the loss function.

[0055] In this embodiment, preferably, the basis The hidden state and its corresponding probability weight for each parallel branch at each time step are used to filter each parallel branch, including: Each The average of the products of the scaled confidence weight and the probability weight corresponding to the hidden state of each parallel branch at time step 1 is used as the confidence score of each parallel branch, as shown in the formula: , in, For parallel branches The confidence score, For time index, for Time-of-flight parallel branch The scaled confidence weights corresponding to the hidden states. for Time-of-flight parallel branch The probability weights.

[0056] The confidence score of each parallel branch can comprehensively consider the input confidence and the model output probability, and measure the reliability of the trajectory-dependent data and the prediction.

[0057] Will The average number of times the data from each sensor in the hidden state of each parallel branch exceeds its threshold range is used as the risk event score for each parallel branch. The difference between 1 and the risk event score of each parallel branch is used as the risk assessment index value of each parallel branch. The calculation formula is: , in, For parallel branches Risk assessment indicator values, For indicator functions, for Time-of-flight parallel branch The data from each sensor corresponding to the hidden state. This represents the lower limit of the threshold range corresponding to the data from each sensor. This represents the upper limit of the threshold range corresponding to the data from each sensor.

[0058] For example, if the temperature is within its safe range, then the temperature contributes zero risk at that moment; if it exceeds the threshold range, it is recorded as a risk event. The value range is [0,1]. The smaller the value, the lower the risk, and it can measure the risk of the trajectory violating the cold chain constraints in the future.

[0059] Based on the arrival time to the target region and the maximum acceptable duration for each parallel branch, the efficiency index value for each parallel branch is obtained, using the following formula: , in, For the first The efficiency index value of parallel branches, For the first The arrival time of each parallel branch to the target region. For the first The starting transport time corresponding to each parallel branch. The maximum acceptable duration; The efficiency index value of each parallel branch is used to quantify transportation efficiency, energy consumption, or arrival time. It is calculated by subtracting the ratio of the predicted actual transportation time of each parallel branch to the maximum allowable time calculated from the start time from 1.

[0060] Will The ratio of the number of times the hidden state of each parallel branch satisfies regulatory requirements to the total number of regulatory requirements is used as the compliance indicator value for each parallel branch. The formula is as follows: , in, For parallel branches Compliance indicator values, The total number required by regulations, For parallel branches Number of times regulatory requirements are met The closer it is to 1, the more compliant it is.

[0061] The system determines whether a trajectory complies with drug administration and cold chain food regulations by defining constraints (regulatory requirements). The judgment logic can be modeled based on actual standard clauses or can be freely set and can be completed automatically. For example: If the transport temperature for frozen foods is defined as not exceeding -18℃, and the transport temperature for refrigerated foods is defined as being maintained between 0℃ and 10℃, then the formula for determining this is: If the temperature exceeds the specified range, it is determined that it does not meet the regulatory requirements. Indicates branch Whether specific regulatory requirements are met For parallel branches Corresponding temperature.

[0062] If it is stipulated that transport vehicles must not deviate from the planned route by more than 2 km and must record trajectory information throughout the entire process, then if a transport vehicle deviates by more than 2 km or loses location data, it will be determined that it does not meet the regulatory requirements.

[0063] The comprehensive score for each parallel branch is obtained by weighted summing of its confidence score, risk assessment index value, efficiency index value, and compliance index value. The formula is as follows: , in, For parallel branches Overall score These are the weighting coefficients.

[0064] To ensure that indicators with different dimensions can be weighted and integrated in the scoring function, all indicators must be normalized before entering the scoring function: for indicators that are proportions (such as...), ... , Its output is naturally in the [0,1] interval and does not require additional normalization; for raw value indicators (such as transportation time, energy consumption, etc.), Min-Max normalization is required: , This is the output after Min-Max normalization. For input data, The minimum value of the input data. The maximum value of the input data is used. After normalization, all index values ​​fall within the range of [0,1], ensuring the comparability and stability of the score calculation.

[0065] To reduce redundant trajectories and improve decision-making efficiency, this invention employs a "two-stage strategy." Based on the comprehensive score of each parallel branch, a pruned branch set is obtained, including: removing branches with a comprehensive score lower than a set comprehensive score threshold. The trajectory branches, DTW (Dynamic Time Warping) or Fréchet distance metric, if the similarity is higher than the set similarity threshold (e.g., 0.8), will only retain the one with the highest overall score.

[0066] The dynamic programming (DP) algorithm is used to search for the trajectory branch with the minimum cost and the highest score in the pruned set, and obtain the optimal branch in the pruned branch set. The transfer cost function for each parallel branch The calculation formula is: , in, For parallel branch temperature changes, For the parallel branch speed variation, Penalties for violating cold chain thresholds, These are the weighting coefficients.

[0067] Choose the trajectory branch with the lowest transition cost and the highest overall score. The parallel branch with the lowest cost and highest overall score is selected as the optimal branch. This is the optimal branch index.

[0068] The optimal branch corresponds to The hidden state at each moment serves as the data from each sensor within a future preset time step. Based on the data from each sensor at each moment within the future preset time step, semantic tags for the cold chain compartment at each moment are generated.

[0069] Compared with traditional single-trajectory prediction methods, this model can not only generate multiple possible future evolution paths, but also dynamically adjust the prediction results by combining semantic and confidence information, thereby improving robustness and interpretability.

[0070] In this embodiment, preferably, after obtaining the data of each sensor and the semantic tags of the cold chain vehicle compartment within a future preset time step, the data within the future preset time step are... The data from each sensor in each parallel branch, the semantic tags of the cold chain compartment, the comprehensive score of each parallel branch, the confidence score of each parallel branch, the risk assessment index value, the efficiency index value, the compliance index value and their corresponding weights, the pruning parameters, the dynamic programming algorithm parameters, and the optimal branch index are used as the original data. Within the preset time step in the future The data from each sensor in each parallel branch, along with the semantic tags of the cold chain vehicle compartment and the optimal branch index, are hashed. The generated hash value, along with the comprehensive score of each parallel branch within a preset future time step, the confidence score of each parallel branch, the risk assessment index value, the efficiency index value, the compliance index value and their corresponding weights, pruning parameters, and dynamic programming algorithm parameters, are used as leaf nodes. Through a Merkle tree structure, the Merkle root hash (compliance certificate) of the original data is generated. When the original data is accessed, the Merkle root hash of the original data is used to verify whether the original data has been tampered with.

[0071] , in, for Merkle root hash of the original data at any given time. It is a Merkle tree structure. For hash processing; Compliance credentials can be stored in the cloud or on the blockchain for subsequent regulatory verification, traceability auditing, and consistency verification across the three ends. Credentials contain both the final result and encapsulate the logic of the entire process. They can not only verify whether the optimal result is consistent, but also verify the rationality of the scoring and elimination process of the discarded trajectory through sampling, thereby achieving full transparency and traceability.

[0072] Existing methods like blockchain or log-based evidence storage typically only store the hashes of the final result or key fields, failing to trace the prediction process and leaving regulators unable to understand "why other trajectories were discarded." This application encapsulates the entire future state set, synthesis function, pruning / DP parameters, and optimal trajectory index in a unified manner, using a Merkle tree to generate a root hash, forming a compliance certificate. Regulators can sample and verify discarded trajectories to confirm the rationality of the scoring and removal process.

[0073] like Figure 2 , 3 As shown in Figures 4 and 5, Figure 2 This is a schematic diagram of the data acquisition and semantic tag generation unit. Figure 3 This is a schematic diagram of a cloud-based future state overlay prediction unit. Figure 4 This is a schematic diagram of the decision collapse and compliance certificate generation unit. Figure 5 This is a schematic diagram of the endpoint-cloud-endpoint consistency and degradation verification unit.

[0074] Graceful degradation methods are common in the IoT field, but they often only retain partial data, are not bound to cold chain regulations, and lack credential support. This invention ensures consistency across the three ends based on credential comparison. In the event of verification failure or network anomalies, only the minimum compliant fields required by regulations (temperature, location, timestamp, semantic tag, confidence level, and credential digest) are transmitted, along with an error code. After network recovery, automatic remediation is performed, ensuring idempotency. Even with network interruption, the regulatory end can still obtain the minimum compliant data, preventing audit gaps and ensuring uninterrupted compliance. The specific solution is as follows: In this embodiment, preferably, the data acquisition and semantic tag generation unit set on the vehicle end collects multi-source data such as temperature, humidity, location, speed, and door status. After filtering and alignment processing, semantic tags and confidence weights are automatically generated to form an enhanced input vector for each moment in the historical time period. By using a cloud-based future state prediction unit and an improved LSTM model, we can achieve multi-trajectory future state distribution prediction, thereby improving prediction robustness and interpretability. Through the decision collapse and compliance certificate generation units of the vehicle terminal, cloud terminal, and regulatory terminal, the results of multiple trajectories are scored and optimized, the optimal branch is selected, and the data of each sensor and the semantic tags of the cold chain compartment within the future preset time step are obtained. At the same time, the optimal branch trajectory set, the comprehensive score of each branch and the selection logic are packaged and generated into hash certificate to ensure the traceability and credibility of the whole process. By using end-to-end consistency and degradation verification units at the vehicle, cloud, and regulatory ends, credential comparison is performed between the three ends to achieve data consistency verification. In the event of network anomalies or verification failures, only the minimum compliant fields (temperature, location, timestamp, semantics, confidence level, and credential digest) are transmitted to ensure that compliant data is not interrupted and is automatically replenished after the network is restored to ensure long-term consistency.

[0075] Within each data cycle, the vehicle terminal, cloud terminal, and regulatory terminal all hold compliance certificates. If the compliance certificates of the three terminals are completely consistent, it indicates that the data has not been tampered with and synchronization has been successful. All three terminals display complete results (optimal branch, summary of the entire branch set, and compliance certificate). If the results of the three terminals are inconsistent, or if network indicators are abnormal (such as end-to-end latency > 30 seconds, packet loss rate > 5%, or certificate verification failure), a degradation mechanism is triggered.

[0076] In degraded mode, only the data from key sensors at each moment of the optimal branch, along with the semantic tags (monitoring parameters), compliance certificates, and error reason codes of the cold chain compartment, are transmitted. This ensures that the cold chain transportation process meets the minimum requirements for regulatory review. The data from key sensors, along with the semantic tags, compliance certificates, and error reason codes of the cold chain compartment, are as follows: ,in, As a compliance certificate, Error reasons are coded (01: hash inconsistency; 02: missing data; 03: communication timeout), and transmission of non-critical fields (such as humidity and energy consumption indicators) is paused to reduce bandwidth consumption. This ensures that the regulatory end can still receive the minimum compliance data without interruption even under extreme conditions.

[0077] Once the network is restored, the system will automatically trigger the recovery process: First, missing interval detection is performed, using the sequence number and timestamp of each data entry to locate the missing data during the anomaly period. Next, data completion is performed, and the missing optimal branch and compliance credentials are retransmitted. The compliance credentials are then compared again across all three ends to ensure final consistency. If the same data appears repeatedly due to retransmission, the system avoids duplicate archiving by verifying the sequence number and hash.

[0078] Even during network outages, regulators can still access key fields such as temperature, location, and timestamps, preventing violations due to missing data. Transparency is enhanced; downgraded data is accompanied by error codes and credential summaries, allowing regulators to trace "why the data is incomplete." Once the network is restored, the system automatically completes the data and credentials, preventing permanent gaps. Minimum compliance fields directly correspond to the requirements of the drug regulatory authority and food cold chain standards, ensuring compliance with regulatory reviews even in extreme circumstances.

[0079] Based on Example 1, this Example 2 will be illustrated using examples of vaccine cold chain transportation and joint distribution by a cross-provincial cold chain vehicle fleet: like Figure 6 As shown, Figure 6 This is a flowchart of the monitoring process for vaccine cold chain transportation. The transport vehicles travel from "pharmacy warehouse → distribution center → vaccination site," with strict monitoring requirements for temperature, humidity, and transportation conditions throughout the entire process. The drug regulatory authorities require: The temperature of the transport compartment must be maintained between 2°C and 8°C; Any node must be able to trace back the complete environmental data and decision-making process; Even in the event of a network outage, it is essential to ensure that regulators can obtain the minimum compliant data (temperature, location, timestamp).

[0080] The refrigerated truck's onboard terminal is equipped with temperature and humidity sensors, GPS, door sensors, and a task scheduling interface. The temperature and humidity data acquisition frequency is set to 1Hz, and the GPS positioning data acquisition frequency is 0.1Hz. The data preprocessing stage then employs a Kalman filter algorithm to eliminate jitter errors caused by external interference from the temperature and humidity sensors. Simultaneously, GPS interpolation is used to supplement the positioning data, ensuring data continuity. Based on this, semantic tags are generated, triggering corresponding tags according to multi-dimensional data combinations: a "loading" tag is generated when the door is detected to be open and the vehicle is located in the warehouse area; a "transporting" tag is generated when the door is closed and the vehicle speed is greater than 5km / h; a "temperature exceeding" tag is generated when the temperature exceeds 8℃ for 10 consecutive minutes; and a "deviation" tag is generated when the vehicle's actual driving trajectory deviates from the planned route by more than 2km. Finally, confidence is calculated, considering data integrity, threshold exceedance, and multi-source data consistency. A confidence score of 0.92 indicates a highly reliable semantic tag determination, generating an enhanced input vector.

[0081] The input vectors are uploaded to the cloud to improve the LSTM model, generating four different branches: For example: Trajectory 1: Temperature remains 5–7℃, probability 65%; Trajectory 2: Temperature rises slightly to 8.5℃, probability 20%; Trajectory 3: Temperature suddenly rises to 10℃ (overheating), probability 10%; Trajectory 4: Temperature drops to 4°C, probability 5%.

[0082] In the cloud, different predicted trajectory branches are uniformly scored. First, branches with a comprehensive score below the pruning threshold of 0.3 are removed. Then, redundant trajectories that are too similar to trajectory 1 (65% temperature stability) are filtered out using Fréchet distance. Finally, the optimal branch with the lowest risk and highest compliance is selected using a dynamic programming algorithm. Then, the certificate generation stage begins. First, the summary of the entire set of branch trajectories, scoring function parameters, pruning parameters, DP parameters, and the index of the optimal branch are packaged. Then, a root hash is generated using a Merkle tree. This certificate is stored in the cloud and synchronized to the regulatory end to support subsequent sampling verification.

[0083] During the edge-cloud-edge consistency and degradation verification phase, the three ends simultaneously receive the optimal branch and compliance credentials, independently calculate hashes, and compare them. If they match, the complete result is displayed, including the optimal branch, a summary of the entire branch trajectory, and the compliance credentials. If inconsistencies or anomalies occur (such as network interruption lasting more than 30 seconds or packet loss exceeding 5%), the system automatically switches to degradation mode, transmitting only the minimum compliance fields. In this case, the regulatory end can still view temperature and location information, ensuring uninterrupted compliance processes. Once the network is restored, the system triggers a recovery mechanism to fill in the missing trajectory set and complete credentials. After the three ends re-compare the compliance credentials and confirm consistency, archiving is completed, effectively preventing compliance vulnerabilities. like Figure 7 As shown, Figure 7 This is a monitoring flowchart for a joint cold chain delivery scenario. In a cross-provincial joint cold chain delivery scenario, multiple cold chain vehicles depart from different warehouses and need to transport vaccines, medicines, and fresh food to several distribution centers and hospitals within a limited time. Unlike single-vehicle transportation, the fleet dispatch center requires unified monitoring and coordinated scheduling of the operational status of multiple vehicles to avoid situations where "some vehicles arrive early and wait" or "some vehicles are delayed, causing a break in the overall delivery chain." Simultaneously, the drug regulatory authorities require the ability to conduct unified random checks of data vouchers in multi-vehicle scenarios to ensure that the environmental parameters, predicted trajectories, and compliance decision-making processes of each vehicle are traceable. Even if a vehicle experiences a communication interruption, the regulatory end must receive its minimum compliance fields in real time.

[0084] Each refrigerated truck is equipped with a temperature and humidity sensor, GPS module, door sensor, and task interface to collect real-time information on temperature, humidity, location, speed, door status, and task stage.

[0085] The temperature and humidity data are collected at a frequency of 1 Hz, while the GPS data is collected at a frequency of 0.2 Hz. During the data preprocessing stage, Kalman filtering is used to smooth the temperature and humidity data, and GPS data is supplemented by interpolation and inertial navigation fusion, while ensuring the alignment of timestamps of data from multiple vehicles. Semantic label generation follows specific rules: a "loading" label is generated when the hatch is open and the vehicle is in the warehouse; a "transporting" label is generated when the hatch is closed and the speed is > 5 km / h; a "temperature exceeding 8℃" label is generated when the temperature is > 8℃ for 10 minutes; and a "deviation from the planned path" label is generated when the deviation is more than 2 km. The system also calculates the confidence level for each data point based on data integrity, consistency, and the degree to which the threshold is exceeded, for example, a confidence level of 0.88.

[0086] In the cloud-based future state overlay prediction stage, the enhanced input vectors of all vehicles are uploaded to the cloud, and the improved LSTM model performs future state prediction. Unlike single-vehicle prediction, this model simultaneously expands the future trajectory branches of multiple vehicles in the temporal overlay layer (TSL). Furthermore, a "cooperative constraint loss" is introduced during the training phase to ensure that the prediction results of multiple vehicles maintain temporal consistency at key nodes such as highway exits and distribution center entrances. Taking the collaborative prediction of two vehicles as an example, the trajectory 1 of vehicle A is stable at 6℃ (probability 70%), and the trajectory 2 is rising to 9℃ (probability 20%). The trajectory 1 of vehicle B is stable at 7℃ (probability 65%), and the trajectory 2 is yaw causing a 15-minute arrival delay (probability 25%). When the system outputs the trajectory, it also explicitly marks the vehicle ID and task stage in the branch label, which is convenient for the subsequent decision-making layer to take into account both the situation of individual vehicles and the overall situation of the fleet.

[0087] Entering the decision collapse and compliance certificate generation stage, the cloud first scores the prediction results of multiple vehicles. In the scoring function of each vehicle's trajectory, the arrival time window of multiple vehicles at checkpoints or distribution centers is constrained by the coordination consistency factor of vehicles at key nodes. Then, pruning and optimization are performed, branches with scores below the pruning threshold of 0.25 are removed, and redundant trajectories with similarity > 0.85 are removed by DTW distance, retaining a candidate set with broad coverage. Then, dynamic programming is used to select the overall optimal branch, which improves the overall efficiency of the fleet while ensuring the compliance of individual vehicles. Finally, a compliance certificate is generated, which encapsulates the complete set of multi-vehicle trajectories, scoring function parameters, pruning and DP parameters, and coordination constraint parameters. A Merkle tree is used to generate a root hash. This certificate can prove the optimal branch selection of each vehicle and also encapsulate the constraint logic of multi-vehicle coordination.

[0088] Regarding end-to-cloud-to-end consistency and degradation verification, the dispatch center, cloud platform, and regulatory platform respectively receive the optimal branch and compliance credentials and compare them with the root hash. If the comparison is consistent, the complete result is displayed, including the optimal trajectory of each vehicle, a summary of the complete set of branch trajectories, and decision credentials. If the comparison is inconsistent or a vehicle's link is abnormal, the system enters degradation mode, transmitting only the minimum compliance fields for that vehicle, and the regulatory interface will indicate "minimum compliance guarantee mode". When the network is restored, the system will trigger a recovery mechanism to fill in the missing complete set of multi-vehicle trajectories and complete credentials, and ensure long-term consistency through idempotency verification.

[0089] This third embodiment provides a cold chain logistics monitoring system, including: A memory for storing computer programs; a processor for executing the computer programs to implement the steps of the aforementioned cold chain logistics monitoring method.

[0090] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0092] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0093] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0094] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A cold chain logistics monitoring method, characterized in that, include: The data integrity index value for each moment in the historical time period is obtained by multiplying the percentage of valid data from each sensor in the cold chain compartment with the importance weight of that sensor, and then summing the product results of all sensors. Based on the preprocessed data of each sensor at each moment within a historical time period and the corresponding set value range of that data, the threshold exceedance index value at each moment within the historical time period is obtained. Based on the standard deviation of the pre-processed data of redundant sensors that measure the same physical quantity as each sensor in the historical time period, the semantic label of the cold chain compartment at each time in the historical time period, and the change in the pre-processed data of each sensor at adjacent times in the historical time period, the multi-source consistency index value at each time in the historical time period is obtained. The confidence weight of each moment in the historical time period is obtained by weighting and summing the data integrity index value, threshold exceedance index value, and multi-source consistency index value at each moment. Using the confidence weights at each moment in the historical time period, the semantic labels of the cold chain vehicle compartment, and the preprocessed data from each sensor, an enhanced input vector is constructed for each moment in the historical time period. Based on the enhanced input vectors at various points in the historical time period, the data from each sensor and the semantic labels of the cold chain vehicle compartment are obtained within a preset time step in the future.

2. The cold chain logistics monitoring method according to claim 1, characterized in that, The process of obtaining the threshold exceedance index value for each moment in the historical time period based on the preprocessed data of each sensor at each moment and the corresponding set numerical range of that data includes: Based on the preprocessed data of each sensor at each moment within a historical time period and the corresponding set numerical range, the exceedance range of each sensor at each moment within the historical time period is calculated using the following formula: ， The exceedance magnitude of each sensor at each moment within the historical time period is normalized to obtain the normalized exceedance magnitude of each sensor at each moment within the historical time period. Based on the importance weight of each sensor and the normalized exceedance magnitude of each sensor at each moment within the historical time period, the threshold exceedance index value at each moment within the historical time period is calculated using the following formula: ， in, Within the historical period The threshold at any given time exceeds the degree indicator value. For the first time in the historical period At the [time]th moment The exceedance of each sensor, For the first Each sensor corresponds to the lower limit of a set numerical range. For the first Each sensor corresponds to the upper limit of a set numerical range. For the first time in the historical period At the [time]th moment Data from one sensor, For the first time in the historical period At the [time]th moment Excess amplitude after normalization of each sensor For the first The importance weight of each sensor This represents a time index within a historical period. Indicates the sensor index. This is a collection of sensors for use inside the cold chain vehicle compartment.

3. The cold chain logistics monitoring method according to claim 1, characterized in that, The multi-source consistency index value for each moment in the historical time period is obtained based on the standard deviation of the preprocessed data of redundant sensors that measure the same physical quantity as each sensor within the historical time period, the semantic tags of the cold chain compartment at each moment in the historical time period, and the change in the preprocessed data of each sensor at adjacent moments in the historical time period. This includes: The standard deviation of the preprocessed data of redundant sensors that measure the same physical quantity as each sensor within the historical time period is normalized at each moment to obtain the redundant sensor consistency index value of each sensor within the historical time period. The importance weight of each sensor is weighted and summed with the consistency index value of each sensor in the historical time period to obtain the consistency index value of redundant sensors at each moment in the historical time period. Based on the semantic labels of each moment in the historical time period and its previous moment, determine whether there is an anomaly in the amount of change of the data of each sensor after preprocessing at each moment in the historical time period. If there is, set the cross-modal consistency index value of that moment in the historical time period to 0; if there is no, set the cross-modal consistency index value of that moment in the historical time period to 1. The absolute value of the difference between the preprocessed data of each sensor at adjacent moments within the historical time period is normalized to obtain the temporal smoothing consistency index value of each sensor at each moment within the historical time period. The importance weight of each sensor is weighted and summed with the temporal smoothing consistency index value of each sensor in the historical time period to obtain the temporal smoothing consistency index value at each moment in the historical time period. The weighted summation of the redundant sensor consistency index value, cross-modal consistency index value, and time smoothing consistency index value at each moment within the historical time period yields the multi-source consistency index value at each moment within the historical time period.

4. The cold chain logistics monitoring method according to claim 1, characterized in that, The process of obtaining data from each sensor and semantic labels for the cold chain vehicle compartment based on enhanced input vectors at various moments within a historical time period includes: Will arrive The confidence weights at each time point within the time period are scaled using a normalization function to obtain... arrive The scaled confidence weights for each time point within the time period; where... This marks the beginning of a historical period. The total number of moments in the historical time period; Will arrive The scaled confidence weights at each time step within the time period are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... arrive Forget gate confidence correction term and input gate confidence correction term at each time point within the time period; based on arrive The augmented input vector, forget gate confidence correction term, and previous hidden state at each time step within the time period are passed through the forget gate to obtain... arrive The output of the forget gate at each moment within the time period; based on arrive The augmented input vector, input gate confidence correction term, and previous hidden state at each time step within the time period are passed through the input gate to obtain... arrive The input gate output at each time point within the time period; Will arrive The forget gate output, input gate output, candidate cell state, and previous cell state at each time step within the time period are obtained through cell state updates. arrive Cell state at each moment within the time period; Will arrive The augmented input vector, cell state, and previous hidden state at each time step within the time period are passed through the output gate to obtain... arrive The output gate outputs at each time point within the time period; based on arrive The output gate output and cell state at each time point within the time period are calculated. arrive The hidden state at each moment within the time period; where the hidden state at each moment represents the data of each sensor at that moment; by The hidden states at time t and their corresponding semantic labels and confidence weights are used to construct... The boosted input vector at time t; where, , Predetermine the time step for the future; Will The scaled confidence weights corresponding to the hidden states at time t are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... Forget gate confidence correction term and input gate confidence correction term at different times; based on Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at a given time is obtained through the forget gate. The forget gate outputs at any given moment; based on Time-based boosting input vector, input gate confidence correction term, The hidden state at time t is obtained through the input gate. Input gate output at any given time; Will Forget gate output, input gate output, candidate cell state at any given time The cell state at any given time is obtained through cell state updates. Cellular state at any given moment; Will Time-dependent augmented input vector, cell state, The hidden state at time t is obtained through the output gate. The output gate outputs at any given time; based on The output gate output and cell state at each time point are calculated. The hidden state at any given moment; The hidden state at a given moment is the data from each sensor within a future preset time step; Based on the data from each sensor at each moment within a future preset time step, semantic tags are generated for the cold chain compartment at each moment within the future preset time step.

5. A cold chain logistics monitoring method according to claim 4, characterized in that, The basis Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at a given time is obtained through the forget gate. The forget gate output at time step is given by the following formula: ， in, for The output of the forget gate at any moment, It is the Sigmoid activation function. This is the first weight matrix of the forget gate. This is the second weight matrix of the forget gate. For the bias term of the forget gate, for The boosted input vector at time step, for The hidden state at any given moment. for The forgetting gate confidence correction term at any given time. The learnable confidence correction coefficient corresponds to the forget gate. for Confidence weights after time scaling.

6. A cold chain logistics monitoring method according to claim 4, characterized in that, The basis Time-based boosting input vector, input gate confidence correction term, The hidden state at time 1 is obtained through the input gate. The input gate output at time t is given by the formula: ， in, for Input gate output at any time, It is the Sigmoid activation function. This is the first weight matrix of the input gate. This is the second weight matrix of the input gate. For the bias term of the input gate, for The boosted input vector at time step, for The hidden state at any given moment. for Input gate confidence correction term at time step. This is the learnable confidence correction coefficient corresponding to the input gate. for Confidence weights after time scaling.

7. A cold chain logistics monitoring method according to claim 4, characterized in that, In the calculation arrive After the hidden state at each time point within the time period, The hidden states at time t and their corresponding semantic labels and confidence weights are used to construct... time The enhanced input vector of parallel branches; where, The number of parallel branches, The weight matrices of the forget gate, input gate, and output gate corresponding to each parallel branch are the same; Will The scaled confidence weights corresponding to the hidden states at time t are multiplied by the learnable confidence correction coefficients corresponding to the forget gate and input gate, respectively, and used as... Forget gate confidence correction term and input gate confidence correction term at different times; based on Time-based augmentation of input vector, forget gate confidence correction term, The hidden state at any given moment, through The forget gate with parallel branches yields... time The forget gate outputs a parallel branch; based on Time-based boosting input vector, input gate confidence correction term, The hidden state at any given moment, through The input gate of the parallel branches yields... time Input gate output with parallel branches; Will time Forget gate output, input gate output, candidate cell state, and parallel branch. The cell state at any given time is obtained through cell state updates. time Cell states with parallel branches; Will Time-dependent augmented input vector, cell state, The hidden state at any given moment, through The output gate of the parallel branch obtains time The output gate has multiple parallel branches; based on At any given moment, the output gate of each parallel branch, the cell state, and the calculation are performed. The hidden state of each parallel branch at any given time; based on The hidden state of each parallel branch at any given time is determined, and the probability weight corresponding to each parallel branch is obtained. based on Given the hidden state of each parallel branch at each time step and the probability weight corresponding to that branch, filter each parallel branch and select the branch corresponding to the hidden state at each time step. The hidden state at a given moment serves as the data from each sensor within a future preset time step; Based on the data from each sensor at each moment within a future preset time step, semantic tags are generated for the cold chain compartment at each moment within the future preset time step.

8. A cold chain logistics monitoring method according to claim 7, characterized in that, The basis The hidden state of each parallel branch at each time step and the probability weight corresponding to that branch are used to filter each parallel branch, including: Each The average of the product of the scaled confidence weight and the probability weight corresponding to the hidden state of each parallel branch at time t is used as the confidence score of each parallel branch. Will The average number of times the data from each sensor exceeds the threshold range corresponding to the hidden state of each parallel branch at any given time is used as the risk event score for each parallel branch. The difference between 1 and the risk event score of each parallel branch is used as the risk assessment index value of each parallel branch. Based on the arrival time to the target region and the maximum acceptable duration for each parallel branch, the efficiency index value for each parallel branch is obtained, using the following formula: ， in, For the first The efficiency index value of parallel branches. For the first The arrival time of each parallel branch to the target region. For the first The starting transport time corresponding to each parallel branch. For the maximum acceptable duration, For parallel branch indexes; Will The ratio of the number of times the hidden state of each parallel branch satisfies regulatory requirements to the total number of regulatory requirements is used as the compliance indicator value for each parallel branch. The confidence score, risk assessment index value, efficiency index value, and compliance index value of each parallel branch are weighted and summed to obtain the comprehensive score of each parallel branch. Based on the comprehensive score of each parallel branch, obtain the pruned branch set; Using dynamic programming, the optimal branch in the pruned branch set is obtained, and the corresponding branch is... The hidden state at a given moment serves as the data from each sensor within a future preset time step; Based on the data from each sensor at each moment within a future preset time step, semantic tags are generated for the cold chain compartment at each moment within the future preset time step.

9. A cold chain logistics monitoring method according to claim 8, characterized in that, After obtaining the data from each sensor and the semantic tags of the cold chain vehicle compartment within a future preset time step, the data within the future preset time step will be... The data from each sensor in each parallel branch, the semantic tags of the cold chain compartment, the comprehensive score of each parallel branch, the confidence score of each parallel branch, the risk assessment index value, the efficiency index value, the compliance index value and their corresponding weights, the pruning parameters, the dynamic programming algorithm parameters, and the optimal branch index are used as the original data. Within the preset time step in the future The data from each sensor in each parallel branch, along with the semantic tags of the cold chain vehicle compartment and the optimal branch index, are hashed. The generated hash value, along with the comprehensive score of each parallel branch within a preset future time step, the confidence score of each parallel branch, the risk assessment index value, the efficiency index value, the compliance index value and their corresponding weights, pruning parameters, and dynamic programming algorithm parameters, are used as leaf nodes. Through a Merkle tree structure, the Merkle root hash of the original data is generated. When the original data is accessed, the Merkle root hash of the original data is used to verify whether the original data has been tampered with.

10. A cold chain logistics monitoring system, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the cold chain logistics monitoring method according to any one of claims 1 to 9.

Images