Methods and systems for monitoring the storage environment during the cold chain distribution of fresh ingredients

By constructing a heterogeneous sensor network and a comprehensive disturbance index, the accuracy problem of environmental monitoring of fresh food under multi-source excitation was solved, enabling precise assessment and efficient handling of the storage environment of fresh food and reducing the loss rate.

CN123117960APending Publication Date: 2026-11-17GUANGDONG CANCHENG AGRI PROD GRP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611298920.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies cannot accurately monitor the true steady-state environment of fresh food under multi-source excitation, resulting in an inability to effectively assess its storage quality.

Method used

A heterogeneous sensor network is constructed, which combines road bump index, refrigeration equipment operating status and environmental data. By integrating disturbance index and virtual steady-state environment vector, the degree of non-steadiness of cold chain environment is accurately quantified, the remaining shelf life is estimated and graded early warning is given.

Benefits of technology

It has improved the accuracy of cold chain environment monitoring, reduced misjudgment interference, improved the efficiency of abnormal handling, reduced the loss rate of fresh food, accurately assessed the quality degradation pattern, and reduced the risk of missed judgment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117960A_ABST
    Figure CN123117960A_ABST
Patent Text Reader

Abstract

This invention provides a method and system for monitoring the storage environment during the cold chain distribution of fresh food, relating to the field of cold chain storage environment technology. The method includes: constructing a heterogeneous sensor network; acquiring the road surface bump index of a road segment and determining the collection frequency of the collection cycle; collecting a set of environmental data and generating a standardized environmental state vector; acquiring operating power and fan speed; determining a comprehensive disturbance index to measure the degree of non-steady-state condition of the cold chain environment under multi-source excitation; acquiring a virtual steady-state environment vector; determining the final estimated remaining shelf life based on the virtual steady-state environment vector; determining a dynamic risk coefficient based on the final estimated remaining shelf life and the delivery time at the current collection moment; and performing graded early warning output based on the dynamic risk coefficient and the virtual steady-state environment vector. According to this invention, the real steady-state environment of fresh food can be monitored based on the disturbance situation under multi-source excitation, reducing the monitoring misjudgment problem caused by disturbances.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of cold chain storage environment monitoring technology, and in particular to a method and system for monitoring the storage environment during the cold chain distribution of fresh food. Background Technology

[0002] Current technologies, while real-time monitoring of the storage environment involves collecting temperature data and performing simple threshold comparisons to determine whether temperatures are too high, do not consider disturbances caused by multiple sources of excitation during transportation, such as road bumps and the start-up and shutdown of refrigeration equipment. This results in monitoring data failing to accurately reflect the steady-state environment of fresh food, meaning that the true steady-state environment of fresh food cannot be monitored based on disturbances caused by multiple sources of excitation.

[0003] The information disclosed in the background section of this application is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0004] This invention provides a method and system for monitoring the storage environment during the cold chain distribution of fresh ingredients, which can solve the technical problem that related technologies cannot monitor the real steady-state environment of fresh ingredients based on disturbances under multi-source excitation.

[0005] According to a first aspect of the present invention, a method for monitoring the storage environment during the cold chain distribution of fresh food is provided, comprising:

[0006] In the storage space of cold chain distribution, multiple wireless sensor nodes are deployed according to a preset spatial distribution to construct a heterogeneous sensor network, wherein the heterogeneous sensor network includes a temperature sensor subnet, a humidity sensor subnet, and a volatile gas sensor subnet.

[0007] At the beginning of each collection cycle, the road surface bump index of the road segment is obtained based on the real-time spatial coordinates of the delivery unit, and the collection frequency of the collection cycle is determined.

[0008] According to the acquisition frequency, a set of environmental data is synchronously acquired from the heterogeneous sensor network to generate a standardized environmental state vector at the current acquisition time, wherein the environmental data includes temperature, humidity and characteristic gas concentration;

[0009] Obtain the operating power of the refrigeration equipment and the fan speed inside the delivery unit at the current data collection time;

[0010] Based on the road surface bump index, the operating power, the fan speed, and the standardized environmental state vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness of the cold chain environment under multi-source excitation.

[0011] Based on the comprehensive disturbance index and the standardized environmental state vector, the virtual steady-state environment vector is obtained;

[0012] Based on the virtual steady-state environment vector, the final estimated remaining shelf life is determined;

[0013] Based on the final estimated remaining shelf life and the delivery time at the current collection time, a dynamic risk coefficient is determined.

[0014] Based on the dynamic risk coefficient and the virtual steady-state environment vector, a graded early warning output is performed.

[0015] Furthermore, at the beginning of each collection cycle, the road surface bump index of the road segment is obtained based on the real-time spatial coordinates of the delivery unit, and the collection frequency of the collection cycle is determined, including:

[0016] Based on the real-time spatial coordinates of the delivery unit, query the cloud-based road condition database to obtain the road surface bump index of the road segment;

[0017] If the road surface bump index is greater than the first bump threshold, then the sampling frequency of the sampling period is the first sampling frequency.

[0018] If the road surface bump index is less than or equal to the first bump threshold and greater than the second bump threshold, then the sampling frequency of the sampling period is the second sampling frequency.

[0019] If the road surface bump index is less than or equal to the second bump threshold, then the sampling frequency of the sampling cycle is the third sampling frequency.

[0020] Furthermore, based on the acquisition frequency, a set of environmental data is synchronously acquired from the heterogeneous sensor network to generate a standardized environmental state vector at the current acquisition time, including:

[0021] Extract historical environmental data up to a preset sliding window length prior to the current acquisition time, wherein the historical environmental data includes historical temperature, historical humidity, and historical characteristic gas concentration;

[0022] Calculate the mean and standard deviation of each historical environment data within a preset sliding window length;

[0023] If the absolute value of the average value of a certain environmental data at the current collection time minus the corresponding historical environmental data is greater than the product of a preset coefficient and the standard deviation of the corresponding historical environmental data, then the corresponding environmental data at the current collection time will be removed and replaced with the historical environmental data corresponding to the previous collection time.

[0024] If the absolute value of the average value of the corresponding historical environmental data minus the current environmental data is less than or equal to the product of the preset coefficient and the standard deviation of the corresponding historical environmental data, then the corresponding environmental data at the current collection time is retained.

[0025] The environmental data collected at the current time, after being discarded or retained, are used to form a standardized environmental state vector, which consists of three components: standard temperature, standard humidity, and standard characteristic gas concentration.

[0026] Furthermore, based on the road surface bump index, the operating power, the fan speed, and the standardized environmental state vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness in the cold chain environment under multi-source excitation, including:

[0027] Obtain the reference values ​​for the operating power of the refrigeration equipment and the reference values ​​for the fan speed inside the compartment;

[0028] The ratio of the operating power to the operating power reference value is taken as the first ratio.

[0029] The ratio of the fan speed to the fan speed reference value is used as the second ratio.

[0030] The active disturbance coefficient at the current acquisition time is obtained by summing and averaging the first ratio and the second ratio.

[0031] The passive disturbance coefficient at the current acquisition time is obtained based on the ratio of the road surface bump index to the full value of the road surface bump index.

[0032] The difference vector between the standardized environmental state vector and the standardized environmental state vector at the previous acquisition time is taken as the fluctuation vector, wherein the fluctuation vector consists of three components: fluctuation temperature, fluctuation humidity, and fluctuation characteristic gas concentration.

[0033] Based on the active disturbance coefficient, the passive disturbance coefficient, and the fluctuation vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness of the cold chain environment under multi-source excitation.

[0034] Further, based on the comprehensive disturbance index and the standardized environmental state vector, a virtual steady-state environment vector is obtained, including:

[0035] When the comprehensive disturbance index is greater than the disturbance threshold, the historical standardized environmental state vector of the M consecutive historical acquisition times before the current acquisition time is obtained. The historical standardized environmental state vector consists of three components: historical standard temperature, historical standard humidity and historical standard characteristic gas concentration, where M is a positive integer.

[0036] For each dimension of the historical standardized environmental state vector, calculate the average slope of change over M intervals to generate the slope vector at the current acquisition time. The slope vector consists of three components: the average slope of change in the temperature dimension, the average slope of change in the humidity dimension, and the average slope of change in the characteristic gas concentration dimension.

[0037] The ratio of the comprehensive disturbance index to the sum of 1 and the comprehensive disturbance index is used as the scaling factor for the current acquisition time.

[0038] The compensation vector at the current acquisition moment is obtained by multiplying the negative value of the scaling factor with the slope vector.

[0039] The standardized environmental state vector and the compensation vector are added together to obtain the compensated virtual steady-state environmental vector, wherein the virtual steady-state environmental vector consists of three components: steady-state temperature, steady-state humidity, and steady-state characteristic gas concentration.

[0040] When the comprehensive disturbance index is less than or equal to the disturbance threshold, the standardized environmental state vector is used as the virtual steady-state environmental vector.

[0041] Further, based on the virtual steady-state environment vector, the final estimated remaining shelf life is determined, including:

[0042] Based on the steady-state temperature and steady-state humidity in the virtual steady-state environment vector, the decay rate constant of the microbial growth dimension is obtained by querying the first rate constant mapping table;

[0043] Based on the steady-state temperature and steady-state characteristic gas concentration in the virtual steady-state environment vector, the lipid oxidation dimension decay rate constant is obtained by querying the second rate constant mapping table.

[0044] Based on the steady-state temperature, steady-state characteristic gas concentration, and steady-state characteristic gas concentration in the virtual steady-state environment vector, the decay rate constant of the water activity dimension is obtained by querying the third rate constant mapping table.

[0045] Based on the decay rate constants of the microbial growth dimension, the lipid oxidation dimension, and the water activity dimension, the cumulative quality degradation that has occurred from the start of delivery to the current collection time is obtained.

[0046] Starting from the current collection time, and using the preset estimated time step as the increment, forward loop iterative calculation is performed on the time axis. In each iteration, the decay rate constants of the microbial growth dimension, lipid oxidation dimension, and water activity dimension at the current collection time are used as the decay rate constants of the three dimensions at the iteration time, and a deterioration increment of a preset estimated time step is calculated.

[0047] The predicted cumulative quality degradation is obtained based on the cumulative quality degradation, the degradation increment, and the preset estimated time step.

[0048] The iteration terminates when the predicted cumulative quality degradation first exceeds or equals the preset unacceptable quality threshold.

[0049] Record the total number of steps in the iterative calculation, and multiply the total number of steps by the calculation time step to obtain the basic remaining shelf life;

[0050] The total shelf life of all goods that match the current delivery type, origin and initial quality level in the historical cold chain distribution database is selected and averaged to obtain the statistical average shelf life.

[0051] Subtract the delivery time at the current collection time from the statistical average shelf life to obtain the average remaining shelf life;

[0052] If the difference between the base remaining shelf life and the average remaining shelf life is greater than a preset deviation threshold, then the weighted average of the base remaining shelf life and the average remaining shelf life is determined as the final estimated remaining shelf life.

[0053] If the difference between the basic remaining shelf life and the average remaining shelf life is less than or equal to a preset deviation threshold, then the basic remaining shelf life is determined as the final estimated remaining shelf life.

[0054] Furthermore, based on the final estimated remaining shelf life and the delivery time at the current collection time, a dynamic risk coefficient is determined, including:

[0055] Obtain the highest and average temperature values ​​from historical data collection times within a preset historical time window prior to the current data collection time.

[0056] The ratio of the highest temperature value to the average temperature value is used as the temperature peak factor;

[0057] The dynamic risk coefficient is determined by multiplying the ratio of the delivery time at the current collection moment to the final estimated remaining shelf life by the temperature peak factor.

[0058] Furthermore, based on the dynamic risk coefficient and the virtual steady-state environment vector, a tiered early warning output is performed, including:

[0059] The dynamic risk coefficient is mapped to generate a multi-level cold chain risk status code, wherein the multi-level cold chain risk status code includes a safety level code, a concern level code, and a critical level code;

[0060] Calculate the vector difference between the virtual steady-state environment vector at the current acquisition time and the virtual steady-state environment vector at the previous acquisition time;

[0061] By performing time differentiation on the L2 norm of the vector difference, the characteristic quantity of environmental mutation trend can be obtained;

[0062] If the environmental mutation trend feature is greater than the preset mutation alarm threshold, the multi-level cold chain risk status code will be forcibly set to the critical level code.

[0063] If the environmental mutation trend feature is less than or equal to the preset mutation alarm threshold, then a graded early warning output is performed based on the multi-level cold chain risk status code.

[0064] According to a second aspect of the present invention, a storage environment monitoring system for the cold chain distribution of fresh food is provided, comprising:

[0065] A heterogeneous sensor network module is used to deploy multiple wireless sensor nodes in a preset spatial distribution within the storage space of cold chain distribution to construct a heterogeneous sensor network, wherein the heterogeneous sensor network includes a temperature sensing subnet, a humidity sensing subnet, and a volatile gas sensing subnet.

[0066] The sampling frequency module is used to obtain the road surface bump index of the road segment based on the real-time spatial coordinates of the delivery unit at the beginning of each sampling cycle, and to determine the sampling frequency of the sampling cycle.

[0067] The standardized environmental state vector module is used to synchronously collect a set of environmental data from the heterogeneous sensor network according to the acquisition frequency, and generate a standardized environmental state vector at the current acquisition time, wherein the environmental data includes temperature, humidity and characteristic gas concentration;

[0068] The delivery unit data module is used to obtain the operating power of the refrigeration equipment and the fan speed inside the delivery unit at the current data collection time.

[0069] The comprehensive disturbance index module is used to determine a comprehensive disturbance index that measures the degree of unsteadiness of the cold chain environment under multi-source excitation based on the road surface bump index, the operating power, the fan speed and the standardized environmental state vector.

[0070] The virtual steady-state environment vector module is used to obtain the virtual steady-state environment vector based on the comprehensive disturbance index and the standardized environment state vector.

[0071] The final estimated remaining shelf life module is used to determine the final estimated remaining shelf life based on the virtual steady-state environment vector.

[0072] The dynamic risk coefficient module is used to determine the dynamic risk coefficient based on the final estimated remaining shelf life and the delivery time at the current collection time.

[0073] The graded early warning output module is used to output graded early warnings based on the dynamic risk coefficient and the virtual steady-state environment vector.

[0074] Technical Effects: According to this invention, by adjusting the sampling frequency through the road bump index, the computational power required for sampling in low-disturbance scenarios can be reduced, improving the timeliness of anomaly detection in bumpy road sections. By comprehensively analyzing the disturbance index, the degree of non-steady-state condition in the cold chain environment can be quantified, eliminating interference from misjudgments of single environmental parameters. By correcting the deviations caused by measured environmental data under non-steady-state disturbances using a virtual steady-state environment vector, the final calculated remaining shelf life more closely reflects the actual quality degradation pattern of fresh food, demonstrating the true steady-state environment in which fresh food exists. Based on dynamic risk coefficients and virtual steady-state environment vectors, a tiered early warning system can be implemented, pushing different handling strategies according to risk levels, reducing invalid alarms, significantly improving the efficiency of cold chain anomaly handling, and reducing the loss rate of fresh food. When determining the comprehensive disturbance index, active disturbances can be precisely bound to the temperature-sensitive dimension, and passive disturbances to the humidity dimension. This reduces assessment bias caused by mismatches between different disturbance sources and environmental parameters, aligning with the actual operating conditions of the cold chain. A new cross-coupling term for the product of active and passive disturbances is added to accurately recreate the nonlinear multiplication effect of characteristic gas diffusion when two disturbances act simultaneously. The comprehensive disturbance index can quantify the overall unsteady state of the cold chain environment under multi-source excitation, improving the accuracy of unsteady state assessment. When determining the final estimated remaining shelf life, the degradation of fresh food quality can be precisely quantified across multiple dimensions. A preset deviation threshold judgment mechanism integrates the basic remaining shelf life and the statistical average remaining shelf life to determine the final estimated remaining shelf life. Weights are adaptively adjusted: the basic remaining shelf life is given greater weight when disturbances are severe, while the statistical average remaining shelf life is emphasized when the environment is stable. This reduces the bias caused by occasional sensor errors in single real-time data and corrects the lag in adapting purely historical statistics to the specific operating conditions of the current batch, significantly improving the stability of the estimated results under extreme scenarios. When determining the dynamic risk coefficient, the cumulative consumption of quality over time and the impact of instantaneous high temperature can be used to determine the dynamic risk coefficient, so as to more comprehensively assess the risk level of the storage environment during the delivery process. By amplifying the damage weight of instantaneous extreme high temperature through the temperature peak factor, the risk of short-term temperature change that is completely covered by traditional average temperature indicators can be accurately captured, reducing the hidden dangers of short-term high temperature rapid spoilage in the fresh food cold chain.

[0075] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Other features and aspects of the invention will become clearer from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0076] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0077] Figure 1 An exemplary flowchart illustrates a method for monitoring the storage environment during the cold chain distribution of fresh ingredients according to an embodiment of the present invention.

[0078] Figure 2 A flowchart for determining the acquisition frequency according to an embodiment of the present invention is shown as an example;

[0079] Figure 3 An exemplary flowchart illustrating the calculation of a standardized environment state vector according to an embodiment of the present invention is shown;

[0080] Figure 4 A flowchart for calculating the comprehensive disturbance index according to an embodiment of the present invention is shown as an example;

[0081] Figure 5 An exemplary flowchart illustrating the calculation of a virtual steady-state environment vector according to an embodiment of the present invention is shown;

[0082] Figure 6 An exemplary flowchart illustrating the calculation of the final estimated remaining shelf life according to an embodiment of the present invention is shown;

[0083] Figure 7 An exemplary flowchart illustrating the calculation of dynamic risk coefficients according to an embodiment of the present invention is shown;

[0084] Figure 8 A flowchart of the graded early warning output according to an embodiment of the present invention is shown as an example;

[0085] Figure 9 A block diagram of a fresh food cold chain distribution process storage environment monitoring system according to an embodiment of the present invention is shown as an example. Detailed Implementation

[0086] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0087] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0088] Figure 1 An exemplary flowchart illustrates a method for monitoring the storage environment during the cold chain distribution of fresh food according to an embodiment of the present invention, the method comprising:

[0089] Step S1: In the storage space of cold chain distribution, multiple wireless sensor nodes are deployed according to a preset spatial distribution to construct a heterogeneous sensor network, wherein the heterogeneous sensor network includes a temperature sensor subnet, a humidity sensor subnet, and a volatile gas sensor subnet.

[0090] Step S2: At the beginning of each collection cycle, the road surface bump index of the road segment is obtained based on the real-time spatial coordinates of the delivery unit, and the collection frequency of the collection cycle is determined.

[0091] Step S3: Based on the acquisition frequency, a set of environmental data is synchronously acquired from the heterogeneous sensor network to generate a standardized environmental state vector at the current acquisition time, wherein the environmental data includes temperature, humidity and characteristic gas concentration;

[0092] Step S4: Obtain the operating power of the refrigeration equipment of the delivery unit and the fan speed inside the compartment at the current data collection time;

[0093] Step S5: Determine the comprehensive disturbance index, which measures the degree of unsteadiness of the cold chain environment under multi-source excitation, based on the road surface bump index, the operating power, the fan speed, and the standardized environmental state vector.

[0094] Step S6: Obtain the virtual steady-state environment vector based on the comprehensive disturbance index and the standardized environmental state vector;

[0095] Step S7: Determine the final estimated remaining shelf life based on the virtual steady-state environment vector;

[0096] Step S8: Determine the dynamic risk coefficient based on the final estimated remaining shelf life and the delivery time at the current collection time;

[0097] Step S9: Based on the dynamic risk coefficient and the virtual steady-state environment vector, perform hierarchical early warning output.

[0098] The method for monitoring the storage environment during the cold chain distribution of fresh food according to embodiments of the present invention, by adjusting the sampling frequency through the road bump index, can reduce the computing power required for sampling in low-disturbance scenarios and improve the timeliness of anomaly detection in bumpy road sections. By comprehensively considering the disturbance index, the non-steady-state degree of the cold chain environment can be quantified, eliminating interference from misjudgments of single environmental parameters. By correcting the deviations caused by measured environmental data under non-steady-state disturbances through a virtual steady-state environment vector, the final calculated remaining shelf life more closely reflects the actual quality decay law of fresh food, thus reflecting the true steady-state environment in which the fresh food exists. Based on dynamic risk coefficients and virtual steady-state environment vectors, a tiered early warning system can be implemented, pushing different handling strategies according to risk levels, reducing invalid alarms, significantly improving the efficiency of cold chain anomaly handling, and reducing the loss rate of fresh food.

[0099] According to one embodiment of the present invention, in step S1, in the storage space of cold chain distribution, due to the need to pack fresh food, multiple wireless sensor nodes are deployed above the loading line, and are evenly arranged in the space above the loading line at 10cm intervals vertically and 40cm intervals horizontally, thereby constructing a heterogeneous sensor network. The heterogeneous sensor network includes a temperature sensing subnetwork, a humidity sensing subnetwork, and a volatile gas sensing subnetwork. The temperature sensing subnetwork uses a PT1000 platinum resistance thermometer, the humidity sensing subnetwork uses a capacitive polymer humidity sensor, and the volatile gas sensing subnetwork uses a metal oxide semiconductor sensor, which is suspended from the top of the container.

[0100] According to one embodiment of the present invention, in step S2, at the beginning of each collection cycle (e.g., 10 minutes), the road surface bump index of the road segment is obtained based on the real-time spatial coordinates of the delivery unit, and the collection frequency of the collection cycle is determined.

[0101] Figure 2 A flowchart for determining the acquisition frequency according to an embodiment of the present invention is shown as an example.

[0102] According to an embodiment of the present invention, step S2 includes: step S21, querying the cloud-based road condition database based on the real-time spatial coordinates of the delivery unit to obtain the road surface bump index of the road segment; step S22, if the road surface bump index is greater than a first bump threshold, then the collection frequency of the collection cycle is a first collection frequency; step S23, if the road surface bump index is less than or equal to the first bump threshold and greater than a second bump threshold, then the collection frequency of the collection cycle is a second collection frequency; step S24, if the road surface bump index is less than or equal to the second bump threshold, then the collection frequency of the collection cycle is a third collection frequency.

[0103] According to one embodiment of the present invention, the real-time spatial coordinates of the delivery unit are obtained via a built-in GPS. Based on the real-time spatial coordinates, a cloud-based road condition database is queried. If the real-time spatial coordinates belong to highway conditions, the road bump index is 1; if the real-time spatial coordinates belong to urban arterial road conditions, such as asphalt roads, the road bump index is 4; if the real-time spatial coordinates belong to district-level roads, such as cement roads, the road bump index is 8; and if the real-time spatial coordinates belong to gravel / unpaved road conditions, the road bump index is 10. If the road bump index is greater than a first bump threshold (e.g., 6), the collection frequency of the collection cycle is a first collection frequency (e.g., 180 seconds); if the road bump index is less than or equal to the first bump threshold and greater than a second bump threshold (e.g., 3), the collection frequency of the collection cycle is a second collection frequency (e.g., 60 seconds); and if the road bump index is less than or equal to the second bump threshold, the collection frequency of the collection cycle is a third collection frequency (e.g., 15 seconds).

[0104] According to an embodiment of the present invention, in step S3, the storage environment of the cold chain distribution process of fresh food is monitored once every preset time interval. The preset time interval can be set to one hour, one and a half hours, etc. The present invention does not limit this. According to the acquisition frequency, that is, according to the acquisition frequency of the acquisition cycle in which the current monitoring time is located, a set of environmental data is synchronously acquired from the heterogeneous sensor network. The set of environmental data includes the temperature, humidity and characteristic gas concentration at the current acquisition time (the acquisition time closest to the current monitoring time). The temperature, humidity and characteristic gas concentration (for example, the characteristic gas concentration of leafy vegetables is ethylene concentration) are the average values ​​of the temperature value, humidity value and characteristic gas concentration value acquired at multiple wireless sensor nodes, respectively, thereby generating a standardized environmental state vector at the current acquisition time.

[0105] Figure 3 A flowchart illustrating the calculation of a standardized environment state vector according to an embodiment of the present invention is shown.

[0106] According to an embodiment of the present invention, step S3 includes: step S31, extracting historical environmental data within a preset sliding window length prior to the current acquisition time, wherein the historical environmental data includes historical temperature, historical humidity, and historical characteristic gas concentration; step S32, calculating the average value and standard deviation of each type of historical environmental data within the preset sliding window length; step S33, if the absolute value of the average value of a certain environmental data at the current acquisition time minus the corresponding historical environmental data is greater than the product of a preset coefficient and the standard deviation of the corresponding historical environmental data, then the corresponding environmental data at the current acquisition time is removed and replaced with the historical environmental data corresponding to the previous acquisition time; step S34, if the absolute value of the average value of a certain environmental data at the current acquisition time minus the corresponding historical environmental data is less than or equal to the product of a preset coefficient and the standard deviation of the corresponding historical environmental data, then the corresponding environmental data at the current acquisition time is retained; step S35, forming a standardized environmental state vector from the environmental data at the current acquisition time after removal or retention, wherein the standardized environmental state vector consists of three components: standard temperature, standard humidity, and standard characteristic gas concentration.

[0107] According to one embodiment of the present invention, historical environmental data within a preset sliding window length (e.g., 11 historical acquisition times with a step size of 1) prior to the current acquisition time is extracted. Specifically, historical environmental data from 11 historical acquisition times prior to the current acquisition time is obtained, where the last of these 11 historical acquisition times is adjacent to the current acquisition time. The historical environmental data includes historical temperature, historical humidity, and historical characteristic gas concentration. The mean and standard deviation of each type of historical environmental data within the preset sliding window length are calculated. Specifically, the mean and standard deviation of each type of historical environmental data from the 11 historical acquisition times, the mean and standard deviation of historical humidity from the 11 historical acquisition times, and the mean and standard deviation of historical characteristic gas concentration from the 11 historical acquisition times are calculated. If the absolute value of the current environmental data minus the mean of the corresponding historical environmental data is greater than the product of a preset coefficient (e.g., 3) and the standard deviation of the corresponding historical environmental data (for example, temperature), then... ,in, The temperature at the current time of data collection. This is the average of the historical temperatures from the 11 historical data collection times preceding the current data collection time. The standard deviation of the historical temperature from the 11 historical data points preceding the current data collection time is used. If a certain environmental data point at the current data collection time is unreliable, the corresponding environmental data point is discarded and replaced with historical environmental data from the previous data collection time. If the absolute value of the difference between the current environmental data point and the average of the corresponding historical environmental data is less than or equal to the product of a preset coefficient and the standard deviation of the corresponding historical environmental data, then the environmental data point at the current data collection time is considered reliable and is retained. The environmental data from the current data collection time, after discarding or retaining data, are used to form a standardized environmental state vector. This standardized environmental state vector consists of three components: standard temperature, standard humidity, and standard characteristic gas concentration.

[0108] According to one embodiment of the present invention, in step S4, the operating power of the refrigeration equipment (e.g., refrigeration unit) of the delivery unit (e.g., refrigeration unit) and the fan speed inside the compartment are obtained at the current acquisition time by an AC power transmitter installed on the power supply circuit side of the refrigeration unit and a non-contact Hall speed sensor installed on the motor output shaft end.

[0109] According to an embodiment of the present invention, in step S5, a comprehensive disturbance index is determined based on the road surface bump index, the operating power, the fan speed, and the standardized environmental state vector to measure the degree of unsteadiness of the cold chain environment under multi-source excitation.

[0110] Figure 4 A flowchart for calculating the comprehensive disturbance index according to an embodiment of the present invention is shown as an example.

[0111] According to an embodiment of the present invention, step S5 includes: step S51, obtaining the operating power reference value of the refrigeration equipment and the fan speed reference value inside the compartment; step S52, taking the ratio of the operating power to the operating power reference value as a first ratio; step S53, taking the ratio of the fan speed to the fan speed reference value as a second ratio; step S54, summing and averaging the first ratio and the second ratio to obtain the active disturbance coefficient at the current acquisition time; step S55, obtaining the passive disturbance coefficient at the current acquisition time based on the ratio of the road bump index to the full value of the road bump index; step S56, taking the difference vector between the standardized environmental state vector and the standardized environmental state vector at the previous acquisition time as a fluctuation vector, wherein the fluctuation vector consists of three components: fluctuating temperature, fluctuating humidity, and fluctuating characteristic gas concentration; step S57, determining a comprehensive disturbance index that measures the degree of unsteadiness of the cold chain environment under multi-source excitation based on the active disturbance coefficient, the passive disturbance coefficient, and the fluctuation vector.

[0112] According to one embodiment of the present invention, a reference value for the operating power of the refrigeration equipment (e.g., 2000W) and a reference value for the fan speed inside the compartment (e.g., 1600r / min) are obtained, and the ratio of the operating power to the reference value is used as a first ratio. ,Right now, ,in, The operating power at the current data acquisition moment. As the operating power reference value, the ratio of the fan speed to the fan speed reference value is used as the second ratio. ,Right now, ,in, The fan speed at the current data collection moment. Using the wind turbine speed as a reference value, the first ratio and the second ratio are summed and averaged to obtain the active disturbance coefficient at the current acquisition time. ,Right now, The passive disturbance coefficient at the current data collection time is obtained by comparing the road surface bump index with the full value of the road surface bump index. ,Right now, ,in, This represents the road bump index at the current data collection time. The full value of the road surface bump index is given, for example, 10. Based on the active disturbance coefficient, passive disturbance coefficient, and fluctuation vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness in the cold chain environment under multi-source excitation.

[0113] According to one embodiment of the present invention, a comprehensive disturbance index measuring the degree of unsteadiness of the cold chain environment under multi-source excitation is determined based on the active disturbance coefficient, the passive disturbance coefficient, and the fluctuation vector, including: determining the comprehensive disturbance index at the current sampling time according to formula (1). ,

[0114] (1),

[0115] in, This represents the active perturbation coefficient at the current data acquisition moment. This represents the passive disturbance coefficient at the current acquisition moment. This represents the temperature component of the fluctuation vector at the current acquisition moment. As a reference scalar for temperature fluctuations, This represents the fluctuating humidity component of the fluctuation vector at the current acquisition time. As a reference scalar for humidity fluctuations, The fluctuation characteristic gas concentration component of the fluctuation vector at the current acquisition time. t is the reference scalar for the characteristic gas fluctuation, and t is the current acquisition time.

[0116] According to an embodiment of the present invention, in formula (1), The active disturbance coefficient at the current acquisition time is the product of the fluctuation temperature component of the fluctuation vector at the current acquisition time and the temperature fluctuation reference scalar (e.g., leafy vegetables accept a temperature fluctuation of 2℃). It represents the influence of active disturbance on the temperature dimension. Temperature fluctuation is mainly caused by active disturbance (e.g., the start and stop of refrigeration equipment, fan agitation). Passive disturbance (e.g., road bumps) has a very small direct impact on temperature (bumps do not directly generate heat or cool). The higher the value, the greater the temperature fluctuation. The passive disturbance coefficient at the current acquisition time is the product of the ratio of the fluctuating humidity component of the fluctuation vector at the current acquisition time to the humidity fluctuation reference scalar (e.g., leafy vegetables accept a 5% humidity fluctuation). This represents the impact of passive disturbance on the humidity dimension. Humidity fluctuations are mainly caused by passive disturbances (bumps causing instantaneous changes in the airtightness of the chamber, and brief exchanges of cold air with the outside environment). The higher the value, the greater the humidity fluctuation. The product of the active disturbance coefficient, the passive disturbance coefficient, and the ratio of the characteristic gas concentration component of the fluctuation vector at the current acquisition time to the characteristic gas fluctuation reference scalar (e.g., ethylene, a marker gas for the ripening and senescence of leafy vegetables, 2 ppm) represents the influence of the cross-coupling term of active and passive disturbances on the characteristic gas concentration dimension. When both vibration (passive disturbance) and airflow change (active disturbance) exist in the compartment, the diffusion rate of the characteristic gas concentration (e.g., ethylene) in the cold storage compartment exhibits a nonlinear multiplication effect; therefore, the product of the two disturbance coefficients is used. The above calculation... , and The square root of the sum of the three squares yields the comprehensive disturbance index at the current acquisition time, representing the degree of unsteadiness of the two disturbance sources—active and passive—in the three dimensions of temperature, humidity, and characteristic gas concentration. The larger the comprehensive disturbance index, the greater the degree of unsteadiness of the cold chain environment under multi-source excitation.

[0117] In this way, by precisely binding active disturbances to the temperature-sensitive dimension and passive disturbances to the humidity dimension, the assessment bias caused by the mismatch between different disturbance sources and environmental parameters can be reduced, thus conforming to the actual working conditions of the cold chain. A new cross-coupling term for the product of active and passive disturbances can be added to accurately restore the nonlinear multiplication effect of characteristic gas diffusion when two disturbances act simultaneously. The comprehensive disturbance index can quantify the overall unsteady state of the cold chain environment under multi-source excitation, thereby improving the accuracy of unsteady state assessment.

[0118] According to an embodiment of the present invention, in step S6, a virtual steady-state environment vector is obtained based on the comprehensive disturbance index and the standardized environmental state vector.

[0119] Figure 5A flowchart for calculating a virtual steady-state environment vector is shown as an example according to an embodiment of the present invention.

[0120] According to an embodiment of the present invention, step S6 includes: step S61, when the comprehensive disturbance index is greater than the disturbance threshold, obtaining the historical standardized environmental state vector for M consecutive historical acquisition times prior to the current acquisition time, wherein the historical standardized environmental state vector consists of three components: historical standard temperature, historical standard humidity, and historical standard characteristic gas concentration, and M is a positive integer; step S62, calculating the average change slope of M intervals for each dimension of the historical standardized environmental state vector to generate the slope vector at the current acquisition time, wherein the slope vector consists of the average change slope of the temperature dimension, the average change slope of the humidity dimension, and the average slope of the characteristic gas concentration dimension. The system consists of three components; Step S63: The ratio of the comprehensive disturbance index to the sum of 1 and the comprehensive disturbance index is used as the scaling factor for the current acquisition time; Step S64: The compensation vector for the current acquisition time is obtained by multiplying the negative value of the scaling factor with the slope vector; Step S65: The standardized environmental state vector and the compensation vector are added together to obtain the compensated virtual steady-state environmental vector, wherein the virtual steady-state environmental vector consists of three components: steady-state temperature, steady-state humidity, and steady-state characteristic gas concentration; Step S66: When the comprehensive disturbance index is less than or equal to the disturbance threshold, the standardized environmental state vector is used as the virtual steady-state environmental vector.

[0121] According to one embodiment of the present invention, when the comprehensive disturbance index is greater than the disturbance threshold (e.g., 0.9), the current storage environment is in an unsteady state, the sensor readings are spurious, and environmental pre-compensation needs to be triggered. This involves acquiring a historically standardized environmental state vector from M (e.g., 5) consecutive historical acquisition times prior to the current acquisition time. The last historical acquisition time among these M consecutive historical acquisition times is adjacent to the current acquisition time. The historically standardized environmental state vector consists of three components: historically standardized temperature, historically standardized humidity, and historically standardized characteristic gas concentration. The average slope of change over M intervals is calculated for each dimension of the historically standardized environmental state vector, i.e., the average slope of change in the temperature dimension, the average slope of change in the humidity dimension, and the average slope of change in the characteristic gas concentration dimension. This constitutes the slope vector at the current acquisition time. This slope vector represents the predicted directional change of the environmental state at the next acquisition time in the three dimensions of temperature, humidity, and characteristic gas concentration without any intervention, i.e., the prediction deviation, quantifying the direction and magnitude of the deviation. The ratio of the comprehensive disturbance index to the sum of 1 and the comprehensive disturbance index is used as the scaling factor for the current acquisition time. ,Right now, The slope vector can be scaled, and the compensation magnitude for each dimension follows the influence of the physical disturbance source. A larger scaling factor indicates a larger disturbance, suggesting that the prediction bias is more likely to be a spurious fluctuation caused by a strong disturbance; therefore, stronger reverse compensation is needed to offset it. The compensation vector at the current acquisition time is obtained by multiplying the negative value of the scaling factor by the slope vector. ,Right now, ,in, Let be the slope vector at the current acquisition time. The compensation vector is in the opposite direction to the slope vector. For example, if the environmental state at the next acquisition time changes positively in the temperature dimension (i.e., the temperature at the next acquisition time is higher than the temperature at the current acquisition time), indicating that the environment is being pushed towards higher temperatures by a disturbance, then the compensation amount is negative (in the direction of cooling) to offset the predicted upward trend, thus obtaining a virtual steady-state value unaffected by the disturbance. Adding the standardized environmental state vector and the compensation vector yields the compensated virtual steady-state environmental vector, i.e., Obtain the virtual steady-state environment vector after compensation at the current acquisition time. ,in, The standardized environmental state vector at the current data collection moment is used as the virtual steady-state environment vector. This vector consists of three components: steady-state temperature, steady-state humidity, and steady-state characteristic gas concentration. Due to a time-scale mismatch between the sensor's "fast response" and the food's "slow response"—for example, the sensor measures air temperature (which has extremely low heat capacity and is highly volatile), while the deterioration of fresh food depends on its core temperature (which has extremely high heat capacity and is almost unaffected by second-level disturbances)—the sensor readings amplify the false impact of instantaneous disturbances on the quality of fresh food. That is, because fresh food has high thermal inertia, if the sensor detects a sudden increase in temperature, the true state of the fresh food will be misjudged as accelerated deterioration. The virtual steady-state environment vector is the equivalent value of the slow changes in temperature, humidity, and characteristic gas concentration of the fresh food. When the comprehensive disturbance index is less than or equal to the disturbance threshold, the current storage environment is in a steady state, and the standardized environmental state vector is used as the virtual steady-state environment vector. The virtual steady-state environment vector reflects the true steady-state environment of the fresh food.

[0122] According to one embodiment of the present invention, in step S7, the final estimated remaining shelf life is determined based on the virtual steady-state environment vector.

[0123] Figure 6 An exemplary flowchart illustrating the calculation of the final estimated remaining shelf life according to an embodiment of the present invention is shown.

[0124] According to an embodiment of the present invention, step S7 includes: step S71, obtaining the decay rate constant of the microbial growth dimension by querying a first rate constant mapping table based on the steady-state temperature and steady-state humidity in the virtual steady-state environment vector; step S72, obtaining the decay rate constant of the lipid oxidation dimension by querying a second rate constant mapping table based on the steady-state temperature and steady-state characteristic gas concentration in the virtual steady-state environment vector; and step S73, obtaining the decay rate constant of the water activity dimension by querying a third rate constant mapping table based on the steady-state temperature, steady-state characteristic gas concentration, and steady-state characteristic gas concentration in the virtual steady-state environment vector. Deceleration rate constant; Step S74, based on the decay rate constant of the microbial growth dimension, the decay rate constant of the lipid oxidation dimension, and the decay rate constant of the water activity dimension, obtain the cumulative quality degradation amount that has occurred from the start of delivery to the current collection time; Step S75, taking the current collection time as the starting point and using a preset calculation time step as the increment, perform forward iterative calculation on the time axis. In each iteration, the decay rate constants of the microbial growth dimension, the lipid oxidation dimension, and the water activity dimension at the current collection time are used as the decay rates of the three dimensions at the iteration time. A constant is used to calculate the degradation increment over a preset estimated time step; Step S76: Based on the cumulative quality degradation, the degradation increment, and the preset estimated time step, the predicted cumulative quality degradation is obtained; Step S77: When the predicted cumulative quality degradation first exceeds or equals a preset unacceptable quality threshold, the iteration is terminated; Step S78: The total number of iterations is recorded, and the total number of iterations is multiplied by the estimated time step to obtain the basic remaining shelf life; Step S79: From the historical cold chain delivery database, all goods matching the current delivery type, origin, and initial quality level are selected. Step S710: Subtract the delivery time already taken at the current collection time from the statistical average shelf life to obtain the average remaining shelf life; Step S711: If the difference between the basic remaining shelf life and the average remaining shelf life is greater than a preset deviation threshold, then the weighted average of the basic remaining shelf life and the average remaining shelf life is determined as the final estimated remaining shelf life; Step S712: If the difference between the basic remaining shelf life and the average remaining shelf life is less than or equal to the preset deviation threshold, then the basic remaining shelf life is determined as the final estimated remaining shelf life.

[0125] According to one embodiment of the present invention, the first rate constant mapping table, the second rate constant mapping table, and the third rate constant mapping table are discrete numerical tables pre-constructed through a finite number of standard experiments (measuring microbial growth rate, lipid oxidation rate, and water loss rate under different steady-state temperature, humidity, and gas concentration combinations). For example, the mapping table of the decay rate constant of microbial growth dimension (the first rate constant mapping table) is constructed as follows: In a laboratory constant temperature and humidity chamber, using leafy vegetables (e.g., lettuce) as samples, temperature gradients of -2℃, 0℃, 2℃, 4℃, 6℃, 8℃, and 10℃ (7 levels in total) and humidity gradients of 80%, 85%, 90%, 95%, and 98% (5 levels in total) are set, forming 7×5=35 orthogonal experimental points. At each orthogonal experimental point, the total number of psychrophilic bacteria (CFU / g) on ​​the sample surface is measured every 24 hours using the plate counting method. This measurement is repeated for 7 consecutive days. The slope of the logarithm of the total number of colonies versus time is fitted by linear regression, and this slope is the decay rate constant under this temperature-humidity combination. In actual monitoring, based on the steady-state temperature and humidity in the input virtual steady-state environment vector, bilinear interpolation is performed among 35 discrete points to obtain the microbial growth dimension decay rate constant at the current collection time. Similarly, a mapping table for the lipid oxidation dimension decay rate constant (second rate constant mapping table) is constructed: 7×6=42 orthogonal experimental points are formed using temperature gradients (the same 7 levels as above) and gas concentration gradients of 0ppm, 2ppm, 5ppm, 10ppm, 20ppm, and 50ppm (a total of 6 levels, for example, prepared using ethylene standard gas). At each experimental point, the malondialdehyde (MDA) content (μmol / g) of the sample is measured every 12 hours using the thiobarbituric acid method (TBA method). This is repeated for 5 consecutive days. The slope of the MDA content versus time is fitted by linear regression, which is the lipid oxidation dimension decay rate constant. A mapping table for the decay rate constant of water activity dimension (third rate constant mapping table) is constructed: 7 × 5 × 6 = 210 three-dimensional orthogonal experimental points are formed using temperature gradient (7 levels), humidity gradient (5 levels), and gas concentration gradient (6 levels). (This can be reduced to 42 representative points using a uniform design method to save experimental costs.) At each experimental point, the water loss rate (%) of the sample is measured every 6 hours using a weighing method. This measurement is repeated for 3 consecutive days. The slope of the water loss rate versus time is fitted using linear regression, which is the decay rate constant of the water activity dimension. Based on the decay rate constants of the microbial growth dimension, lipid oxidation dimension, and water activity dimension, the cumulative quality degradation from the start of distribution to the current collection time is obtained, i.e., according to the formula: Obtain the cumulative quality degradation that has occurred from the start of delivery to the current collection time. ,in, Let be the constant of the decay rate of microbial growth dimension at the i-th acquisition time. Let be the lipid oxidation dimension decay rate constant at the i-th acquisition time. Let be the constant of the decay rate of water activity dimension at the i-th sampling time. Let be the time interval between the (i-1)th and the ith time interval, and N be the number of data collections from the delivery start time to the current time, where i ≤ N, and both i and N are positive integers. No data is collected at the delivery start time. The time interval between the start of delivery and the first data collection time. Let be the geometric mean of the decay rate constants of microbial growth, lipid oxidation, and water activity at the i-th sampling time. This represents the result of the deterioration of fresh food quality under the synergistic effect of multiple dimensions. Microbial reproduction will damage cell structure and accelerate lipid oxidation. The free radicals generated by lipid oxidation will inhibit microbial growth (antagonistic effect). Water loss will simultaneously exacerbate the reaction rates of both. The three dimensions have both synergistic and antagonistic effects. Mathematically, the geometric mean can reflect the "product effect". That is, when the decay rate constant of any dimension approaches zero, the geometric mean will also approach zero, which means that the overall deterioration is suppressed. This represents the cumulative quality degradation of fresh food products due to the synergistic effects of multiple dimensions from the start of delivery to the current collection time. The larger the cumulative quality degradation, the more severe the quality degradation. Starting from the current collection time and using a preset calculation time step (e.g., 0.5 hours) as the increment, iterative calculations are performed forward along the time axis. In each iteration, the decay rate constants for the three dimensions at the current collection time are calculated by looking up the virtual steady-state environment vector corresponding to the current collection time (assuming the future storage environment remains at the current steady-state level, i.e., the virtual steady-state environment vector remains unchanged at future iteration times). Specifically, the decay rate constants for the microbial growth dimension, lipid oxidation dimension, and water activity dimension at the current collection time are used as the decay rate constants for the three dimensions at the iteration time. According to the formula: Calculate the deterioration increment at a preset estimated time step for the current acquisition time. ,in, This represents the constant of the decay rate of microbial growth dimensions at the current data collection time. This represents the lipid oxidation dimension decay rate constant at the current acquisition time. Let be the constant representing the decay rate of water activity dimension at the current sampling time. The preset calculation time step is used. Based on the cumulative quality degradation, the increment of degradation, and the preset calculation time step, the predicted cumulative quality degradation is obtained, i.e., according to the formula: To obtain the predicted cumulative quality degradation ,in, The total number of steps is the number of preset estimated time steps that have been taken after the iteration terminates when the predicted cumulative quality degradation first exceeds or equals the preset unacceptable quality threshold (2.8 for leafy vegetables). Multiplying the total number of steps by the estimated time step yields the basic remaining shelf life, which is the estimated time from the current collection time to the point where the quality becomes unacceptable. All historical records matching the current delivery type, origin, and initial quality level are selected from the historical cold chain delivery database. Each historical record is marked with a total shelf life, i.e., the total time from factory delivery to the point of unacceptability (fresh ingredients arriving at each delivery are sampled, and the total time from factory delivery to quality failure is recorded and stored in the historical cold chain delivery database). The total shelf life is averaged to obtain the statistical average shelf life. Subtracting the delivery time at the current collection time from the statistical average shelf life yields the average remaining shelf life, i.e., the remaining shelf life obtained from historical statistics relative to the delivery time at the current collection time. If the difference between the baseline remaining shelf life and the average remaining shelf life is greater than a preset deviation threshold (e.g., 12 hours for leafy vegetables), then the weighted average of the baseline remaining shelf life and the average remaining shelf life is determined as the final estimated remaining shelf life. If the current batch has special anomalies (e.g., the current environment is better or worse than the historical average), the two need to be combined, that is, according to the formula: To obtain the final estimated remaining shelf life at the current data collection time. ,in, The remaining shelf life at the current collection time. The average remaining shelf life at the current collection time. This is a dynamic weighting factor. The value of the dynamic weighting factor depends on the road surface bump index of the current road segment; that is, if the road surface bump index is 10, then... The current disturbance is severe, and the measured and estimated values ​​are ( This better reflects the true acceleration and deterioration, and is given a higher weight. If the road bump index is 8, then... If the road surface bump index is 4, then If the road surface bump index is 1, then The environment is stable, and historical statistics ( The reliability of the measured data is relatively high, so the weight of the measured data should be appropriately reduced. If the difference between the basic remaining shelf life and the average remaining shelf life is less than or equal to the preset deviation threshold, the basic remaining shelf life will be determined as the final estimated remaining shelf life. That is, the deterioration rate of the current batch is basically consistent with the historical average level, and the measured estimated results can be directly trusted.

[0126] In this way, the degradation of fresh food quality can be accurately quantified from multiple dimensions. By using a preset deviation threshold judgment mechanism, the basic remaining shelf life and the statistical average remaining shelf life are integrated to determine the final estimated remaining shelf life. The weights are adaptively adjusted, increasing the weight of the basic remaining shelf life when there is severe disturbance, and emphasizing the reference value of the statistical average remaining shelf life when the environment is stable. This reduces the deviation of single real-time data due to occasional sensor errors, and corrects the lag of pure historical statistics that cannot adapt to the special working conditions of the current batch, greatly improving the stability of the estimated results in extreme scenarios.

[0127] According to one embodiment of the present invention, in step S8, a dynamic risk coefficient is determined based on the final estimated remaining shelf life and the delivery time at the current collection time.

[0128] Figure 7 A flowchart for calculating dynamic risk coefficients according to an embodiment of the present invention is shown as an example.

[0129] According to an embodiment of the present invention, step S8 includes: step S81, obtaining the highest temperature value and average temperature value of historical collection times within a preset historical time window before the current collection time; step S84, using the ratio of the highest temperature value to the average temperature value as a temperature peak factor; step S83, multiplying the ratio of the delivery time at the current collection time to the final estimated remaining shelf life by the temperature peak factor to determine the dynamic risk coefficient.

[0130] According to one embodiment of the present invention, the highest and average temperature values ​​of historical collection times within a preset historical time window (e.g., the past 30 minutes) prior to the current collection time are obtained. The ratio of the delivery time at the current collection time to the final estimated remaining shelf life is multiplied by a temperature peak factor to determine the dynamic risk coefficient, i.e., according to the formula: Determine the dynamic risk coefficient at the current data collection moment. ,in, This represents the delivery time since the current collection time. This is the final estimated remaining shelf life at the current collection time. This refers to the highest temperature value from historical data collection times within a preset historical time window prior to the current data collection time. This is the average temperature value of historical data collected within a preset historical time window prior to the current data collection time. The temperature peak factor represents the additional impact damage to quality caused by recent instantaneous extreme high temperatures. The higher the temperature peak factor, the higher the risk level of the storage environment during the delivery process. This is the ratio of the delivery time at the current collection moment to the estimated remaining shelf life, representing the cumulative quality loss over time. The greater this cumulative quality loss, the higher the risk. and Multiplying the two results in a dynamic risk coefficient. The higher the dynamic risk coefficient, the higher the risk level of the storage environment during the delivery process.

[0131] In this way, dynamic risk coefficients can be determined by the cumulative consumption of quality over time and the impact of instantaneous high temperatures, so as to more comprehensively assess the risk level of the storage environment during the delivery process. By amplifying the damage weight of instantaneous extreme high temperatures through the temperature peak factor, the risk of sudden temperature changes in a short period of time, which is completely covered by traditional average temperature indicators, can be accurately captured, reducing the hidden dangers of rapid spoilage of fresh food in the cold chain due to short-term high temperatures.

[0132] According to one embodiment of the present invention, in step S9, a graded early warning output is performed based on the dynamic risk coefficient and the virtual steady-state environment vector.

[0133] Figure 8 A flowchart of the graded early warning output according to an embodiment of the present invention is shown as an example.

[0134] According to an embodiment of the present invention, step S9 includes: step S91, mapping the dynamic risk coefficient to generate a multi-level cold chain risk status code, wherein the multi-level cold chain risk status code includes a safety level code, a concern level code, and a critical level code; step S92, calculating the vector difference between the virtual steady-state environment vector at the current acquisition time and the virtual steady-state environment vector at the previous acquisition time; step S93, performing time differentiation processing on the L2 norm of the vector difference to obtain an environmental mutation trend feature; step S94, if the environmental mutation trend feature is greater than a preset mutation alarm threshold, then forcibly setting the multi-level cold chain risk status code to the critical level code; step S95, if the environmental mutation trend feature is less than or equal to the preset mutation alarm threshold, then performing a graded early warning output based on the multi-level cold chain risk status code.

[0135] According to one embodiment of the present invention, a multi-level cold chain risk status code is generated by mapping the dynamic risk coefficient. Specifically, if the dynamic risk coefficient is less than a first risk threshold (e.g., 0.5), a safety-level code (00) is generated; if the dynamic risk coefficient is greater than or equal to the first risk threshold and less than a second risk threshold (e.g., 0.9), a concern-level code (01) is generated; and if the dynamic risk coefficient is greater than or equal to the second risk threshold, a critical-level code (10) is generated. The L2 norm of the vector difference is then subjected to time differentiation to obtain the environmental mutation trend characteristic, i.e., according to the formula: To obtain the characteristic quantity of environmental mutation trend. ,in, This represents the virtual steady-state environment vector at the current acquisition moment. This is the virtual steady-state environment vector from the previous acquisition time. This represents the time interval between the current data collection time and the previous data collection time. It is a 2-norm. This is the vector difference between the virtual steady-state environment vector at the current acquisition time and the virtual steady-state environment vector at the previous acquisition time. The larger this environmental mutation trend feature, the higher the degree of mutation in the storage environment during the delivery process. If the environmental mutation trend feature is greater than the preset mutation alarm threshold (e.g., 0.15 / s), the multi-level cold chain risk status code is forcibly set to the critical level code to achieve a risk level upgrade under extreme environmental mutation scenarios. If the environmental mutation trend feature is less than or equal to the preset mutation alarm threshold, a graded early warning output is performed based on the multi-level cold chain risk status code. The graded early warning output includes: when the multi-level cold chain risk status code is a safety level code, a normal status message is generated and sent to the remote monitoring center at a first frequency (e.g., once every 5 minutes). The message content only includes a timestamp, coordinates, multi-level cold chain risk status code, and temperature summary. When the multi-level cold chain risk status code is a concern level code, a quality concern message is generated and sent to the remote monitoring center at a second frequency (e.g., once every 1 minute). The message content adds a dynamic risk coefficient and a suggestion for priority delivery. When the multi-level cold chain risk status is coded to the critical level, an emergency intervention message is generated and sent to the remote monitoring center at a third frequency (e.g., once every 10 seconds). The message content includes the final estimated remaining shelf life, and at the same time, the local audible and visual alarm device (80-decibel buzzer + red flashing LED) is activated.

[0136] The method for monitoring the storage environment during the cold chain distribution of fresh food according to embodiments of the present invention, by adjusting the sampling frequency through the road bump index, can reduce the computing power required for sampling in low-disturbance scenarios and improve the timeliness of anomaly detection in bumpy road sections. By comprehensively considering the disturbance index, the non-steady-state degree of the cold chain environment can be quantified, eliminating interference from misjudgments of single environmental parameters. By correcting the deviations caused by measured environmental data under non-steady-state disturbances through a virtual steady-state environment vector, the final calculated remaining shelf life more closely reflects the actual quality decay law of fresh food, thus reflecting the true steady-state environment in which the fresh food exists. Based on dynamic risk coefficients and virtual steady-state environment vectors, a tiered early warning system can be implemented, pushing different handling strategies according to risk levels, reducing invalid alarms, significantly improving the efficiency of cold chain anomaly handling, and reducing the loss rate of fresh food. When determining the comprehensive disturbance index, active disturbances can be precisely bound to the temperature-sensitive dimension, and passive disturbances to the humidity dimension. This reduces assessment bias caused by mismatches between different disturbance sources and environmental parameters, aligning with the actual operating conditions of the cold chain. A new cross-coupling term for the product of active and passive disturbances is added to accurately recreate the nonlinear multiplication effect of characteristic gas diffusion when two disturbances act simultaneously. The comprehensive disturbance index can quantify the overall unsteady state of the cold chain environment under multi-source excitation, improving the accuracy of unsteady state assessment. When determining the final estimated remaining shelf life, the degradation of fresh food quality can be precisely quantified across multiple dimensions. A preset deviation threshold judgment mechanism integrates the basic remaining shelf life and the statistical average remaining shelf life to determine the final estimated remaining shelf life. Weights are adaptively adjusted: the basic remaining shelf life is given greater weight when disturbances are severe, while the statistical average remaining shelf life is emphasized when the environment is stable. This reduces the bias caused by occasional sensor errors in single real-time data and corrects the lag in adapting purely historical statistics to the specific operating conditions of the current batch, significantly improving the stability of the estimated results under extreme scenarios. When determining the dynamic risk coefficient, the cumulative consumption of quality over time and the impact of instantaneous high temperature can be used to determine the dynamic risk coefficient, so as to more comprehensively assess the risk level of the storage environment during the delivery process. By amplifying the damage weight of instantaneous extreme high temperature through the temperature peak factor, the risk of short-term temperature change that is completely covered by traditional average temperature indicators can be accurately captured, reducing the hidden dangers of short-term high temperature rapid spoilage in the fresh food cold chain.

[0137] Figure 9 An exemplary block diagram of a fresh food cold chain distribution process storage environment monitoring system according to an embodiment of the present invention is shown, the system comprising:

[0138] A heterogeneous sensor network module is used to deploy multiple wireless sensor nodes in a preset spatial distribution within the storage space of cold chain distribution to construct a heterogeneous sensor network, wherein the heterogeneous sensor network includes a temperature sensing subnet, a humidity sensing subnet, and a volatile gas sensing subnet.

[0139] The sampling frequency module is used to obtain the road surface bump index of the road segment based on the real-time spatial coordinates of the delivery unit at the beginning of each sampling cycle, and to determine the sampling frequency of the sampling cycle.

[0140] The standardized environmental state vector module is used to synchronously collect a set of environmental data from the heterogeneous sensor network according to the acquisition frequency, and generate a standardized environmental state vector at the current acquisition time, wherein the environmental data includes temperature, humidity and characteristic gas concentration;

[0141] The delivery unit data module is used to obtain the operating power of the refrigeration equipment and the fan speed inside the delivery unit at the current data collection time.

[0142] The comprehensive disturbance index module is used to determine a comprehensive disturbance index that measures the degree of unsteadiness of the cold chain environment under multi-source excitation based on the road surface bump index, the operating power, the fan speed and the standardized environmental state vector.

[0143] The virtual steady-state environment vector module is used to obtain the virtual steady-state environment vector based on the comprehensive disturbance index and the standardized environment state vector.

[0144] The final estimated remaining shelf life module is used to determine the final estimated remaining shelf life based on the virtual steady-state environment vector.

[0145] The dynamic risk coefficient module is used to determine the dynamic risk coefficient based on the final estimated remaining shelf life and the delivery time at the current collection time.

[0146] The graded early warning output module is used to output graded early warnings based on the dynamic risk coefficient and the virtual steady-state environment vector.

[0147] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0148] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and explained in the embodiments, and any variations or modifications may be made to the implementation of the present invention without departing from the stated principles.

Claims

1. A method for monitoring the storage environment during the cold chain distribution process of fresh food ingredients, characterized in that, include: In the storage space of cold chain distribution, multiple wireless sensor nodes are deployed according to a preset spatial distribution to construct a heterogeneous sensor network, wherein the heterogeneous sensor network includes a temperature sensor subnet, a humidity sensor subnet, and a volatile gas sensor subnet. At the beginning of each collection cycle, the road surface bump index of the road segment is obtained based on the real-time spatial coordinates of the delivery unit, and the collection frequency of the collection cycle is determined. According to the acquisition frequency, a set of environmental data is synchronously acquired from the heterogeneous sensor network to generate a standardized environmental state vector at the current acquisition time, wherein the environmental data includes temperature, humidity and characteristic gas concentration; Obtain the operating power of the refrigeration equipment and the fan speed inside the delivery unit at the current data collection time; Based on the road surface bump index, the operating power, the fan speed, and the standardized environmental state vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness of the cold chain environment under multi-source excitation. Based on the comprehensive disturbance index and the standardized environmental state vector, the virtual steady-state environment vector is obtained; Based on the virtual steady-state environment vector, the final estimated remaining shelf life is determined; Based on the final estimated remaining shelf life and the delivery time at the current collection time, a dynamic risk coefficient is determined. Based on the dynamic risk coefficient and the virtual steady-state environment vector, a graded early warning output is performed.

2. The fresh food cold chain distribution process storage environment monitoring method of claim 1, wherein, At the beginning of each data collection cycle, the road surface bump index is obtained based on the real-time spatial coordinates of the delivery unit, and the data collection frequency for the cycle is determined, including: Based on the real-time spatial coordinates of the delivery unit, query the cloud-based road condition database to obtain the road surface bump index of the road segment; If the road surface bump index is greater than the first bump threshold, then the sampling frequency of the sampling period is the first sampling frequency. If the road surface bump index is less than or equal to the first bump threshold and greater than the second bump threshold, then the sampling frequency of the sampling period is the second sampling frequency. If the road surface bump index is less than or equal to the second bump threshold, then the sampling frequency of the sampling cycle is the third sampling frequency.

3. The fresh food cold chain delivery process storage environment monitoring method of claim 1, wherein, Based on the acquisition frequency, a set of environmental data is synchronously acquired from the heterogeneous sensor network to generate a standardized environmental state vector at the current acquisition time, including: Extract historical environmental data up to a preset sliding window length prior to the current acquisition time, wherein the historical environmental data includes historical temperature, historical humidity, and historical characteristic gas concentration; Calculate the mean and standard deviation of each historical environment data within a preset sliding window length; If the absolute value of the average value of a certain environmental data at the current collection time minus the corresponding historical environmental data is greater than the product of a preset coefficient and the standard deviation of the corresponding historical environmental data, then the corresponding environmental data at the current collection time will be removed and replaced with the historical environmental data corresponding to the previous collection time. If the absolute value of the average value of the corresponding historical environmental data minus the current environmental data is less than or equal to the product of the preset coefficient and the standard deviation of the corresponding historical environmental data, then the corresponding environmental data at the current collection time is retained. The environmental data collected at the current time, after being discarded or retained, are used to form a standardized environmental state vector, which consists of three components: standard temperature, standard humidity, and standard characteristic gas concentration.

4. The fresh food cold chain delivery process storage environment monitoring method of claim 1, wherein, Based on the road surface bump index, the operating power, the fan speed, and the standardized environmental state vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness in the cold chain environment under multi-source excitation, including: Obtain the reference values ​​for the operating power of the refrigeration equipment and the reference values ​​for the fan speed inside the compartment; The ratio of the operating power to the operating power reference value is taken as the first ratio. The ratio of the fan speed to the fan speed reference value is used as the second ratio. The active disturbance coefficient at the current acquisition time is obtained by summing and averaging the first ratio and the second ratio. The passive disturbance coefficient at the current acquisition time is obtained based on the ratio of the road surface bump index to the full value of the road surface bump index. The difference vector between the standardized environmental state vector and the standardized environmental state vector at the previous acquisition time is taken as the fluctuation vector, wherein the fluctuation vector consists of three components: fluctuation temperature, fluctuation humidity, and fluctuation characteristic gas concentration. Based on the active disturbance coefficient, the passive disturbance coefficient, and the fluctuation vector, a comprehensive disturbance index is determined to measure the degree of unsteadiness of the cold chain environment under multi-source excitation.

5. The fresh food cold chain delivery process storage environment monitoring method of claim 1, wherein, Based on the comprehensive disturbance index and the standardized environmental state vector, a virtual steady-state environment vector is obtained, including: When the comprehensive disturbance index is greater than the disturbance threshold, the historical standardized environmental state vector of the M consecutive historical acquisition times before the current acquisition time is obtained. The historical standardized environmental state vector consists of three components: historical standard temperature, historical standard humidity and historical standard characteristic gas concentration, where M is a positive integer. For each dimension of the historical standardized environmental state vector, calculate the average slope of change over M intervals to generate the slope vector at the current acquisition time. The slope vector consists of three components: the average slope of change in the temperature dimension, the average slope of change in the humidity dimension, and the average slope of change in the characteristic gas concentration dimension. The ratio of the comprehensive disturbance index to the sum of 1 and the comprehensive disturbance index is used as the scaling factor for the current acquisition time. The compensation vector at the current acquisition moment is obtained by multiplying the negative value of the scaling factor with the slope vector. The standardized environmental state vector and the compensation vector are added together to obtain the compensated virtual steady-state environmental vector, wherein the virtual steady-state environmental vector consists of three components: steady-state temperature, steady-state humidity, and steady-state characteristic gas concentration. When the comprehensive disturbance index is less than or equal to the disturbance threshold, the standardized environmental state vector is used as the virtual steady-state environmental vector.

6. The fresh food cold chain delivery process storage environment monitoring method of claim 1, wherein, Based on the virtual steady-state environment vector, the final estimated remaining shelf life is determined, including: Based on the steady-state temperature and steady-state humidity in the virtual steady-state environment vector, the decay rate constant of the microbial growth dimension is obtained by querying the first rate constant mapping table; Based on the steady-state temperature and steady-state characteristic gas concentration in the virtual steady-state environment vector, the lipid oxidation dimension decay rate constant is obtained by querying the second rate constant mapping table. Based on the steady-state temperature, steady-state characteristic gas concentration, and steady-state characteristic gas concentration in the virtual steady-state environment vector, the decay rate constant of the water activity dimension is obtained by querying the third rate constant mapping table. Based on the decay rate constants of the microbial growth dimension, the lipid oxidation dimension, and the water activity dimension, the cumulative quality degradation that has occurred from the start of delivery to the current collection time is obtained. Starting from the current collection time, and using the preset estimated time step as the increment, a forward loop iterative calculation is performed on the time axis. In each iteration, the decay rate constants of the microbial growth dimension, lipid oxidation dimension, and water activity dimension at the current collection time are used as the decay rate constants of the three dimensions at the iteration time, and a deterioration increment of a preset estimated time step is calculated. The predicted cumulative quality degradation is obtained based on the cumulative quality degradation, the degradation increment, and the preset estimated time step. The iteration terminates when the predicted cumulative quality degradation first exceeds or equals the preset unacceptable quality threshold. Record the total number of steps in the iterative calculation, and multiply the total number of steps by the calculation time step to obtain the basic remaining shelf life; The total shelf life of all goods that match the current delivery type, origin and initial quality level in the historical cold chain distribution database is selected and averaged to obtain the statistical average shelf life. Subtract the delivery time at the current collection time from the statistical average shelf life to obtain the average remaining shelf life; If the difference between the base remaining shelf life and the average remaining shelf life is greater than a preset deviation threshold, then the weighted average of the base remaining shelf life and the average remaining shelf life is determined as the final estimated remaining shelf life. If the difference between the basic remaining shelf life and the average remaining shelf life is less than or equal to a preset deviation threshold, then the basic remaining shelf life is determined as the final estimated remaining shelf life.

7. The method for monitoring the storage environment during the cold chain distribution of fresh food according to claim 1, characterized in that, Based on the final estimated remaining shelf life and the delivery time at the current collection time, a dynamic risk coefficient is determined, including: Obtain the highest and average temperature values ​​from historical data collection times within a preset historical time window prior to the current data collection time. The ratio of the highest temperature value to the average temperature value is used as the temperature peak factor; The dynamic risk coefficient is determined by multiplying the ratio of the delivery time at the current collection moment to the final estimated remaining shelf life by the temperature peak factor.

8. The method for monitoring the storage environment during the cold chain distribution of fresh food according to claim 1, characterized in that, Based on the dynamic risk coefficient and the virtual steady-state environment vector, a tiered early warning output is performed, including: The dynamic risk coefficient is mapped to generate a multi-level cold chain risk status code, wherein the multi-level cold chain risk status code includes a safety level code, a concern level code, and a critical level code; Calculate the vector difference between the virtual steady-state environment vector at the current acquisition time and the virtual steady-state environment vector at the previous acquisition time; By performing time differentiation on the L2 norm of the vector difference, the characteristic quantity of environmental mutation trend can be obtained; If the environmental mutation trend feature is greater than the preset mutation alarm threshold, the multi-level cold chain risk status code will be forcibly set to the critical level code. If the environmental mutation trend feature is less than or equal to the preset mutation alarm threshold, then a graded early warning output is performed based on the multi-level cold chain risk status code.

9. A system for monitoring the storage environment during the cold chain distribution of fresh ingredients, used to execute the method for monitoring the storage environment during the cold chain distribution of fresh ingredients as described in any one of claims 1-8, characterized in that, include: A heterogeneous sensor network module is used to deploy multiple wireless sensor nodes in a preset spatial distribution within the storage space of cold chain distribution to construct a heterogeneous sensor network, wherein the heterogeneous sensor network includes a temperature sensing subnet, a humidity sensing subnet, and a volatile gas sensing subnet. The sampling frequency module is used to obtain the road surface bump index of the road segment based on the real-time spatial coordinates of the delivery unit at the beginning of each sampling cycle, and to determine the sampling frequency of the sampling cycle. The standardized environmental state vector module is used to synchronously collect a set of environmental data from the heterogeneous sensor network according to the acquisition frequency, and generate a standardized environmental state vector at the current acquisition time, wherein the environmental data includes temperature, humidity and characteristic gas concentration; The delivery unit data module is used to obtain the operating power of the refrigeration equipment and the fan speed inside the delivery unit at the current data collection time. The comprehensive disturbance index module is used to determine a comprehensive disturbance index that measures the degree of unsteadiness of the cold chain environment under multi-source excitation based on the road surface bump index, the operating power, the fan speed and the standardized environmental state vector. The virtual steady-state environment vector module is used to obtain the virtual steady-state environment vector based on the comprehensive disturbance index and the standardized environment state vector. The final estimated remaining shelf life module is used to determine the final estimated remaining shelf life based on the virtual steady-state environment vector. The dynamic risk coefficient module is used to determine the dynamic risk coefficient based on the final estimated remaining shelf life and the delivery time at the current collection time. The graded early warning output module is used to output graded early warnings based on the dynamic risk coefficient and the virtual steady-state environment vector.