Stem cell transportation remote management system based on environment prediction

Abstract

The invention discloses a stem cell transportation remote management system based on environmental prediction, and relates to the field of biomedical transportation, and the system comprises a visualization module which discretizes alternative paths into path units according to a comprehensive data set, marks spatial-temporal characteristics for each path unit by using a spatial-temporal data association method, and stores the spatial-temporal characteristics; a stem cell characteristic enhancement mechanism is utilized to set the weight of the spatial-temporal characteristics, a biological-environment coupling thermodynamic diagram and path unit characteristic data are generated, an evaluation module is used for constructing a stem cell transportation risk model, and the environment interaction between the biological-environment coupling thermodynamic diagram and the path unit characteristic data is analyzed; and obtaining a path unit risk feature set, calculating the risk score of the alternative path, and outputting a path risk assessment report. According to the method, the injury risk during stem cell transportation is reduced, and sudden decrease of the cell survival rate under composite environmental factors is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of biomedical transportation, in particular to a stem cell transportation remote management system based on environmental prediction. BACKGROUND

[0002] The development of stem cell transportation technology has gone through a process from basic transportation equipment to intelligent management. Early stem cell transportation mainly relied on cold chain logistics, using constant temperature transportation boxes and manual monitoring equipment, and recording temperature and humidity by artificial to ensure cell survival rate. The introduction of vehicle-mounted sensors and GIS improves the accuracy of transportation environment monitoring. Models such as multilayer perception based on machine learning are also applied to transportation risk prediction, which optimizes path selection by analyzing the relationship between environmental factors and cell survival rate.

[0003] The existing stem cell transportation remote management system has several points that need to be improved. Conventional stem cell transportation mostly monitors single environmental factors, lacks integration and interactive analysis of multi-dimensional environmental data, and the comprehensiveness of risk assessment is insufficient. The selection of transportation path is static planning, which cannot dynamically respond to real-time weather warning and feedback of biological sensors, resulting in low cell survival rate of stem cells during transportation. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a stem cell transportation remote management system based on environmental prediction, which solves the problems of insufficient interaction analysis of environmental factors and lack of dynamic adjustment of transportation path in existing stem cell transportation.

[0006] To solve the above technical problems, the present application provides the following technical solutions: The present application provides a stem cell transportation remote management system based on environmental prediction, which includes a data acquisition module, a dispatch center selects a candidate path, acquires transportation environment data and pre-processes to obtain a comprehensive data set; A visualization module, according to the comprehensive data set, discretizes the candidate path into path units, uses a space-time data association method to label the space-time features of each path unit, and uses a stem cell feature enhancement mechanism to set the weight of the space-time features, generates a biological-environment coupling heat map and path unit feature data; An evaluation module, a stem cell transportation risk model is constructed, the environmental interaction between the biological-environment coupling heat map and the path unit feature data is analyzed, the path unit risk feature set is obtained and the risk score of the candidate path is calculated, and a path risk assessment report is output; An output module, according to the path risk assessment report, selects the path with the lowest risk as the transportation path, outputs the final navigation path and three-dimensional visual navigation interface.

[0007] As a preferred scheme of the stem cell transportation remote management system based on environment prediction, the transportation environment data comprises road flatness data, precipitation probability distribution, temperature variation curve, wind speed data, road surface vibration data and historical road accident data.

[0008] As a preferred scheme of the stem cell transportation remote management system based on environment prediction, the obtaining of the comprehensive data set comprises the following steps. The road flatness data is denoised, the missing values in the precipitation probability distribution and the wind speed data are filled by using linear interpolation, the temperature mutation points in the temperature variation curve are identified and removed, and the historical road accident data and the road surface vibration data are recorded and screened to output the cleaned transportation environment data. The cleaned transportation environment data is standardized, normalized and spatio-temporally aligned to obtain the comprehensive data set.

[0009] As a preferred scheme of the stem cell transportation remote management system based on environment prediction, the discretization of the alternative path into path units comprises the following steps. The geographic coordinates of the alternative path points in the comprehensive data set are selected as coordinate points, the spherical distance between adjacent coordinate points is calculated using the secant formula to obtain the total length of the alternative path. The fixed interval of the alternative path is set and the number of path points of the alternative path is calculated based on the total length of the alternative path. The distance of the alternative path is accumulated from the starting point of the alternative path, and when the accumulated distance of the alternative path is greater than or equal to the fixed interval of the alternative path, the division point of the alternative path is determined. According to the number of path points of the alternative path and the division point of the alternative path, each alternative path is divided into path units.

[0010] As a preferred scheme of the stem cell transportation remote management system based on environment prediction, the labeling of the spatio-temporal features for each path unit means that the altitude change rate, the average number of road surface vibrations and the average number of accidents of the path unit are calculated respectively using the road flatness data, the road surface vibration data and the historical road accident data in the comprehensive data set, and are marked as spatio-temporal features.

[0011] As a preferred scheme of the stem cell transportation remote management system based on environment prediction, the generation of the biological-environment coupling thermodynamic map and the path unit feature data comprises the following steps. A biological tolerance threshold is set, the temperature fluctuation amplitude and the road surface vibration frequency of the transportation box are collected by using a biological sensor and compared with the biological tolerance threshold to obtain a stem cell survival rate risk judgment result. According to the risk judgment result of stem cell survival rate, corresponding weights are set for the space-time characteristics to generate path unit feature data and calculate the comprehensive risk value of each path unit; The comprehensive risk value of each path unit is mapped into the geographic space of the alternative path using the Kriging spatial interpolation method, and the risk level is labeled to form a two-dimensional heat map; Set the risk adjustment rule and adjust the risk level in the two-dimensional heat map to generate a bio-environment coupling heat map.

[0012] As a preferred scheme of the stem cell transportation remote management system based on environmental prediction, wherein: the stem cell transportation risk model is constructed by selecting a multi-layer perception machine as the core architecture, collecting historical stem cell transportation data and preprocessing, and training the multi-layer perception machine using the preprocessed historical stem cell transportation data to obtain the stem cell transportation risk model.

[0013] As a preferred scheme of the stem cell transportation remote management system based on environmental prediction, wherein: the path unit risk feature set is obtained, specifically including the following steps, Set the interaction intensity threshold, calculate the environmental interaction intensity value using the stem cell transportation risk model, and compare it with the interaction intensity threshold to judge the high-risk combination feature, and output the interaction data set; According to the risk judgment result of stem cell survival rate, set the vibration sensitivity coefficient and adjust the weight of the average vibration frequency of the path unit road surface; Integrate the path unit feature data, the weight of the average vibration frequency of the adjusted path unit road surface, and the interaction data set into JSON format to form the path unit risk feature set.

[0014] As a preferred scheme of the stem cell transportation remote management system based on environmental prediction, wherein: the output path risk assessment report specifically includes the following steps, Using the path unit risk feature set, calculate the vibration cumulative value, temperature fluctuation standard deviation and risk period coincidence degree of each alternative path, and after normalization processing, input into the stem cell transportation risk model to generate the comprehensive risk score of the alternative path; Sort the comprehensive risk score of each alternative path, and select the path with the lowest comprehensive risk score as the candidate path; Record the geographic coordinates and environmental factors of the path unit risk area in the bio-environment coupling risk heat map to generate avoidance measure suggestions; Integrate the comprehensive risk score of the alternative path, the avoidance measure suggestions and the candidate path into PDF format to output the path risk assessment report.

[0015] As a preferred scheme of the stem cell transportation remote management system based on environment prediction, the output final navigation path and three-dimensional visual navigation interface specifically comprises the following steps, Calculate the total length and predicted transportation time of the candidate path, and mark the turning points of the candidate path to generate turning point navigation instructions; Integrate the total length, predicted transportation time and turning point navigation instructions of the candidate path to form navigation path data; Collect the geographic information of the refrigeration supply station and the weather warning information of the navigation path, and use the Cesium three-dimensional rendering engine to mark the geographic information of the navigation path and the weather warning information of the navigation path. Generate a three-dimensional visual navigation page.

[0016] The present application has the advantages that: by setting different vibration sensitivity coefficients and real-time feedback of the transportation box biological sensor, the weight of the parameters in the path characteristic data is dynamically adjusted, the biological characteristics of stem cells are converted into quantitative indicators for risk assessment, the drawbacks of ignoring biological characteristics and environmental coupling in conventional stem cell transportation path planning are overcome, the risk assessment is changed from general cold chain transportation to stem cell specific protection, and the risk of stem cell damage is greatly reduced. And by analyzing the nonlinear relationship between precipitation probability and vibration frequency through a multilayer perception machine, the weight of the vibration frequency is automatically increased, and the generated high-risk combination features break through the limitations of manually setting weights, and key risk combinations are found through machine learning. Avoid sudden drops in cell survival rate under complex environmental factors. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0018] Fig. 1 The flowchart of the stem cell transportation remote management system.

[0019] Fig. 2 The flowchart of path unit discretization and thermal map generation.

[0020] Fig. 3 The flowchart of constructing a stem cell transportation risk model.

[0021] Fig. 4 The schematic diagram of generating a three-dimensional visual navigation page. DETAILED DESCRIPTION

[0022] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0023] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. The present application, however, can be practiced in a variety of ways other than those specifically described herein, and the present application can be practiced with other apparatus than those explicitly described herein, and by individuals other than those who have been described using the present application. Accordingly, the present application is not intended to be limited to the specific embodiments disclosed below, which are included as examples only.

[0024] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an embodiment that is separate or alternative to other embodiments.

[0025] Reference Figs. 1-4 For one embodiment of the present application, the embodiment provides a stem cell transportation remote management system based on environmental prediction, comprising the following steps: A data acquisition module, the dispatch center selects alternative paths, acquires transportation environment data and pre-processes to obtain a comprehensive data set.

[0026] Comprising the following steps, The alternative path refers to the multiple feasible transportation routes from the starting point to the ending point provided by the dispatch center at the beginning of the stem cell transportation task, mainly including highways, rural roads or paths around specific areas. The dispatch center usually provides 2-5 alternative paths to balance transportation efficiency and risk avoidance. Each alternative path contains complete geographic coordinate points from the transportation starting point to the ending point.

[0027] The collected transportation environment data refers to the transportation environment data on the alternative path, including road flatness data, precipitation probability distribution map, temperature change curve, wind speed data, road surface vibration data and historical road accident data. The specific collection process is: using the vehicle-mounted laser radar equipment to scan the road surface on each alternative path at high frequency, generating three-dimensional point cloud data, i.e. road flatness data, recording the characteristics of road surface pits, cracks and elevation slope, etc. Through the API interface of the meteorological bureau, the precipitation probability distribution map, temperature change curve and wind speed data of the alternative path along the way in the future period of time are obtained, and each geographic coordinate point on the alternative path is covered. Extract the typical accident coordinate points related to the transportation season from the historical database, such as the geographic location and frequency of winter bridge icing accidents or summer high temperature equipment failure.

[0028] The transportation environment data is pre-processed.

[0029] Specifically, the three-dimensional point cloud data is denoised to remove abnormal points caused by device jitter or external interference, such as removing three-dimensional point cloud data deviating from the normal height of the road surface. The precipitation probability distribution map and wind speed provided by the meteorological bureau API are checked for missing values, and linear interpolation is used to fill in the missing values, such as when a certain candidate path point has missing wind speed data, the average value of the previous and next time points is used to fill in. Identify temperature mutation points in the temperature change curve, such as temperature change exceeding 5°C within 1 hour, which is beyond the reasonable range of stem cell transportation, and use the median filter to remove. The historical road accident data and road surface vibration data extracted from the historical database are screened for redundant records, and invalid records such as values exceeding the range of the vibration sensor are deleted, and the cleaned transportation environment data is output.

[0030] The cleaned transportation environment data is standardized, normalized and spatio-temporally aligned, and a comprehensive data set is output.

[0031] Specifically, all cleaned transportation environment data is converted to a unified structured format, stored in JSON format, including fields such as road flatness represented by the standard deviation of three-dimensional point cloud data, precipitation probability distribution represented as a percentage, and wind speed represented as meters per second. Normalization is performed using the min-max normalization method, which maps all cleaned data to the value range of 0-1.

[0032] The normalized transportation environment data is spatio-temporally aligned.

[0033] Specific operations are: using GIS (Geographic Information System), matching the geographic coordinate points of each candidate path with the cleaned and normalized transportation environment data, ensuring that each candidate path point corresponds to a unique type of transportation environment data. Through time stamp alignment, the precipitation probability distribution map, temperature change curve and wind speed data are associated with the seasonal information of the historical road accident data to the real-time time window of the candidate path point. The aligned transportation environment data is stored in JSON format to generate a comprehensive data set, including the geographic coordinates of each candidate path point, environmental parameters (road flatness, precipitation probability, temperature, wind speed, road surface vibration) and risk labels (accident frequency).

[0034] The visualization module discretizes the candidate path into path units based on the comprehensive data set, labels the spatio-temporal features of each path unit using spatio-temporal data association method, and sets the weight of the spatio-temporal features using stem cell feature enhancement mechanism to generate bio-environment coupling heat map and path unit feature data.

[0035] The method comprises the following steps, The geographic coordinates of the candidate path points are selected from the comprehensive data set as coordinate points, and the spherical distance between adjacent coordinate points is calculated using the haversine formula to obtain the total length of the candidate path.

[0036] In particular, the haversine formula is used to calculate the shortest path distance between any two points on the earth's surface, i.e. the great circle distance, and the formula is: ; ; ; wherein, represents the haversine value of the spherical angle between adjacent coordinate points, represents the latitude of the first coordinate point, represents the latitude of the second coordinate point, represents the latitude difference between the first coordinate point and the second coordinate point, represents the longitude difference between the first coordinate point and the second coordinate point, represents the spherical angle between the first coordinate point and the second coordinate point, represents the spherical distance between the first coordinate point and the second coordinate point, represents the average radius of the earth.

[0037] Example: Assuming that one coordinate point on a candidate path is (31°, 121°) and the second coordinate point is (31.1°, 121.1°), the spherical distance between the two coordinate points is calculated using the haversine formula to be approximately 16.68 meters. The total length of the candidate path is obtained by repeating the calculation for all adjacent coordinate points on the candidate path and adding them up.

[0038] The fixed interval of the candidate path is set according to the actual requirements of the stem cell transportation task, such as calculation accuracy and data processing capacity.

[0039] The number of path points of the candidate path is calculated according to the total length of the candidate path, i.e. by dividing the total length of the candidate path by the fixed interval of the candidate path. For example, if the total length of the candidate path is 50 kilometers and the fixed interval is 500 meters, then 100 path points are needed.

[0040] According to the fixed interval of the candidate path, the distance of the candidate path is accumulated from the starting point of the candidate path. When the accumulated distance of the candidate path is greater than or equal to the fixed interval of the candidate path, the division point of the candidate path is obtained. For example, if the fixed interval of the candidate path is 500 meters, the distance is accumulated from the starting point of the candidate path to 500 meters, and the corresponding geographic coordinate point is recorded as the first division point. Then continue to accumulate to 1000 meters, mark the second division point, and so on.

[0041] According to the split points of the alternative paths and the path point numbers of the alternative paths, each alternative path is divided into continuous path units.

[0042] Specifically, taking each split point as the starting point of the path unit and the next split point as the termination point of the path unit, a corresponding number of path units are generated, for example, 100 units correspond to 99 split points plus one path endpoint, and each path unit records the starting coordinates, termination coordinates, center point coordinates and corresponding environmental parameters in the comprehensive data set.

[0043] The path units are labeled with spatiotemporal features.

[0044] Specifically, the road flatness data in the comprehensive data set is used to calculate the elevation change rate of the path unit. For example, the elevation slope of the starting point of an alternative path point in a path unit is 100 meters, the terminal elevation slope is 120 meters, and the length of the path unit is 500 meters, then the elevation change rate of the path unit is 4%.

[0045] According to the road surface vibration data in the comprehensive data set, the average road surface vibration frequency of the path unit is calculated. That is, according to the road surface vibration data, the annual average vibration frequency of the road section in each path unit can be obtained.

[0046] According to the historical road accident data in the comprehensive data set, the average accident occurrence frequency in the path unit is calculated. That is, the typical accident coordinate points related to the transportation season are extracted from the comprehensive data set, the accident occurrence frequency of each path unit coverage area is counted, and the annual average frequency is converted, for example, a path unit coverage area has occurred 5 accidents in the past 10 years, then the annual average frequency is 0.5 times / year.

[0047] The elevation change rate of the path unit, the average road surface vibration frequency of the path unit and the average accident occurrence frequency in the path unit are marked as spatiotemporal features.

[0048] The weights of the spatiotemporal features are set by using the stem cell feature enhancement mechanism to generate path unit feature data.

[0049] Specifically, a biosensor is deployed in the transport box of stem cells, which collects the amplitude of temperature fluctuation inside the transport box and the real-time road vibration frequency of the transport box during transportation in real time, and sets a biological tolerance threshold according to the sensitivity of different stem cell types to vibration frequency and temperature fluctuation. Stem cell types are divided into neural stem cells and mesenchymal stem cells, and the biological tolerance threshold is divided into a vibration tolerance threshold and a temperature tolerance threshold. The vibration tolerance threshold is for neural stem cells, and the value can be set to 3-5 Hz, with 3 Hz as the lower limit. Vibration below 3 Hz, such as slight shaking of flat asphalt roads, has negligible impact on the survival rate of neural stem cells. 5 Hz is the upper limit, and vibration above 5 Hz, such as gravel roads or sharp turns, will cause a significant decrease in the survival rate of neural stem cells, as the cell membrane of neural stem cells is sensitive to high-frequency vibration. For mesenchymal stem cells, the value is set to 5-8 Hz, with 5 Hz as the lower limit. Vibration below 5 Hz has less impact on mesenchymal stem cells, as the cell structure of mesenchymal stem cells is relatively strong. The upper limit is 8 Hz, and vibration above 8 Hz, such as continuous bumpy sections, will also cause a decrease in the survival rate of mesenchymal stem cells. The basis for setting the value is that the vibration frequency of stem cell transport vehicles on highways is usually 1-3 Hz, while rural roads or gravel roads are mostly 3-10 Hz, and sharp turns or bridges may reach 10-15 Hz.

[0050] The temperature tolerance threshold for neural stem cells is in the range of 1-2°C, and for mesenchymal stem cells, it is in the range of 2-3°C. The basis for setting the value is that the transport box usually maintains a constant temperature environment of 4-8°C, and the actual transportation is affected by external factors such as high summer temperatures and low winter temperatures. The temperature fluctuation on highway sections is about 0.5-1.5°C, and on rural roads or when stopping, it may reach 2-4°C.

[0051] The amplitude of temperature fluctuation of the transport box, the road vibration frequency, and the biological tolerance threshold are compared to obtain the risk judgment result of stem cell survival rate.

[0052] Specifically, when the temperature fluctuation range of the transport container and the road vibration frequency are less than 50% of the biological tolerance threshold, the current transport environment is classified as low-risk. Low risk indicates that environmental conditions have a relatively small impact on stem cells, and the possibility of cell survival being threatened is low. When the temperature fluctuation range of the transport container and the road vibration frequency are greater than or equal to the vibration tolerance threshold but less than the temperature tolerance threshold, the current transport environment is classified as medium-risk. For example, assuming the vibration tolerance threshold is 5 Hz and the temperature tolerance threshold is 2 degrees Celsius, if the road vibration frequency is 4 Hz and the temperature fluctuation range of the transport container is 1.5 degrees Celsius, then the current transport environment is medium-risk. When both the temperature fluctuation range of the transport container and the road vibration frequency are greater than or equal to the biological tolerance threshold, the current transport environment is classified as high-risk. For example, assuming the vibration tolerance threshold is 5 Hz and the temperature tolerance threshold is 2 degrees Celsius, if the road vibration frequency is 6 Hz and the temperature fluctuation range of the transport container is 3 degrees Celsius, then the current transport environment is high-risk.

[0053] Based on the risk assessment results of stem cell survival rate in the environment, corresponding weights are assigned to spatiotemporal features to generate path unit feature data.

[0054] Specifically, the corresponding weights should be set based on the actual risk level. For example, a weight of 1 is assigned to low-risk areas, 1.3 to medium-risk areas, and 1.5 to high-risk areas. The set weights are then applied to the spatiotemporal characteristics of the path units, generating path unit feature data, which is stored in JSON format and includes the path unit number, geographic coordinates, and weight value.

[0055] In stem cell transport scenarios where all three risk weights are greater than 1, the purpose of these weights is to amplify features that significantly impact stem cell survival, such as the impact of vibration frequency in high-risk environments, rather than allocating fixed proportions among spatiotemporal features. A weight sum of 1 is typically used in normalization scenarios. A weight greater than 1 reflects the unique characteristics of stem cell transport, where high-risk environments significantly increase the threat to cell survival, necessitating higher weights to highlight relevant spatiotemporal features. Forcing a weight sum of 1 would weaken the impact of high-risk features, resulting in insufficient sensitivity in risk assessment.

[0056] Based on the path unit characteristic data, the comprehensive risk value of each path unit is calculated, as expressed by: ; in, This represents the overall risk value of a path unit. This represents the rate of elevation change of the path unit. This represents the average number of vibrations on the road surface within the path element. This represents the average number of accidents within a path unit. a weight representing the rate of elevation change of the path unit, a weight representing the average vibration frequency of the path unit, a weight representing the average number of accidents in the path unit.

[0057] The integrated risk value of each path unit is mapped into the geographic space of the alternative path using the Kriging spatial interpolation method, and the risk level is marked to form a two-dimensional heat map.

[0058] Specifically, the path unit feature data is read, and the risk value of any point on the alternative path is estimated using Variogram (Variogram). Specifically, the geographic coordinates of any target point are selected and a Kriging weight value is set according to the distance from the surrounding path unit center point. When the geographic coordinates of the target point are close to the center point of the path unit, the Kriging weight value is larger, and vice versa. The Kriging weight value is used to reflect the contribution of the path unit to the risk value of the target point.

[0059] According to the Kriging weight value, the average risk value of the target point is calculated, and the expression is: ; Among them, the average risk value of the target point, the Kriging weight value.

[0060] A two-dimensional grid is set to cover the geographic range of the alternative path, and the grid point spacing is a fixed value, such as 10 meters per grid point. The calculated average risk value of the target point is stored as grid data and sequentially added to the two-dimensional grid, and the risk level is marked with color, such as blue for low risk, yellow for medium risk, and red for high risk, to form a two-dimensional heat map.

[0061] According to the precipitation probability distribution map in the integrated data set, set the risk adjustment rule and adjust the risk level in the two-dimensional heat map to generate a biological-environmental coupling heat map.

[0062] Specifically, the precipitation probability distribution map is read, and the risk adjustment rule needs to be set according to the actual transportation demand of stem cells, that is, to define the incremental influence of precipitation probability on the integrated risk value of the path unit. For example, when the precipitation probability exceeds 60%, the wet road increases the vibration and accident risk, and the integrated risk value is increased by 0.2. When the precipitation probability is in the interval of 30%~60%, the integrated risk value of the path unit is increased by 0.1. When the precipitation probability is less than 30%, the integrated risk value of the path unit is not adjusted.

[0063] The risk adjustment rule is applied to the integrated risk value of each path unit, and the expression is: ; Among them, a comprehensive risk value of each path unit after adjustment, an increment of the comprehensive risk value, determined according to a risk adjustment rule a specific numerical value.

[0064] The adjusted comprehensive risk value is mapped to the geographic space of the alternative path again by using the Kriging spatial interpolation method, the risk surface of the two-dimensional heat map is updated, and a bio-environment coupling risk heat map is generated, including the risk area of the path unit.

[0065] The evaluation module builds a stem cell transportation risk model, analyzes the environmental interaction between the bio-environment coupling heat map and the path unit feature data, obtains a path unit risk feature set and calculates the risk score of the alternative path, and outputs a path risk assessment report.

[0066] comprising the steps of, The multi-layer perception (MLP) is selected as the core architecture, which includes the input layer (receiving path unit features), the hidden layer (processing feature interaction), and the output layer (generating risk score).

[0067] The historical stem cell transportation data is collected, including the altitude change rate of the transportation route, the average road vibration frequency, the historical accident frequency, the precipitation probability, the temperature, and the transportation failure rate.

[0068] The historical stem cell transportation data is preprocessed, i.e. standardized to the range of 0-1 by using the min-max normalization method, for example, the annual average road vibration frequency ranges from 50 to 200 times.

[0069] The multi-layer perception is initialized, i.e. the input layer nodes are set to the feature dimensions of the altitude change rate of the transportation route, the average road vibration frequency, the historical accident frequency, the precipitation probability, and the temperature, the hidden layer is two layers with 50 nodes in each layer, the output layer is a single risk score, and the initial weights of the multi-layer perception are randomly set. The preprocessed historical stem cell transportation data is used as the training set to train the multi-layer perception, specifically using the gradient descent method, the goal is to make the predicted risk score close to the actual transportation failure rate. The training process is iterated 1000 times, the learning rate is set to 0.01, and the loss function is mean square error, which is used to compare the comprehensive risk score of the alternative path with the actual transportation failure rate. When the gradient descent method is iterated 1000 times, the change of the mean square error is less than a fixed value such as 0.001, indicating that the weights of the stem cell transportation risk model converge, and the training is completed.

[0070] The environmental interaction between the bio-environment coupling thermal map and the path unit feature data is analyzed using the stem cell transportation risk model, for example, the hidden layer of the stem cell transportation risk model captures the nonlinear relationship between the precipitation probability and the vibration frequency. Taking the precipitation probability and the road surface vibration frequency as an example, the environmental interaction intensity value is calculated, and the expression is as follows: ; wherein, represents the environmental interaction intensity value, the numerical range is [-1, 1], represents the average vibration frequency of the road surface of the path unit, represents the average vibration frequency of the road surface in actual transportation, represents the precipitation probability, represents the average precipitation probability.

[0071] The interaction intensity threshold is set according to the industry standards of the stem cell transportation industry, such as the logistics standards of the biological cold chain. The specific value can be set to 0.6, because the cell survival rate decreases by more than 10% which meets the industry high risk standard, 0.6 sensitivity is moderate, to avoid too high to miss the risk or too low to mark too many irrelevant combinations.

[0072] When the absolute value of the environmental interaction intensity value is greater than or equal to the interaction intensity threshold, the parameters in the environmental interaction intensity value are marked as high-risk combination features, such as precipitation probability ≥ 60% and vibration frequency ≥ 100 times / year. The high-risk combination features are stored as the interaction data set, saved in JSON format, containing path unit number, geographic coordinates and environmental interaction intensity value.

[0073] According to the interaction data set, the weight of the average vibration frequency of the road surface of the path unit is adjusted.

[0074] The reason for choosing the average vibration frequency of the road surface of the path unit as the weight adjustment object is that the sensitivity of stem cells to vibration is much higher than that of temperature or altitude change. When the vibration frequency of the road surface exceeds the biological tolerance threshold, it will cause damage to the cell membrane of stem cells or apoptosis, and the survival rate will decrease significantly. In contrast, the altitude change rate mainly affects the stability of the transportation equipment, the historical accident frequency only reflects the long-term risk trend, and the precipitation probability and temperature are indirectly included in the risk calculation through the thermal map. The changes of these four features are relatively slow, so the dynamic demand of weight adjustment is low, and therefore the average vibration frequency of the road surface is the core condition affecting the survival of stem cells.

[0075] Specifically, according to the previously obtained stem cell survival rate risk judgment result, the vibration sensitivity coefficient is set, that is, when the road surface vibration frequency is judged as low, medium and high risk, the values of the vibration sensitivity coefficient are different, for example, 0.2 for low risk, 0.3 for medium risk, and 0.5 for high risk.

[0076] The expression for adjusting the weight of the average vibration number of the road surface in the path element is: ; in, The weight representing the average number of road surface vibrations in the adjusted path element. The weight representing the average number of road surface vibrations in the path element, i.e., the original weight. This represents the vibration sensitivity coefficient. For example, assuming the original weight is 1.3 and the vibration sensitivity coefficient is 0.5, the weight of the average vibration frequency of the road surface in the adjusted path unit is 1.95.

[0077] The path unit feature data, the weights of the average road surface vibration frequency of the adjusted path unit, and the interaction dataset are integrated and stored in JSON format to form a path unit risk feature set.

[0078] Using the risk feature set of the path unit, the cumulative vibration value, standard deviation of temperature fluctuation, and overlap of risk periods for each candidate path are calculated and input into the stem cell transportation risk model to generate a comprehensive risk score for the candidate paths.

[0079] Specifically, the cumulative vibration value for each candidate path is calculated using the following expression: ; in, This represents the cumulative vibration value of the alternative paths. This represents the average number of road surface vibrations per path unit. For example, if there are 100 path units, the weight of the average number of road surface vibrations per path unit is 1.8, and the average number of road surface vibrations is 100, then the cumulative vibration value of this candidate path is 18000.

[0080] The standard deviation of temperature fluctuation for each candidate path is calculated using the following expression: ; in, This represents the standard deviation of temperature fluctuations along the alternative routes. Indicates the first Temperature values ​​of each path unit Indicates the number of path units. This represents the average temperature across the path unit.

[0081] Calculate the overlap of risk periods for each candidate route. This involves obtaining the comprehensive risk value of each route unit from the bio-environment coupling risk heatmap, classifying route units with a comprehensive risk value greater than or equal to 0.7 as high-risk route units, and counting their number. The expression is: ; in, a degree of coincidence of the risk period representing the alternative path, a number of high-risk path units. Example: assuming that there are 100 path units on a certain alternative path, and the number of high-risk path units is 30, then the degree of coincidence of the risk period is 30%.

[0082] The vibration cumulative value, temperature fluctuation standard deviation, and degree of coincidence of the risk period of each alternative path are normalized, that is, according to the minimum and maximum values of the vibration cumulative value, temperature fluctuation standard deviation, and degree of coincidence of the risk period of each alternative path, the three indicators are mapped to the range of 0 to 1.

[0083] The vibration cumulative value, temperature fluctuation standard deviation, and degree of coincidence of the risk period of each alternative path after normalization are input into the stem cell transportation risk model, and a comprehensive risk score of the alternative path is generated by using a nonlinear mapping function, and the expression is: ; wherein, the comprehensive risk score of the alternative path, the nonlinear mapping function, the vibration cumulative value of the normalized alternative path, the temperature fluctuation standard deviation of the normalized alternative path, the degree of coincidence of the risk period of the normalized alternative path. Example: assuming that the vibration cumulative value of the normalized alternative path is 0.6, the temperature fluctuation standard deviation of the normalized alternative path is 0.4, and the degree of coincidence of the risk period of the normalized alternative path is 0.3, then the comprehensive risk score of the alternative path calculated by the nonlinear mapping function is 0.48.

[0084] The comprehensive risk scores of each alternative path are sorted, and the path with the lowest comprehensive risk score is selected as the candidate path, for example, there are five alternative paths, and the comprehensive risk scores are 0.62, 0.55, 0.57, 0.48, and 0.6, then the alternative path with the comprehensive risk score of 0.48 is selected as the candidate path, including the geographic coordinates, number, etc. of the candidate path.

[0085] The geographical coordinates and environmental factors of the risk areas of path units in the bio-environment coupling risk heatmap are recorded. For example, the geographical coordinates of the entrance to a certain overpass are (121.06°, latitude 31.07°). Environmental factors include the probability of precipitation, road vibration frequency, temperature, frequency of historical accidents, and road type of the transportation section. Mitigation measures are generated. For example, based on the road vibration frequency in the environmental factors, it is recommended that the transport vehicle use shock-absorbing pads or buffer devices to wrap the transport box to reduce vibration damage to stem cells from high-vibration road sections. Alternatively, the vehicle can be detoured to a low-vibration road section. Regarding the temperature of the transportation section, it is recommended to equip the transport box with constant temperature and set the temperature control within ±0.5°C to offset the impact of temperature fluctuations in the road section. At the same time, transportation should be avoided during high-temperature periods. Regarding the road type, such as gravel road, it is recommended that the transport vehicle reduce its speed to below 20 km / h to reduce the additional vibration caused by the unevenness of the gravel road or choose an alternative route such as asphalt road to reduce transportation risk.

[0086] The comprehensive risk scores of alternative routes, mitigation measures recommendations, and candidate routes are integrated into a PDF format to output a route risk assessment report.

[0087] The output module selects the route with the lowest risk as the transportation route based on the route risk assessment report, and outputs the final navigation route and a 3D visualization navigation interface.

[0088] Includes the following steps, Calculate the total length and estimated transit time of the candidate routes, mark the turning points of the candidate routes, and generate turning point navigation instructions.

[0089] Specifically, based on the geographical coordinates of the path units, the total length of the candidate paths is calculated using the Euclidean distance formula, expressed as: ; in, Indicates the total length of the candidate paths. Indicates the number of path units. Indicates the first Longitude of each path unit Indicates the first The latitude of each path unit Indicates the first Longitude of each path unit Indicates the first The latitude of each path unit. For example, assuming there are 100 path units and the average distance between path units is 0.1 kilometers, the total length of the candidate paths calculated by this formula is approximately 9.9 kilometers, while the actual total length is approximately 600 kilometers.

[0090] The expected transportation time of the candidate path is calculated. According to the avoidance measure suggestions in the path risk assessment report, that is, the recommended speed of the road section, assuming 20 km / h, the driving time of each section of the candidate path is calculated, and the expected transportation time of the candidate path is accumulated.

[0091] The turning points of the candidate path are marked, and the turning point navigation instructions are generated. That is, the geographic coordinates of the path unit are analyzed, and the path unit with an angle change ≥ 30° (such as an intersection) is identified, and the turning point navigation instruction is generated, such as the transportation vehicle will pass through the right turn at 121.055°, 31.055° into XX Road.

[0092] The path unit with an angle change ≥ 30° is selected as the turning point, because a smaller angle change usually only indicates a slight bending of the path, that is, the natural curvature of the road, and no special navigation instruction is needed, while a change ≥ 30° reflects a clear turning requirement, such as a 90° intersection or a 45° curve, which has practical guiding significance for the driver's navigation.

[0093] The total length of the candidate path, the expected transportation time, and the turning point navigation instruction are integrated and stored in JSON format to form the navigation path data.

[0094] The geographic information of the refrigeration supply station is collected, including the supply station number, geographic coordinates, and expected arrival time, and the weather warning information of the navigation path is obtained through the API of the meteorological bureau, such as the risk of icing on a certain bridge section, suggesting that the vehicle reduce speed to 15 km / h.

[0095] The geographic information of the navigation path, the refrigeration supply station, and the weather warning information of the navigation path are marked using the Cesium three-dimensional rendering engine to generate a three-dimensional visual navigation page.

[0096] Specifically, the navigation path is drawn, the geographic coordinates of the path unit are connected using green lines, the total length and expected transportation time of the candidate path are displayed, the path unit with a comprehensive risk value ≥ 0.7 is marked with red, and avoidance measure suggestions are added, such as reducing speed to 20 km / h on a slippery road section. Mark the turning points, use orange arrows for path units with an angle change ≥ 30°, and add turning point navigation instructions; mark the refrigeration supply station, mark the supply station location with a blue icon, that is, the geographic coordinates, number, and expected arrival time of the supply station, display weather warnings, and use a yellow warning box to mark the warning section such as a bridge section, and add a prompt information such as the risk of icy road surface, suggesting to reduce speed to 10 km / h. The generated three-dimensional visual navigation interface is stored in HTML format and presented through the vehicle display screen or the remote management platform (such as Web).

[0097] To sum up, the application converts the biological characteristics of stem cells into a quantitative index of risk assessment by setting different vibration sensitivity coefficients and real-time feedback of the transport box biosensor, dynamically adjusting the weight of parameters in path characteristic data, overcoming the drawbacks of ignoring the coupling effect of biological characteristics and environment in conventional stem cell transport path planning, making risk assessment from general cold chain transportation to stem cell specific protection, and greatly reducing the risk of stem cell damage. And through the analysis of the nonlinear relationship between precipitation probability and vibration frequency by the multilayer perception machine, the weight of the vibration frequency is automatically improved, and the generated high-risk combination features break through the limitations of artificial experience setting weight, and through machine learning, the key risk combination is found automatically, avoiding the sudden drop of cell survival rate under the combined environmental factors.

[0098] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A stem cell transport remote management system based on environmental prediction, characterized by: The application relates to a biological-transportation environment coupling risk assessment system and method. The data acquisition module is used for selecting an alternative path by a dispatch center, collecting transportation environment data and preprocessing the transportation environment data to obtain a comprehensive data set; The visualization module is used for discretizing the alternative path into path units according to the comprehensive data set, labeling space-time features of each path unit by using a space-time data correlation method, setting weights of the space-time features by using a stem cell feature enhancement mechanism, generating a biological-transportation environment coupling thermodynamic map and path unit feature data, and outputting a path risk assessment report by calculating a risk score of the alternative path. The output module is used for selecting a path with the lowest risk as a transportation path according to the path risk assessment report, and outputting a final navigation path and a three-dimensional visual navigation interface. The transportation environment data includes road flatness data, a rainfall probability distribution map, a temperature change curve, wind speed data, road surface vibration data and historical road accident data.

2. The environmental prediction-based stem cell transport remote management system of claim 1, wherein: The comprehensive data set is obtained by the following steps, 3. The environmental prediction-based stem cell transport remote management system of claim 2, wherein: The road flatness data is denoised, missing values in the rainfall probability distribution map and the wind speed data are filled by using a linear interpolation method, temperature mutation points in the temperature change curve are identified and removed, and the historical road accident data and the road surface vibration data are recorded and screened to output cleaned transportation environment data; The cleaned transportation environment data is standardized, normalized and space-time aligned to obtain the comprehensive data set. The alternative path is discretized into path units by the following steps, 4. The environmental prediction based stem cell transport remote management system as claimed in claim 3, wherein: The geographic coordinates of the alternative path points in the comprehensive data set are selected as coordinate points, the spherical distance between adjacent coordinate points is calculated by using a secant formula to obtain the total length of the alternative path; The path point number of the alternative path is calculated by setting a fixed interval of the alternative path and the total length of the alternative path; The distance of the alternative path is accumulated from the starting point of the alternative path, and when the accumulated distance of the alternative path is greater than or equal to the fixed interval of the alternative path, the division point of the alternative path is determined; Each alternative path is divided into path units according to the path point number of the alternative path and the division point of the alternative path. The space-time features of each path unit are calculated by using the road flatness data, the road surface vibration data and the historical road accident data in the comprehensive data set, and the space-time features are marked as the space-time features. The biological-transportation environment coupling thermodynamic map and the path unit feature data are generated by the following steps, 5. The environmental prediction-based stem cell transport remote management system of claim 4, wherein: A biological tolerance threshold is set, the temperature fluctuation amplitude and the road surface vibration frequency of the transportation box are collected by using a biological sensor, and the temperature fluctuation amplitude and the road surface vibration frequency are compared with the biological tolerance threshold to obtain a stem cell survival rate risk judgment result; 6. The environmental prediction-based stem cell transport remote management system of claim 5, wherein: According to the stem cell survival rate risk judgment result, corresponding weights are set for the space-time features, path unit feature data are generated, and the comprehensive risk value of each path unit is calculated; The comprehensive risk value of each path unit is mapped to the geographic space of the alternative path by using a Kriging spatial interpolation method, and a risk level is marked to form a two-dimensional thermodynamic map. ​ ​ The risk adjustment rule is set, and the risk level in the two-dimensional heat map is adjusted to generate a bio-environment coupling heat map.

7. The environmental prediction-based stem cell transport remote management system of claim 6, wherein: The construction of the stem cell transportation risk model includes selecting a multi-layer perception as a core architecture, collecting historical stem cell transportation data and preprocessing, training the multi-layer perception using the preprocessed historical stem cell transportation data, and obtaining the stem cell transportation risk model.

8. The environmental prediction-based stem cell transport remote management system of claim 7, wherein: The path unit risk feature set is obtained, specifically including the following steps, An interaction intensity threshold is set, the environmental interaction intensity value is calculated using the stem cell transportation risk model and compared with the interaction intensity threshold, the high-risk combination feature is judged, and the interaction data set is output; According to the stem cell survival rate risk judgment result, a vibration sensitivity coefficient is set and the weight of the average vibration frequency of the road surface of the path unit is adjusted; The path unit feature data, the weight of the average vibration frequency of the road surface of the adjusted path unit and the interaction data set are integrated into JSON format to form the path unit risk feature set.

9. The environmental prediction-based stem cell transport remote management system of claim 8, wherein: The output path risk assessment report specifically includes the following steps, Using the path unit risk feature set, the vibration cumulative value, temperature fluctuation standard deviation and risk period coincidence degree of each alternative path are calculated, and after normalization processing, they are input into the stem cell transportation risk model to generate the comprehensive risk score of the alternative path; The comprehensive risk score of each alternative path is sorted, and the path with the lowest comprehensive risk score is selected as the candidate path; The geographic coordinates and environmental factors of the path unit risk area in the bio-environment coupling risk heat map are recorded to generate avoidance measure suggestions; The comprehensive risk score of the alternative path, the avoidance measure suggestions and the candidate path are integrated into PDF format to output the path risk assessment report.

10. The environmental prediction-based stem cell transport remote management system of claim 9, wherein: The output final navigation path and three-dimensional visual navigation interface specifically includes the following steps, The total length and the estimated transportation time of the candidate path are calculated, and the turning point of the candidate path is marked to generate the turning point navigation instruction; The total length, the estimated transportation time and the turning point navigation instruction of the candidate path are integrated to form the navigation path data; The geographic information of the refrigeration supply station and the weather warning information of the navigation path are collected, and the navigation path, the geographic information of the refrigeration supply station and the weather warning information of the navigation path are labeled using the Cesium three-dimensional rendering engine to generate a three-dimensional visual navigation page.

Images