Refrigerated container control method and device based on artificial intelligence

Abstract

The invention provides a refrigerated container control method and device based on artificial intelligence, and relates to the technical field of refrigerated containers. According to the method, the refrigeration temperature needed by refrigeration of stacked goods is determined according to the type of the refrigeration stacked goods, the type, the weight and the refrigeration temperature of the refrigeration stacked goods are input into an operation parameter determination model, and multiple sets of operation parameters of the intelligent refrigeration container enabling all positions on the refrigeration stacked goods to keep the refrigeration temperature are determined. Inputting the refrigeration power and the air outlet speed in each group of operation parameters into a power consumption determination model, and determining the power consumption of the refrigeration equipment corresponding to each group of operation parameters; and selecting the minimum power consumption of the refrigeration equipment and the corresponding group of operation parameters. According to the refrigeration power, the air outlet speed and the air outlet direction of the refrigeration equipment in the selected set of operation parameters, the refrigeration equipment is controlled to operate, and all the electric air deflectors are controlled to reach the included angle between the extending direction, facing the interior of the box body, in the selected operation parameters and the horizontal direction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of refrigerated containers, and in particular to a refrigerated container control method and device based on artificial intelligence. BACKGROUND

[0002] A container refers to a large cargo container specially used for circulation, which has certain strength, rigidity and specifications. Using a container to transport goods can directly load goods in the warehouse of the sender and unload goods in the warehouse of the receiver. During the transportation, the goods do not need to be taken out of the container for reloading. For some goods that need to be refrigerated during transportation (such as medical biological agents, vaccines, and food that needs to be refrigerated), the goods that need to be refrigerated during transportation can be loaded into a cargo box, and then the cargo box is stacked into a refrigerated container. The refrigerated container can refrigerate the goods in the cargo box according to the fixed refrigeration power corresponding to the type of the goods, so as to prevent the goods from deteriorating.

[0003] At present, in order to save the space of the transportation vehicle, the refrigerated container that is not fully loaded with goods can be folded to a certain extent. However, after the refrigerated container is folded, the ventilation channel in the refrigerated container will change, which will cause the cold and hot degrees of the goods at different positions to be different, and thus still cause some goods to deteriorate. SUMMARY

[0004] The present application provides a refrigerated container control method and device based on artificial intelligence, which is used to solve the problem that in the prior art, after the refrigerated container is folded, the ventilation channel in the refrigerated container will change, which will cause the cold and hot degrees of the goods at different positions to be different, and thus still cause some goods to deteriorate.

[0005] In a first aspect, the present application provides a refrigerated container control method based on artificial intelligence, which is applied to a main controller of an intelligent refrigerated container. The intelligent refrigerated container includes a box body, the box body includes oppositely arranged top and bottom plates, two oppositely arranged first side panels, and two oppositely arranged second side panels. Each first side panel is foldable towards the inside of the box body, and the folding line of each first side panel when folded is parallel to the top plate or the bottom plate of the box body. The folding line of the inner side of the two first side panels is respectively provided with a first distance sensor and a second distance sensor, the inner side of the top plate is provided with a third distance sensor, and the outer side of the top plate is provided with a human-computer interaction panel. Each first side panel is respectively provided with an electrically operated air deflector on both sides of the folding line thereof. The inner side of one of the second side panels is provided with a refrigeration device. The method comprises the following steps: receive a type and a weight of the refrigerated stacked goods in the box input by a user on a human-computer interaction panel, and receive a first distance value of the refrigerated stacked goods collected by a first distance sensor, a second distance value of the refrigerated stacked goods collected by a second distance sensor, and a third distance value of the refrigerated stacked goods collected by a third distance sensor; subtract the first distance value from a preset distance threshold value to obtain a first distance difference, subtract the second distance value from the preset distance threshold value to obtain a second distance difference, and subtract the third distance value from the preset distance threshold value to obtain a third distance difference; determine a foldable angle corresponding to each of the first distance difference, the second distance difference, and the third distance difference, and drive each first side panel along a folding line to fold a minimum foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference; determine a refrigeration temperature required by the refrigerated stacked goods according to the type of the refrigerated stacked goods; input the type, the weight, and the refrigeration temperature of the refrigerated stacked goods into a pre-trained operation parameter determination model to determine a plurality of sets of operation parameters of the intelligent refrigerated container for maintaining the refrigeration temperature at each position on the refrigerated stacked goods, each set of operation parameters including a refrigeration power, an air outlet speed, and an air outlet direction of the refrigeration equipment, and an angle between an extension direction of each electrically-driven air deflector towards the inside of the box and a horizontal direction, the operation parameter determination model being trained by inputting a plurality of first training samples into a first to-be-trained network, each first training sample including a historical type, a historical weight, and a historical refrigeration temperature of a historical refrigerated stacked goods, and historical operation parameters of an intelligent refrigerated container for maintaining the historical refrigeration temperature at each position on the historical refrigerated stacked goods; input the refrigeration power and the air outlet speed in each set of operation parameters into a pre-trained power consumption determination model to determine a power consumption of the refrigeration equipment corresponding to each set of operation parameters, wherein the power consumption determination model is trained by inputting a plurality of second training samples into a second to-be-trained network, each second training sample including a historical refrigeration power and a historical air outlet speed in historical operation parameters of the refrigeration equipment, and a corresponding historical power consumption; select a set of operation parameters corresponding to a minimum power consumption of the refrigeration equipment; control the refrigeration equipment to operate according to the refrigeration power, the air outlet speed, and the air outlet direction of the refrigeration equipment in the selected set of operation parameters, and control each electrically-driven air deflector to rotate so as to make each electrically-driven air deflector reach the angle between the extension direction towards the inside of the box and the horizontal direction in the selected set of operation parameters.

[0006] In some embodiments, each first side panel is provided with a non-contact temperature sensor on both sides of the folding line thereof, and the geometric center of the box body is provided with a contact temperature sensor. After the refrigeration device is controlled to operate according to the refrigeration power, the air outlet speed and the air outlet direction of the refrigeration device, and each electrically-driven air deflector is controlled to rotate so as to make each electrically-driven air deflector reach a determined included angle between the extension direction thereof towards the inside of the box body and the horizontal direction, the method provided by the application further comprises: receiving the temperature of the refrigerated stacked goods collected by each non-contact temperature sensor and receiving the temperature of the refrigerated stacked goods collected by the contact temperature sensor; determining an actual average temperature of the received multiple temperatures of the refrigerated stacked goods; determining a difference temperature between the actual average temperature and the refrigeration temperature required by the refrigerated stacked goods; in the case where the difference temperature is greater than a set temperature threshold, updating the first network parameter of the running parameter determination model according to the first reinforcement learning model until the difference temperature is less than or equal to the set temperature threshold, wherein the difference temperature is the reward of the first reinforcement learning model, and the lower the difference temperature, the higher the reward, the first network parameter of the running parameter determination model is the action of the first reinforcement learning model, and the first network parameter is the state of the first reinforcement learning model.

[0007] In some embodiments, in the case where the difference temperature is less than or equal to the set temperature threshold, the method provided by the application further comprises: determining the variance of the received multiple temperatures of the refrigerated stacked goods; in the case where the variance of the multiple temperatures of the refrigerated stacked goods is greater than a preset variance threshold, updating the second network parameter of the running parameter determination model according to the second reinforcement learning model until the variance of the multiple temperatures of the refrigerated stacked goods is less than or equal to the preset variance threshold, wherein the variance of the multiple temperatures of the refrigerated stacked goods is the reward of the second reinforcement learning model, and the smaller the variance of the multiple temperatures of the refrigerated stacked goods, the higher the reward, the second network parameter of the running parameter determination model is the action of the second reinforcement learning model, and the second network parameter is the state of the second reinforcement learning model.

[0008] In some embodiments, determining the foldable angles corresponding to the first distance difference, the second distance difference and the third distance difference respectively comprises: according to a preset first mapping relationship, searching for the foldable angles corresponding to the first distance difference and the second distance difference from a preset database; according to a preset second mapping relationship, searching for the foldable angle corresponding to the third distance difference from the preset database.

[0009] In some embodiments, the first network to be trained is a neural network.

[0010] In a second aspect, the application also provides an artificial intelligence-based refrigerated container control device configured in a main controller of an intelligent refrigerated container, the intelligent refrigerated container comprising a box body, the box body comprising a top plate and a bottom plate arranged oppositely, two first side panels arranged oppositely, and two second side panels arranged oppositely, each first side panel being foldable towards the inside of the box body, and the folding line of each first side panel when folded being parallel to the top plate or the bottom plate of the box body; a first distance sensor and a second distance sensor being arranged on the folding line of the inner side of each first side panel respectively, a third distance sensor being arranged on the inner side of the top plate, and a human-computer interaction panel being arranged on the outer side of the top plate, one of the second side panels having a refrigeration device arranged on the inner side thereof, and the device comprising: a data receiving unit configured to receive the type and weight of the refrigerated stacked goods to be transported in the box body input by a user on the human-computer interaction panel, and receive the first distance value of the refrigerated stacked goods collected by the first distance sensor, the first distance value of the refrigerated stacked goods collected by the second distance sensor, and the third distance value of the refrigerated stacked goods collected by the third distance sensor; a distance difference determining unit configured to subtract a preset distance threshold value from the first distance value to obtain a first distance difference, subtract the preset distance threshold value from the second distance value to obtain a second distance difference, and subtract the preset distance threshold value from the third distance value to obtain a third distance difference; a foldable angle determining unit configured to determine the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference respectively, and control the driving motor to drive each first side panel to fold along the folding line by the smallest foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference; a refrigeration temperature determining unit configured to determine the refrigeration temperature required by the refrigerated stacked goods according to the type of the refrigerated stacked goods; a running parameter determining unit configured to input the type, weight, and refrigeration temperature of the refrigerated stacked goods into a pre-trained running parameter determining model to determine a plurality of sets of running parameters of the intelligent refrigerated container for maintaining the refrigeration temperature on each position of the refrigerated stacked goods, each set of running parameters comprising the refrigeration power, air outlet speed, and air outlet direction of the refrigeration device, and the included angle between the extension direction of each electrically-driven air deflector towards the inside of the box body and the horizontal direction, the running parameter determining model being obtained by inputting a plurality of first training samples into a first training network, each first training sample comprising the historical type, historical weight, and historical refrigeration temperature of a historical refrigerated stacked goods, and the historical running parameters of the intelligent refrigerated container for maintaining the historical refrigeration temperature on each position of the historical refrigerated stacked goods; The power consumption determination unit is configured to input the refrigeration power and the air outlet speed in each set of operation parameters into a pre-trained power consumption determination model to determine the power consumption of the refrigeration equipment corresponding to each set of operation parameters, wherein the power consumption determination model is obtained by inputting a plurality of second training samples into a second to-be-trained network for training, each second training sample includes a historical refrigeration power and a historical air outlet speed in historical operation parameters of the refrigeration equipment, and a corresponding historical power consumption; The operation parameter selection unit is configured to select a set of operation parameters corresponding to the minimum power consumption of the refrigeration equipment. The control unit is configured to control the refrigeration equipment to operate according to the refrigeration power, the air outlet speed and the air outlet direction of the refrigeration equipment in the selected set of operation parameters, and control each electrically-driven air deflector to rotate so that each electrically-driven air deflector reaches an extension direction towards the inside of the cabinet and a horizontal direction at an included angle in the selected set of operation parameters.

[0011] In some embodiments, each first side panel is provided with a non-contact temperature sensor on both sides of the folding line thereof, and the geometric center of the cabinet is provided with a contact temperature sensor, The data receiving unit is further configured to receive the temperature of the refrigerated stacked goods collected by each non-contact temperature sensor and receive the temperature of the refrigerated stacked goods collected by the contact temperature sensor. The device provided in the application further includes: The average temperature determination unit is configured to determine an actual average temperature of the plurality of received temperatures of the refrigerated stacked goods. The difference temperature determination unit is configured to determine a difference temperature between the actual average temperature and the refrigeration temperature required by the refrigerated stacked goods. The parameter updating unit is configured to update the first network parameter of the operation parameter determination model according to the first reinforcement learning model when the difference temperature is greater than a set temperature threshold, until the difference temperature is less than or equal to the set temperature threshold, wherein the difference temperature is the reward of the first reinforcement learning model, and the lower the difference temperature, the higher the reward, the first network parameter of the operation parameter determination model is the action of the first reinforcement learning model, and the first network parameter is the state of the first reinforcement learning model.

[0012] In some embodiments, the device provided in the application further includes: The temperature variance determination unit is configured to determine the variance of the plurality of received temperatures of the refrigerated stacked goods. The parameter updating unit is further configured to update the second network parameter of the running parameter determination model according to the second reinforcement learning model in a case where the variance of the plurality of temperatures of the refrigerated stacked goods is greater than the preset variance threshold, until the variance of the plurality of temperatures of the refrigerated stacked goods is less than or equal to the preset variance threshold, wherein the variance of the plurality of temperatures of the refrigerated stacked goods is a reward of the second reinforcement learning model, and the smaller the variance of the plurality of temperatures of the refrigerated stacked goods is, the higher the reward is, the second network parameter of the running parameter determination model is an action of the second reinforcement learning model, and the second network parameter is a state of the second reinforcement learning model.

[0013] In some embodiments, the foldable angle determination unit is specifically configured to: find, according to a preset first mapping relationship, a foldable angle corresponding to the first distance difference and the second distance difference from a preset database; and find, according to a preset second mapping relationship, a foldable angle corresponding to the third distance difference from the preset database.

[0014] In some embodiments, the first network to be trained is a neural network.

[0015] The application provides a refrigerated container control method and device based on artificial intelligence. The refrigerated stacked goods can be stacked from the center to the periphery of the refrigerated container, and the refrigerated stacked goods need to maintain a certain distance from the inner wall of the refrigerated container to ensure the air circulation channel inside. Then, the type and weight of the refrigerated stacked goods to be transported in the container can be received by the user through the man-machine interaction panel, and the first distance value of the refrigerated stacked goods collected by the first distance sensor, the first distance value of the refrigerated stacked goods collected by the second distance sensor, and the third distance value of the refrigerated stacked goods collected by the third distance sensor can be received.

[0016] The foldable angle corresponding to the first distance difference, the second distance difference, and the third distance difference is determined, and each first side panel is driven by the control driving motor to fold the minimum foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference along the folding line. In this way, each first side panel and the top plate can maintain a certain distance from the inner wall of the refrigerated container after the refrigerated container is folded.

[0017] The refrigeration temperature required by the refrigerated stacked goods is determined according to the type of the refrigerated stacked goods. The refrigeration temperature required can save power consumption while the refrigerated stacked goods are not deteriorated.

[0018] The type, weight and refrigeration temperature of the refrigerated stacked goods are input into the pre-trained operation parameter determination model to determine a plurality of sets of operation parameters of the intelligent refrigerated container for maintaining the refrigeration temperature at each position on the refrigerated stacked goods, each set of operation parameters including the refrigeration power, air outlet speed and air outlet direction of the refrigeration equipment, and the angle between the extension direction of each electrically-driven air deflector towards the inside of the container and the horizontal direction. In this way, each position on the refrigerated stacked goods can be at the required refrigeration temperature and uniform in cold and heat.

[0019] The refrigeration power and air outlet speed in each set of operation parameters are input into the pre-trained power consumption determination model to determine the power consumption of the refrigeration equipment corresponding to each set of operation parameters; and a set of operation parameters corresponding to the minimum power consumption of the refrigeration equipment is selected. The refrigeration power, air outlet speed and air outlet direction of the refrigeration equipment in the selected set of operation parameters are used to control the operation of the refrigeration equipment; and each electrically-driven air deflector is controlled to rotate so as to reach the angle between the extension direction of each electrically-driven air deflector towards the inside of the container and the horizontal direction in the selected set of operation parameters. In this way, each position on the refrigerated stacked goods can be at the required refrigeration temperature and uniform in cold and heat, and the power consumption is minimized, while the occupied space of the refrigerated container is saved, and the practicability is high. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative labor.

[0021] Figure 1 A sectional view of the intelligent refrigerated container in an unfolded state according to an embodiment of the present application; Figure 2 One of the flowcharts of the refrigerated container control method based on artificial intelligence according to an embodiment of the present application; Figure 3 A sectional view of the intelligent refrigerated container in a folded state according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments made by ordinary technicians in this field based on the inspiration of these embodiments fall within the scope of protection of this application.

[0023] The terms "first," "second," "third," "fourth," and so forth (if any) in the specification and claims of this application and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequential sequence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the application described herein can be implemented in orders other than those illustrated or described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not necessarily limited to those steps or elements explicitly listed, but may include other steps or elements not explicitly listed or inherent to such process, method, product, or apparatus.

[0024] The embodiment of the present application provides a refrigerated container control method based on artificial intelligence, which is applied to the main controller of the intelligent refrigerated container. Figure 1 As shown, the intelligent refrigerated container includes a box body, and the geometric shape of the box body is a rectangular parallelepiped. The box body includes a top plate 101 and a bottom plate 102 arranged opposite to each other, two first side panels 103 arranged opposite to each other, and two second side panels arranged opposite to each other (not shown in the drawings). Each first side panel 103 is foldable toward the inside of the box body, and the folding line of each first side panel 103 when folded is parallel to the top plate 101 or the bottom plate 102 of the box body; a first distance sensor 105 and a second distance sensor 106 are respectively provided on the folding line on the inner side of the two first side panels 103, a third distance sensor 107 is provided on the inner side of the top plate 101, and a human-machine interaction panel 108 is provided on the outer side of the top plate 101. Each first side panel 103 is provided with an electric air deflector 109 on both sides of its folding line, and a refrigeration device is provided on the inner side of one of the second side panels (not shown in the drawings). As shown Figure 2 As shown, the method provided in the embodiment of the present application includes: S201: Receive the type and weight of the refrigerated stacked goods 111 to be transported in the box input by the user on the human-computer interaction panel 108, and receive the first distance value from the refrigerated stacked goods 111 collected by the first distance sensor 105, the first distance value from the refrigerated stacked goods 111 collected by the second distance sensor 106, and the third distance value from the refrigerated stacked goods 111 collected by the third distance sensor 107.

[0025] It should be noted that the refrigerated stacked goods 111 can be stacked from the center of the refrigerated container to the periphery, and the refrigerated stacked goods 111 needs to maintain a certain distance from the inner wall of the refrigerated container to ensure the internal air circulation channel.

[0026] For example, the types of the refrigerated stacked goods 111 can be, but are not limited to, ice cream packaged by small boxes, biological agents (such as insulin), and vaccines (such as the new crown vaccine), etc. It can be understood that there are gaps between each refrigerated stacked good 111 packaged by small boxes, which can be ventilated.

[0027] S202: subtracting the first distance value by a preset distance threshold to obtain a first distance difference, subtracting the second distance value by the preset distance threshold to obtain a second distance difference, and subtracting the third distance value by the preset distance threshold to obtain a third distance difference.

[0028] S203: determining the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference, respectively, and driving each first side panel 103 according to the control driving motor to fold the minimum foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference along the folding line.

[0029] Exemplarily, the foldable angles corresponding to the first distance difference and the second distance difference can be found from the preset database according to a preset first mapping relationship, and the foldable angle corresponding to the third distance difference can be found from the preset database according to a preset second mapping relationship. It should be noted that the first to-be-trained network is a neural network.

[0030] wherein, Figure 2 is a cross-sectional view of the intelligent refrigerated container after folding, and the foldable angle can be the included angle between the two parts of the first side panel 103 after folding.

[0031] S204: determining the refrigeration temperature required by the refrigerated stacked goods 111 according to the type of the refrigerated stacked goods 111.

[0032] For example, the refrigeration temperature required by the refrigerated stacked goods 111 can be found from the preset database.

[0033] S205: inputting the type, weight, and refrigeration temperature of the refrigerated stacked goods 111 into the pre-trained operation parameter determination model to determine a plurality of sets of operation parameters of the intelligent refrigerated container for maintaining the refrigeration temperature at each position on the refrigerated stacked goods 111, each set of operation parameters including the refrigeration power of the refrigeration equipment, the air outlet speed and direction, and the angle between the extension direction of each electrically operated air deflector 109 to the inside of the box and the horizontal direction.

[0034] The operation parameter determination model is obtained by inputting a plurality of first training samples into a first to-be-trained network for training, each first training sample including a historical type, a historical weight and a historical refrigeration temperature of a historical refrigerated stacked cargo 111, and historical operation parameters of an intelligent refrigerated container for maintaining the historical refrigeration temperature at each position on the historical refrigerated stacked cargo 111.

[0035] S206: input the refrigeration power and the air outlet speed in each group of operation parameters into the pre-trained power consumption determination model to determine the power consumption of the refrigeration equipment corresponding to each group of operation parameters.

[0036] The power consumption determination model is obtained by inputting a plurality of second training samples into a second to-be-trained network for training, each second training sample including a historical refrigeration power and a historical air outlet speed in historical operation parameters of the refrigeration equipment, and a corresponding historical power consumption.

[0037] S207: select a group of operation parameters corresponding to the smallest power consumption of the refrigeration equipment.

[0038] S208: control the refrigeration equipment to operate according to the refrigeration power, the air outlet speed and the air outlet direction of the refrigeration equipment in the selected group of operation parameters; and control each electrically-driven air deflector 109 to rotate so that each electrically-driven air deflector 109 reaches an extension direction towards the inside of the box body and the horizontal direction in the selected group of operation parameters.

[0039] In summary, the application provides a refrigerated container control method based on artificial intelligence, which can stack refrigerated stacked cargos 111 from the center of the refrigerated container to the periphery, and the refrigerated stacked cargos 111 need to maintain a certain distance from the inner wall of the refrigerated container to ensure the air flow passage inside. Further, the type and weight of the refrigerated stacked cargos 111 to be transported in the box body input by the user on the man-machine interaction panel 108 can be received, and the first distance value collected by the first distance sensor 105, the first distance value collected by the second distance sensor 106 and the third distance value collected by the third distance sensor 107 can be received. The first distance value is subtracted by the preset distance threshold value to obtain the first distance difference (i.e. the distance that one of the first side panels 103 can move towards the inside of the box body after reserving the air flow passage distance), the second distance value is subtracted by the preset distance threshold value to obtain the second distance difference (i.e. the distance that the other first side panel 103 can move towards the inside of the box body after reserving the air flow passage distance), and the third distance value is subtracted by the preset distance threshold value to obtain the third distance difference (i.e. the distance that the other first side panel 103 can move towards the inside of the box body after reserving the air flow passage distance).

[0040] The first distance difference, the second distance difference, and the third distance difference correspond to the foldable angles, and the minimum foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference is folded along the folding line by driving each first side panel 103 according to the control of the driving motor. In this way, after the refrigerated container is folded, each first side panel 103 and the top panel 101 are kept at a certain distance from the inner wall of the refrigerated container.

[0041] According to the type of the refrigerated stacked goods 111, the refrigeration temperature required by the refrigerated stacked goods 111 is determined. The required refrigeration temperature can save power consumption without deteriorating the refrigerated stacked goods 111.

[0042] The type, weight, and refrigeration temperature of the refrigerated stacked goods 111 are input into the pre-trained operation parameter determination model to determine a plurality of sets of operation parameters of the intelligent refrigerated container for keeping each position on the refrigerated stacked goods 111 at the refrigeration temperature, each set of operation parameters including the refrigeration power of the refrigeration equipment, the air outlet speed and direction, and the angle between the extension direction of each electrically operated air deflector 109 to the inside of the box and the horizontal direction. In this way, each position on the refrigerated stacked goods 111 can be kept at the required refrigeration temperature and uniform cooling and heating.

[0043] The refrigeration power and air outlet speed in each set of operation parameters are input into the pre-trained power consumption determination model to determine the power consumption of the refrigeration equipment corresponding to each set of operation parameters; the set of operation parameters corresponding to the minimum power consumption of the refrigeration equipment is selected. According to the refrigeration power of the refrigeration equipment, the air outlet speed and direction in the selected set of operation parameters, the refrigeration equipment is controlled to operate; and each electrically operated air deflector 109 is controlled to rotate so that each electrically operated air deflector 109 reaches the angle between the extension direction to the inside of the box and the horizontal direction in the selected set of operation parameters. In this way, each position on the refrigerated stacked goods 111 can be kept at the required refrigeration temperature and uniform cooling and heating, and the power consumption is minimized, while the occupied space of the refrigerated container is saved, and the practicability is strong.

[0044] In some embodiments, each first side panel 103 is provided with a non-contact temperature sensor 110 on both sides of the folding line thereof, and the geometric center of the box is provided with a contact temperature sensor (for example, provided on a bracket extending into the geometric center of the box, not shown in the drawings). After S208, the method provided by the present application further comprises: Step 1: receiving the temperature of the refrigerated stacked goods 111 collected by each non-contact temperature sensor 110 and receiving the temperature of the refrigerated stacked goods 111 collected by the contact temperature sensor.

[0045] Step 2: determining the actual average temperature of the plurality of received temperatures of the refrigerated stacked goods 111.

[0046] Step 3: determining a difference temperature between the actual average temperature and the refrigeration temperature required by the refrigerated stacked goods 111.

[0047] Step 4: in the case that the difference temperature is greater than the set temperature threshold, updating the first network parameter of the running parameter determination model according to the first reinforcement learning model until the difference temperature is less than or equal to the set temperature threshold, wherein the difference temperature is the reward of the first reinforcement learning model, and the lower the difference temperature, the higher the reward, the first network parameter of the running parameter determination model is the action of the first reinforcement learning model, and the first network parameter is the state of the first reinforcement learning model.

[0048] Based on the above steps 1-4, the actual average temperature of the refrigerated stacked goods 111 can be accurately controlled to the required temperature.

[0049] In some embodiments, in the case that the difference temperature is less than or equal to the set temperature threshold, the method provided by the embodiments of the present application further comprises: Step A: determining the variance of the plurality of temperatures of the received refrigerated stacked goods 111.

[0050] Step B: in the case that the variance of the plurality of temperatures of the refrigerated stacked goods 111 is greater than the preset variance threshold, updating the second network parameter of the running parameter determination model according to the second reinforcement learning model until the variance of the plurality of temperatures of the refrigerated stacked goods 111 is less than or equal to the preset variance threshold, wherein the variance of the plurality of temperatures of the refrigerated stacked goods 111 is the reward of the second reinforcement learning model, and the smaller the variance of the plurality of temperatures of the refrigerated stacked goods 111, the higher the reward, the second network parameter of the running parameter determination model is the action of the second reinforcement learning model, and the second network parameter is the state of the second reinforcement learning model.

[0051] Based on the above, the refrigerated stacked goods 111 at each position can be accurately brought to the required refrigeration temperature and uniformly cooled and heated.

[0052] In addition, the present application also provides an artificial intelligence-based refrigerated container control device configured in the main controller of an intelligent refrigerated container. It should be noted that the basic principle and technical effects of the artificial intelligence-based refrigerated container control device provided by the embodiments of the present application are the same as those of the above-mentioned embodiments. For brevity, the part not mentioned in the embodiments of the present application can refer to the corresponding content in the above-mentioned embodiments.

[0053] As Figure 1As shown, the intelligent refrigerated container includes a box body, the box body includes a top plate 101 and a bottom plate 102 arranged oppositely, two first side panels 103 arranged oppositely, and two second side panels arranged oppositely, each first side panel 103 is foldable towards the inside of the box body, and the folding line of each first side panel 103 when folded is parallel to the top plate 101 or the bottom plate 102 of the box body; a first distance sensor 105 and a second distance sensor 106 are respectively arranged on the folding line of the inner side of the two first side panels 103, a third distance sensor 107 is arranged on the inner side of the top plate 101, a human-computer interaction panel 108 is arranged on the outer side of the top plate 101, and each first side panel 103 is respectively provided with an electrically operated air deflector 109 on both sides of the folding line thereof, and the inner side of one of the second side panels is provided with refrigeration equipment. The device provided by the embodiment of the application comprises: A data receiving unit is configured to receive the type and weight of the refrigerated stacked goods 111 to be transported in the box body input by a user on the human-computer interaction panel 108, and receive the first distance value of the refrigerated stacked goods 111 collected by the first distance sensor 105, the first distance value of the refrigerated stacked goods 111 collected by the second distance sensor 106, and the third distance value of the refrigerated stacked goods 111 collected by the third distance sensor 107; A distance difference determining unit is configured to subtract a preset distance threshold value from the first distance value to obtain a first distance difference, subtract the preset distance threshold value from the second distance value to obtain a second distance difference, and subtract the preset distance threshold value from the third distance value to obtain a third distance difference; A foldable angle determining unit is configured to determine the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference respectively, and control the driving motor to drive each first side panel 103 to fold along the folding line by the smallest foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference; A refrigeration temperature determining unit is configured to determine the refrigeration temperature required by the refrigerated stacked goods 111 according to the type of the refrigerated stacked goods 111; An operating parameter determining unit is configured to input the type, weight, and refrigeration temperature of the refrigerated stacked goods 111 into a pre-trained operating parameter determining model to determine multiple sets of operating parameters of the intelligent refrigerated container for maintaining the refrigeration temperature at each position on the refrigerated stacked goods 111, each set of operating parameters including the refrigeration power, air outlet speed, and air outlet direction of the refrigeration equipment, and the included angle between the extension direction of each electrically operated air deflector 109 towards the inside of the box body and the horizontal direction, and the operating parameter determining model is obtained by training a first training network with multiple first training samples, each first training sample including the historical type, historical weight, and historical refrigeration temperature of a historical refrigerated stacked goods 111, and historical operating parameters of the intelligent refrigerated container for maintaining the historical refrigeration temperature at each position on the historical refrigerated stacked goods 111. a power consumption determination unit, configured to input the refrigeration power and the air outlet speed in each group of operation parameters into a pre-trained power consumption determination model to determine the power consumption of the refrigeration device corresponding to each group of operation parameters; an operation parameter selection unit, configured to select a group of operation parameters corresponding to the minimum power consumption of the refrigeration device; a control unit, configured to control the refrigeration device to operate according to the refrigeration power, the air outlet speed and the air outlet direction of the refrigeration device in the selected group of operation parameters, and control each electrically-driven air deflector 109 to rotate so that each electrically-driven air deflector 109 reaches an extension direction towards the inside of the cabinet and a horizontal direction at an included angle in the selected group of operation parameters.

[0054] In some embodiments, each first side panel 103 is provided with a non-contact temperature sensor 110 on both sides of the folding line thereof, and the geometric center of the cabinet is provided with a contact temperature sensor, The data receiving unit is further configured to receive the temperature of the refrigerated stacked goods 111 collected by each non-contact temperature sensor 110 and receive the temperature of the refrigerated stacked goods 111 collected by the contact temperature sensor; The apparatus provided by the embodiments of the present application further includes: an average temperature determination unit, configured to determine an actual average temperature of the plurality of received temperatures of the refrigerated stacked goods 111; a difference temperature determination unit, configured to determine a difference temperature between the actual average temperature and the refrigeration temperature required by the refrigerated stacked goods 111; a parameter updating unit, configured to update the first network parameter of the operation parameter determination model according to the first reinforcement learning model when the difference temperature is greater than a set temperature threshold, until the difference temperature is less than or equal to the set temperature threshold, wherein the difference temperature is the reward of the first reinforcement learning model, and the lower the difference temperature, the higher the reward, the first network parameter of the operation parameter determination model is the action of the first reinforcement learning model, and the first network parameter is the state of the first reinforcement learning model.

[0055] In some embodiments, the apparatus provided by the embodiments of the present application further includes: a temperature variance determination unit, configured to determine the variance of the plurality of received temperatures of the refrigerated stacked goods 111; The parameter updating unit is further configured to update the second network parameter of the running parameter determination model according to the second reinforcement learning model in a case where the variance of the temperatures of the plurality of refrigerated stacked goods 111 is greater than a preset variance threshold, until the variance of the plurality of temperatures of the refrigerated stacked goods 111 is less than or equal to the preset variance threshold, wherein the variance of the plurality of temperatures of the refrigerated stacked goods 111 is a reward of the second reinforcement learning model, and the smaller the variance of the plurality of temperatures of the refrigerated stacked goods 111, the higher the reward, the second network parameter of the running parameter determination model is an action of the second reinforcement learning model, and the second network parameter is a state of the second reinforcement learning model.

[0056] In some embodiments, the foldable angle determination unit is specifically configured to: find, according to a preset first mapping relationship, a foldable angle corresponding to the first distance difference and the second distance difference from a preset database; and find, according to a preset second mapping relationship, a foldable angle corresponding to the third distance difference from the preset database.

[0057] In some embodiments, the first network to be trained is a neural network.

[0058] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A refrigerated container control method based on artificial intelligence, characterized in that: A main controller for an intelligent refrigerated container, the intelligent refrigerated container comprising a box body, the box body comprising an oppositely arranged top plate and bottom plate, two oppositely arranged first side panels, and two oppositely arranged second side panels, each of the first side panels being foldable toward the interior of the box body, and a folding line of each first side panel when folded is parallel to the top plate or the bottom plate of the box body; a first distance sensor and a second distance sensor are respectively provided on the folding lines on the inner sides of the two first side panels, a third distance sensor is provided on the inner side of the top plate, a human-machine interaction panel is provided on the outer side of the top plate, each of the first side panels is provided with an electric air deflector on both sides of its folding line, and a refrigeration device is provided on the inner side of one of the second side panels, the method comprising: receiving the type and weight of the refrigerated stacked goods to be transported in the box input by the user on the human-computer interaction panel, and receiving the first distance value from the refrigerated stacked goods collected by the first distance sensor, the first distance value from the refrigerated stacked goods collected by the second distance sensor, and the third distance value from the refrigerated stacked goods collected by the third distance sensor; Subtracting a preset distance threshold from the first distance value to obtain a first distance difference, subtracting the preset distance threshold from the second distance value to obtain a second distance difference, and subtracting the preset distance threshold from the third distance value to obtain a third distance difference; determining foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference, respectively, and controlling a drive motor to drive each first side panel to fold the first side panel along the folding line to a minimum foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference; Determining the refrigeration temperature required for the refrigerated stacked goods according to the type of the refrigerated stacked goods; Inputting the type, weight, and refrigeration temperature of the refrigerated stacked goods into a pre-trained operating parameter determination model to determine multiple sets of operating parameters of the intelligent refrigerated container that enable each position on the refrigerated stacked goods to maintain the refrigerated temperature, each set of operating parameters including the refrigeration power, air outlet speed, and air outlet direction of the refrigeration equipment, and the angle between the extension direction of each electric air deflector toward the interior of the box and the horizontal direction. The operating parameter determination model is obtained by inputting multiple first training samples into a first to-be-trained network for training, each first training sample including the historical type, historical weight, and historical refrigeration temperature of the historical refrigerated stacked goods, and the historical operating parameters of the intelligent refrigerated container that enable each position on the historical refrigerated stacked goods to maintain the historical refrigeration temperature; Inputting the cooling power and air outlet speed in each set of the operating parameters into a pre-trained power consumption determination model to determine the power consumption of the refrigeration equipment corresponding to each set of the operating parameters; wherein the power consumption determination model is trained by inputting a plurality of second training samples into a second to-be-trained network, each of the second training samples including a historical cooling power and a historical air outlet speed in the historical operating parameters of the refrigeration equipment, and the corresponding historical power consumption; Select the minimum power consumption of the cooling equipment and the corresponding set of operating parameters; The operation of the refrigeration equipment is controlled according to the refrigeration power, air outlet speed and air outlet direction of the refrigeration equipment in a selected set of operating parameters; and the rotation of each of the electric air guide plates is controlled so that each of the electric air guide plates reaches the angle between the extension direction toward the inside of the box and the horizontal direction in the selected operating parameters.

2. The method according to claim 1, characterized in that Each of the first side panels is provided with a non-contact temperature sensor on both sides of the folding line thereof, and a contact temperature sensor is provided at the geometric center of the box. After controlling the operation of the refrigeration device according to the refrigeration power, air outlet speed, and air outlet direction of the refrigeration device, and controlling the rotation of each of the electric air deflectors so that each of the electric air deflectors reaches a determined angle between its extension direction toward the inside of the box and the horizontal direction, the method further includes: receiving the temperature of the refrigerated stacked goods collected by each of the non-contact temperature sensors and receiving the temperature of the refrigerated stacked goods collected by the contact temperature sensor; determining an actual average temperature of a plurality of temperatures received for the refrigerated stacked goods; determining a temperature difference between the actual average temperature and a required refrigeration temperature for the refrigerated stacked goods; When the temperature difference is greater than a set temperature threshold, a first network parameter of the operating parameter determination model is updated according to the first reinforcement learning model until the temperature difference is lower than or equal to the set temperature threshold, wherein the temperature difference is a reward of the first reinforcement learning model, and the lower the temperature difference, the higher the reward. Updating the first network parameter of the operating parameter determination model is an action of the first reinforcement learning model, and the first network parameter is the state of the first reinforcement learning model.

3. The method according to claim 2, characterized in that When the temperature difference is lower than or equal to a set temperature threshold, the method further includes: determining a variance of a plurality of temperatures received for the refrigerated stacked items; When the variance of the temperatures of the multiple refrigerated stacked goods is greater than a preset variance threshold, the second network parameter of the operating parameter determination model is updated according to the second reinforcement learning model until the variance of the multiple temperatures of the refrigerated stacked goods is less than or equal to the preset variance threshold, wherein the variance of the multiple temperatures of the refrigerated stacked goods is the reward of the second reinforcement learning model, and the smaller the variance of the temperatures of the multiple refrigerated stacked goods, the higher the reward, updating the second network parameter of the operating parameter determination model is the action of the second reinforcement learning model, and the second network parameter is the state of the second reinforcement learning model.

4. The method according to claim 1, wherein The determining the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference, respectively, includes: According to a preset first mapping relationship, searching a preset database for foldable angles corresponding to the first distance difference and the second distance difference, respectively; According to a preset second mapping relationship, the foldable angle corresponding to the third distance difference is searched from a preset database.

5. The method according to claim 1, wherein The first network to be trained is a neural network.

6. A refrigerated container control device based on artificial intelligence, characterized in that: A main controller configured for an intelligent refrigerated container, the intelligent refrigerated container comprising a box body, the box body comprising an oppositely arranged top plate and bottom plate, two oppositely arranged first side panels, and two oppositely arranged second side panels, each of the first side panels being foldable toward the interior of the box body, and a folding line of each first side panel when folded is parallel to the top plate or the bottom plate of the box body; a first distance sensor and a second distance sensor are respectively provided on the folding lines on the inner sides of the two first side panels, a third distance sensor is provided on the inner side of the top plate, a human-machine interaction panel is provided on the outer side of the top plate, each of the first side panels is respectively provided with an electric air deflector along both sides of its folding line, and a refrigeration device is provided on the inner side of one of the second side panels, the device comprising: a data receiving unit, configured to receive the type and weight of the refrigerated stacked goods to be transported in the box input by the user on the human-computer interaction panel, and receive the first distance value from the refrigerated stacked goods collected by the first distance sensor, the first distance value from the refrigerated stacked goods collected by the second distance sensor, and the third distance value from the refrigerated stacked goods collected by the third distance sensor; a distance difference determining unit, configured to subtract a preset distance threshold from the first distance value to obtain a first distance difference, subtract the preset distance threshold from the second distance value to obtain a second distance difference, and subtract the preset distance threshold from the third distance value to obtain a third distance difference; a foldable angle determining unit, configured to determine foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference, respectively, and drive each of the first side panels according to a control driving motor to fold the first side panels along the folding line to a minimum foldable angle among the foldable angles corresponding to the first distance difference, the second distance difference, and the third distance difference; a refrigeration temperature determination unit, configured to determine a refrigeration temperature required for the refrigerated stacked goods according to the type of the refrigerated stacked goods; an operating parameter determination unit, configured to input the type, weight, and refrigeration temperature of the refrigerated stacked goods into a pre-trained operating parameter determination model, and determine multiple sets of operating parameters of the intelligent refrigerated container that enable each position on the refrigerated stacked goods to maintain the refrigerated temperature, each set of operating parameters including the refrigeration power, air outlet speed, and air outlet direction of the refrigeration equipment, and an angle between an extension direction of each of the electric air deflectors toward the interior of the box and a horizontal direction, the operating parameter determination model being trained by inputting multiple first training samples into a first to-be-trained network, each first training sample including a historical type, historical weight, and historical refrigeration temperature of historical refrigerated stacked goods, and historical operating parameters of the intelligent refrigerated container that enable each position on the historical refrigerated stacked goods to maintain the historical refrigeration temperature; a power consumption determination unit, configured to input the cooling power and air flow rate in each set of the operating parameters into a pre-trained power consumption determination model, and determine the power consumption of the refrigeration equipment corresponding to each set of the operating parameters, wherein the power consumption determination model is trained by inputting a plurality of second training samples into a second to-be-trained network, each of the second training samples including a historical cooling power and a historical air flow rate in the historical operating parameters of the refrigeration equipment, and the corresponding historical power consumption; An operating parameter selection unit for selecting a set of operating parameters corresponding to the minimum power consumption of the refrigeration equipment; A control unit is used to control the operation of the refrigeration equipment according to the refrigeration power, air outlet speed and air outlet direction of the refrigeration equipment in a selected set of operating parameters; and to control the rotation of each of the electric air guide plates so that each of the electric air guide plates reaches the angle between the extension direction toward the inside of the box and the horizontal direction in the selected operating parameters.

7. The device according to claim 6, characterized in that Each of the first side panels is provided with a non-contact temperature sensor on both sides of its folding line, and a contact temperature sensor is provided at the geometric center of the box. The data receiving unit is further configured to receive the temperature of the refrigerated stacked goods collected by each of the non-contact temperature sensors and the temperature of the refrigerated stacked goods collected by the contact temperature sensor; The device further comprises: an average temperature determining unit for determining an actual average temperature of the plurality of temperatures received for the refrigerated stacked goods; a temperature difference determination unit, configured to determine a temperature difference between the actual average temperature and a refrigeration temperature required for the refrigerated stacked goods; a parameter updating unit, configured to update, when the temperature difference is greater than a set temperature threshold, a first network parameter of the operating parameter determination model according to the first reinforcement learning model until the temperature difference is lower than or equal to the set temperature threshold, wherein the temperature difference is a reward of the first reinforcement learning model, and a lower the temperature difference is, a higher the reward is; updating the first network parameter of the operating parameter determination model is an action of the first reinforcement learning model, and the first network parameter is a state of the first reinforcement learning model.

8. The device according to claim 7, characterized in that The device further comprises: a temperature variance determining unit, configured to determine a variance of a plurality of temperatures of the received refrigerated stacked goods; The parameter updating unit is further configured to update the second network parameter of the operating parameter determination model according to the second reinforcement learning model when the variance of the temperatures of the multiple refrigerated stacked goods is greater than a preset variance threshold, until the variance of the multiple temperatures of the refrigerated stacked goods is less than or equal to the preset variance threshold, wherein the variance of the multiple temperatures of the refrigerated stacked goods is a reward of the second reinforcement learning model, and the smaller the variance of the temperatures of the multiple refrigerated stacked goods, the higher the reward, updating the second network parameter of the operating parameter determination model is the action of the second reinforcement learning model, and the second network parameter is the state of the second reinforcement learning model.

9. The device according to claim 6, characterized in that The foldable angle determination unit is specifically used to search for the foldable angles corresponding to the first distance difference and the second distance difference from a preset database according to a preset first mapping relationship; and search for the foldable angle corresponding to the third distance difference from a preset database according to a preset second mapping relationship.

10. The device according to claim 6, characterized in that The first network to be trained is a neural network.

Images