KR20240076496A - Unmanned Parcel Storage Method and System for using Loading Optimization and Situation Assessment Module - Google Patents

Abstract

An unmanned parcel storage method and system using a loading optimization and situation determination module is presented. The unmanned parcel storage system using the loading optimization and situation judgment module proposed in the present invention is based on the image around the parcel box collected from the delivery box imaging device when a new parcel is received in the unmanned delivery box, and the user does not use the delivery box. A situation determination module that determines the status and, when in an idle state, a loading optimization module that updates the package loading information according to the loading space, weight of the package, and the package movement path, instructs loading movement, and changes to standby state after performing the loading movement. Includes.

Description

Unmanned Parcel Storage Method and System for using Loading Optimization and Situation Assessment Module}

The present invention relates to an unmanned parcel storage method and system using a loading optimization and situation determination module.

The unmanned delivery device (Public Patent No. 10-2013-0114495), which can dynamically store mail using a variable-structure unmanned delivery device according to the prior art [1], is an unmanned delivery box that dynamically allocates storage space, saving parcel storage space. Because it is allocated dynamically, it has the advantage of being able to use space as efficiently as possible.

In order to prevent the accumulation of errors over time in smart parcel storage boxes according to the prior art [2], an organizing method and device (disclosed to the public) checks through cameras or scanning equipment, obtains idle points through big data, and performs parcel organizing. Technologies such as (Patent No. 10-2021-0090186) can minimize storage-related defects by using variable loading space to organize during idle time, allowing for efficient storage of parcels. For example, errors in the number and size of parcels that occur due to long-term operation can be improved. However, depending on the situation, it is not possible to organize flexibly or receive delivery quickly.

An unmanned delivery management system (Public Patent No. 10-2022-0055250) that can manage the delivery by identifying the identity of the person entering through video and audio information, and outside the door through a doorbell, filming device, and monitoring device. Technologies such as a security system (Patent No. 10-2108185) that can determine the situation of the home and check whether a package has arrived outside the door are systems that can determine the user's information through video or receive the package rather than an unmanned delivery box. This system is used at receiving locations and needs to be applied to delivery boxes.

If the recipient is absent when the package arrives, a place where the recipient can receive the package on his/her behalf is an apartment management office. If there are no facilities such as an apartment management office, nearby convenience stores or shops are used to receive the package in the absence of the recipient. In the case of facilities that store parcels on behalf of others, they are not safe due to the inefficient waste of space and the risk of damage or theft during storage. An unmanned delivery box system that can safely manage parcels is also needed for proxy pickup. Existing unmanned delivery boxes are in the form of lockers with a fixed space, so they cannot store a lot of capacity and may result in waste of space. In urban areas, there may be a lack of space to install unmanned delivery boxes, especially in villas or residential areas. Therefore, there is a need for unmanned delivery boxes that can be space-efficient and optimally organized.

Public Patent No. 10-2013-0114495 Public Patent No. 10-2021-0090186 Public Patent No. 10-2022-0055250 Registered Patent No. 10-2108185

The technical task to be achieved by the present invention is to organize the internal arrangement of the delivery box to suit space efficiency through an optimal loading module, and to efficiently organize and provide service by determining idle time when the user does not use the delivery box through the situation judgment module. The goal is to provide a possible unmanned delivery storage method and system.

In one aspect, the unmanned parcel storage system using the loading optimization and situation determination module proposed in the present invention allows the user to deliver the parcel based on the image around the parcel box collected from the delivery box imaging device when a new parcel is received in the unmanned delivery box. The situation determination module determines the idle state when not in use, and when in the idle state, the parcel loading information is updated according to the loading space, the weight of the parcel, and the parcel movement path, and the loading movement is instructed and the loading movement is performed and then changed to a standby state. Includes a loading optimization module.

The situation determination module determines the existence of an object and its location on the image by detecting objects in the collected images around the delivery box, and the action recognition module recognizes the behavior of each detected object and the behavior of the object obtained through the action recognition module. It includes an idle state determination module that determines the idle state based on behavior recognition information.

The action recognition module is designed based on deep learning, and the internal layer includes Fully Connected Network (FCN), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Attention Module, and the behavior of the detected object. By recognizing and labeling the object's behavior recognition information and image data, it is used as learning data for transfer learning to resolve data dependency.

If the idle state determination module determines that the user is using the delivery box, it determines the idle state as False and switches to the standby state, and if it determines that the user is not using the delivery box, it determines the idle state as True. and provide services.

The loading optimization module includes an optimization algorithm performing unit that optimizes internal parcel placement based on pre-stored parcel loading information when the determination result through the situation determination module is in an idle state, and parcels to be moved after parcel placement is optimized according to the optimization algorithm. It includes a parcel movement path designation unit that performs load optimization by specifying the parcel movement path and instructing the automatic transport device to move the parcel.

The optimization algorithm execution unit receives map information about the delivery arrangement inside the current unmanned delivery box, the map information digitizes the actual delivery arrangement information, and the map structure is designed in 1D, 2D, or 3D depending on the shape of the actual delivery box. , Bin Packing algorithm is used to remove empty space between two parcels from the batch information.

The parcel movement path designation unit specifies the parcel movement path by minimizing the movement line of each parcel, selecting a location according to the weight of the parcel, and using the parcel movement line as a weight to minimize the power consumption required for the automatic transport device used to move the parcel. , Outputs map information that can give movement instructions according to changes in the location of each parcel.

In another aspect, the unmanned parcel storage method using the loading optimization and situation determination module proposed in the present invention determines the situation based on the images around the parcel box collected from the delivery box imaging device when a new parcel is received in the unmanned delivery box. Through the module, the user determines the idle state in which the delivery box is not used. If the delivery box is in an idle state, the loading optimization module updates the parcel loading information according to the loading space, the weight of the parcel, and the delivery route to instruct the loading movement. It includes the step of changing to a standby state after performing a loading movement.

According to embodiments of the present invention, parcel loading can be optimized without interfering with customer service use through an unmanned parcel storage system using image-based behavior recognition. In addition, in order to avoid interfering with the customer's use of the service, the idle time is determined by recognizing the behavior of objects in the video surrounding the delivery box received from the delivery box video device, and through this, the customer can use the unmanned delivery box service without any inconvenience. . In addition, it is possible to provide an energy-efficient and space-efficient unmanned delivery service by optimizing internal loading information by considering loading space, weight, and delivery path during idle time.

Figure 1 is a diagram showing the configuration of an unmanned parcel storage system using a loading optimization and situation determination module according to an embodiment of the present invention.
Figure 2 is a diagram showing the configuration of a situation determination module according to an embodiment of the present invention.
Figure 3 is a diagram for explaining transfer learning of a situation determination module according to an embodiment of the present invention.
Figure 4 is a diagram showing the configuration of a loading optimization module according to an embodiment of the present invention.
Figure 5 is a diagram for explaining the bin packing algorithm of the loading optimization module according to an embodiment of the present invention.
Figure 6 is a flowchart illustrating an unmanned parcel storage method using a loading optimization and situation determination module according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

Figure 1 is a diagram showing the configuration of an unmanned parcel storage system using a loading optimization and situation determination module according to an embodiment of the present invention.

The present invention proposes an unmanned parcel storage system using image-based behavior recognition that can optimize loading considering space and energy without interfering with customer service use. The proposed system is loaded into the situation determination module 110 and bin packing, which determines idle time by recognizing the user's behavior in the video surrounding the delivery box received from the delivery box video device so as not to interfere with the customer's use of the service. It includes a loading optimization module 120 that performs loading optimization to minimize operating energy efficiency in consideration of space, movement line, and weight.

The unmanned parcel storage system using the proposed loading optimization and situation determination module is a situation determination module (110) based on the image information collected from the delivery box imaging device (112) after a new parcel is received in the unmanned delivery box (111). Determine the idle state. In addition, when in an idle state, the loading optimization module 120 specifies a location to store the parcel in conjunction with the current loading information in the storage box, including the loading space, weight of the parcel, and parcel movement path, and automatically instructs the parcel loading movement information. Load movement is performed through the transport device 121. When the movement is completed, the loading information is updated (122) to reflect the organized contents and changes to the standby state (123). If a customer approaches in the middle of loading optimization and the situation determination module outputs that the idle state is False, only the optimization loading movement that is in progress at that time is performed and the system returns to the standby state (123) to minimize user inconvenience.

The unmanned parcel storage system using the proposed loading optimization and situation determination module includes a situation determination module 110 and a loading optimization module 120.

When a new parcel is delivered to the unmanned delivery box, the situation determination module 110 determines an idle state in which the user is not using the delivery box based on the image around the delivery box collected from the delivery box imaging device.

The situation determination module 110 includes an action recognition module and an idle state determination module.

The behavior recognition module detects objects in the collected images around the delivery box, determines the existence of the object and its location in the image, and recognizes the behavior of each detected object.

The action recognition module is designed based on deep learning, and the internal layers include Fully Connected Network (FCN), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Attention Module. The behavior of the detected object is recognized and the object's behavior recognition information and image data are labeled and used as learning data for transfer learning to resolve data dependency.

The idle state determination module determines the idle state based on the object's behavior recognition information obtained through the behavior recognition module.

If the idle state determination module determines that the user is using the delivery box, it determines the idle state as False and switches to the standby state.

If the idle state determination module determines that the user is not using the delivery box, it determines the idle state as True and provides the service.

When in an idle state, the loading optimization module 120 updates the package loading information according to the loading space, the weight of the package, and the package movement path, instructs loading movement, performs the loading movement, and changes to a standby state.

The loading optimization module 120 includes an optimization algorithm execution unit and a parcel movement path designation unit.

The optimization algorithm execution unit optimizes the internal parcel arrangement based on pre-stored parcel loading information when the determination result through the situation determination module is in an idle state.

The optimization algorithm execution unit receives map information about the parcel placement inside the current unmanned delivery box, and the map information digitizes the actual parcel placement information. The map structure is designed in 1D, 2D, or 3D depending on the shape of the actual delivery box, and a bin packing algorithm is used to remove the empty space between two parcels from the placement information.

The parcel movement path designator specifies the parcel movement path for the parcels to be moved after the parcel arrangement is optimized according to an optimization algorithm, and instructs the automatic transport device to move the parcels, thereby optimizing loading.

The parcel movement route designation unit specifies the parcel movement path by minimizing the movement path of each parcel, selecting a location based on the weight of the parcel, and using the parcel movement route as a weight to minimize the power consumption required by the automatic transport device used to move the parcel. It outputs map information that can give movement instructions according to changes in the location of each parcel.

Figure 2 is a diagram showing the configuration of a situation determination module according to an embodiment of the present invention.

When a new parcel is received in an unmanned delivery box, the situation determination module 210 according to an embodiment of the present invention determines an idle state in which the user does not use the delivery box based on the image around the delivery box collected from the delivery box imaging device. .

The situation determination module 210 determines whether the user will or will not use the delivery locker, receives images around the delivery box using imaging devices such as a camera, and determines situations such as whether the user is just passing by, standing, or approaching. . For judgment, the object is detected using a video object detection model, and the object's behavior is judged through a behavior recognition module. Based on the judgment results, the status of the user's use is judged.

The situation determination module 210 includes an action recognition module 211 and an idle state determination module 212.

Here, the delivery box imaging device that captures the image that is input data to the situation determination module is arranged to check the surroundings (e.g., in front) of the unmanned delivery box, and lighting can be used to ensure brightness at night. .

The behavior recognition module 211 determines the existence of an object and its location on the image through object detection in the collected images around the delivery box, and recognizes the behavior of each detected object.

The behavior recognition module 211 detects the presence of a user and its location in the image through object detection in the image surrounding the received delivery box, and recognizes the behavior of each detected user. Here, object (i.e. user) detection and action recognition are designed based on deep learning, and the internal layers include FCN (Fully Connected Network), CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), Attention Module, etc. Use .

The action recognition module 211 of the situation judgment module 210 has data-dependent characteristics because it uses deep learning-based technology. Because of this characteristic, the performance of action recognition may deteriorate in new environments. These factors can be caused by changes in background and lighting. Therefore, a method to improve performance at the installed location is essential. An example of the transfer learning process of the action recognition module 211 will be described with reference to FIG. 3.

The idle state determination module 212 determines the idle state based on the object's behavior recognition information obtained through the behavior recognition module 211.

If it is determined that the user is using the delivery box, the idle state determination module 212 determines the idle state as False and switches to the standby state.

If it is determined that the user is not using the delivery box, the idle state determination module 212 determines the idle state as True and provides the service.

The idle state determination module 212 determines the idle state based on human behavior recognition information obtained through the situation determination module 210. Here, the action recognition information includes text information such as 'approaching', 'standing', and 'walking' output through the action recognition module 211. Based on this information, information that is judged to be content that comes to use the service, such as 'approaching', determines the idle state to be False, and content that is judged to be content that does not use the service, such as 'standing' or 'walking', is determined to be false. The information determines the idle state as True. Through this, the idle time of the delivery box can be flexibly determined depending on the situation, allowing the service to be provided as soon as the user approaches to use the delivery box.

Figure 3 is a diagram for explaining transfer learning of a situation determination module according to an embodiment of the present invention.

The action recognition module 300 of the situation determination module according to an embodiment of the present invention has data-dependent characteristics because it uses deep learning-based technology. Because of this characteristic, the performance of action recognition may deteriorate in new environments. These factors can be caused by changes in background and lighting. Therefore, a method to improve performance at the installed location is essential. To solve this, data dependencies are resolved and a model suitable for the installation environment is updated through transfer learning that can respond to various situations.

The action recognition module 300 receives an input image 310 of a new region and new learning data 320 and performs transfer learning. For example, the data used in transfer learning is video data of a user who came to use a delivery service, and when the user manipulates the UI of the delivery box, the two data can be labeled and used as learning data for transfer learning. Transfer learning can be performed through the behavior recognition module 300 using the learning data, and then the recognized behavior 330 can be output as a result.

Figure 4 is a diagram showing the configuration of a loading optimization module according to an embodiment of the present invention.

If the situation determination module determines that the user is not using the delivery storage box, the loading optimization module 410 according to an embodiment of the present invention organizes the internal delivery arrangement using an optimization algorithm based on the current loading information in the storage box. . Load optimization is performed by creating a movement route for the parcels that will be moved after being organized, instructing the automatic transport device, and moving them. If the situation judgment module determines the user's approach, the optimal loading process that was intended to be executed is canceled and the parcel is moved back to the previous location.

When in an idle state, the loading optimization module 410 updates the package loading information according to the loading space, the weight of the package, and the package movement path, instructs loading movement, performs the loading movement, and changes to the standby state.

The loading optimization module 120 includes an optimization algorithm performing unit 411 and a parcel movement path designation unit 412.

The optimization algorithm performing unit 411 optimizes the internal parcel arrangement based on pre-stored parcel loading information when the determination result through the situation determination module is in the idle state. The optimization algorithm is run using the current storage bin loading information as input. The optimization algorithm operates in a way to increase energy efficiency by moving the minimum number of packages while eliminating the gap between packages on the shelves where the packages will be loaded as much as possible in order to load as many packages as possible into the internal space.

The optimization algorithm performing unit 411 receives map information about the current delivery arrangement inside the unmanned delivery box, and the map information digitizes the actual delivery arrangement information. The map structure is designed in 1D, 2D, or 3D depending on the shape of the actual delivery box, and a bin packing algorithm is used to remove the empty space between two parcels from the placement information. The bin packing algorithm according to an embodiment of the present invention will be described in more detail with reference to FIG. 5.

The parcel movement path designation unit 412 performs loading optimization by designating the parcel movement path for the parcels to be moved after the parcel arrangement is optimized according to an optimization algorithm and instructing the automatic transport device to move the parcels.

In order to minimize the power consumption required for the automatic transport device used to move the parcel, the parcel movement path designation unit 412 minimizes the movement line of each parcel, selects a position according to the weight of the parcel, and uses the parcel movement route as a weight to determine the parcel movement path. Specify and output map information that can give movement instructions according to changes in the location of each parcel.

Once the parcel's movement path is specified through an optimization algorithm, the automatic transport device is instructed to move the parcel. If the situation determination module determines the situation in which the user is approaching during operation of the loading optimization module, the movement of the parcel that was in progress at that point is stopped and loaded back to the previous location to prevent the loading information in the storage box from being stored incorrectly. Through these methods, it is possible to develop a space-efficient unmanned parcel storage system that can load as many parcels as possible.

Figure 5 is a diagram for explaining the bin packing algorithm of the loading optimization module according to an embodiment of the present invention.

The load optimization module to save surplus space is activated when the idle state is True. The load optimization module receives map information about the parcel placement inside the current unmanned delivery box. Here, the map information is stored in a data structure such as a list or tuple of information received from the delivery box control system in order to digitize the actual delivery placement information. The map structure can be designed in 1D, 2D, or 3D depending on the shape of the actual delivery box. And from this batch information, a bin packing algorithm is used to remove the empty space between two parcels.

The bin packing algorithm is an optimization technique that calculates the minimum number of bags needed to put as many items of various sizes as possible in a given space, and can be considered the most appropriate algorithm for optimizing loading space. Here, in order to minimize the power consumption required for the automatic transport device used to move parcels, an algorithm can be designed using weight and movement routes as weights, such as minimizing the movement path of each parcel and storing heavy parcels in a lower location. This not only organizes excess internal space but also saves power consumption. The output of the load optimization module is map information that can give movement instructions according to changes in the location of each parcel. Afterwards, the changed parcel loading information is reflected in the system by updating the loading information through map information, and the status of the unmanned parcel storage system is changed to standby. However, if a customer uses the delivery box during load optimization, it may cause inconvenience. Therefore, in order to minimize user inconvenience, when the situation determination module recognizes the idle state and becomes False, the movement instructions of the loading optimization module are executed only up to the current instruction and quickly return to the standby state to provide service.

Figure 6 is a flowchart illustrating an unmanned parcel storage method using a loading optimization and situation determination module according to an embodiment of the present invention.

The proposed unmanned parcel storage method using the stacking optimization and situation judgment module is based on the situation judgment module based on the image around the box collected from the box imaging device when a new package is received in the unmanned delivery box, preventing the user from using the box. In the step 610 of determining an idle state, and in the case of an idle state, the package loading information is updated according to the loading space, the weight of the package, and the package movement path through the loading optimization module to instruct loading movement, perform the loading movement, and then wait. It includes a step 620 of changing the state.

In step 610, when a new parcel is received in the unmanned delivery box, an idle state in which the user does not use the delivery box is determined through a situation determination module based on the image around the delivery box collected from the delivery box imaging device.

The situation determination module includes an action recognition module and an idle state determination module.

The behavior recognition module detects objects in the collected images around the delivery box, determines the existence of the object and its location in the image, and recognizes the behavior of each detected object.

The action recognition module is designed based on deep learning, and the internal layers include Fully Connected Network (FCN), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Attention Module. The behavior of the detected object is recognized and the object's behavior recognition information and image data are labeled and used as learning data for transfer learning to resolve data dependency.

The idle state determination module determines the idle state based on the object's behavior recognition information obtained through the behavior recognition module.

If the idle state determination module determines that the user is using the delivery box, it determines the idle state as False and switches to the standby state.

If the idle state determination module determines that the user is not using the delivery box, it determines the idle state as True and provides the service.

In step 620, when in an idle state, the parcel loading information is updated according to the loading space, the weight of the parcel, and the parcel movement path through the loading optimization module, a loading movement is instructed, and the loading movement is performed and then changed to a standby state.

The loading optimization module includes an optimization algorithm execution unit and a parcel movement path designation unit.

The optimization algorithm execution unit optimizes the internal parcel arrangement based on pre-stored parcel loading information when the determination result through the situation determination module is in an idle state.

The optimization algorithm execution unit receives map information about the parcel placement inside the current unmanned delivery box, and the map information digitizes the actual parcel placement information. The map structure is designed in 1D, 2D, or 3D depending on the shape of the actual delivery box, and a bin packing algorithm is used to remove the empty space between two parcels from the placement information.

The parcel movement path designator specifies the parcel movement path for the parcels to be moved after the parcel arrangement is optimized according to an optimization algorithm, and instructs the automatic transport device to move the parcels, thereby optimizing loading.

The parcel movement route designation unit specifies the parcel movement path by minimizing the movement path of each parcel, selecting a location based on the weight of the parcel, and using the parcel movement route as a weight to minimize the power consumption required by the automatic transport device used to move the parcel. It outputs map information that can give movement instructions according to changes in the location of each parcel.

The basic cell of the unmanned delivery storage system according to an embodiment of the present invention can be designed to enable time and movement optimization when delivering boxes between cells. For example, through an opening that can be opened and closed, a highly efficient loading and unloading method can be used for delivery drivers (e.g., delivery robots), or the delivery time (based on post office delivery boxes, bulk storage not possible) can be efficiently adjusted. there is.

The basic cell according to an embodiment of the present invention can be manufactured with a desired configuration including room temperature and refrigeration functions in one cell. In the case of room temperature cells, they can be used as individual room temperature cells, and functions can be added or changed. In the case of refrigeration cells, they can be used as individual refrigeration-only cells, and functions can be added or changed.

The basic cell according to an embodiment of the present invention can be designed and manufactured with an artificial intelligence (AI) variable rack structure. In addition, the rack is variable rather than fixed and can be adjusted to suit the size of the cargo, and can be managed by measuring cargo dimensions through a scanner mounted in the cargo reception area and transmitting this information to the operating system. By adding deep learning artificial intelligence (AI) technology to the proposed system, optimized cargo loading may be possible. For example, even after cargo storage, automatic rearrangement of cargo optimized by artificial intelligence (AI) is possible.

The connection type for each cell according to an embodiment of the present invention may be arranged so that each cell is single or multiple cells are connected. In addition, it can be deployed in a form of cell-by-cell connection using a more expanded area by utilizing multi-space including above ground and underground.

The device described above may be implemented with hardware components, software components, and/or a combination of hardware components and software components. For example, devices and components described in embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), etc. , may be implemented using one or more general-purpose or special-purpose computers, such as a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. A processing device may execute an operating system (OS) and one or more software applications that run on the operating system.　 Additionally, a processing device may access, store, manipulate, process, and generate data in response to the execution of software.　 For ease of understanding, a single processing device may be described as being used; however, those skilled in the art will understand that a processing device includes multiple processing elements and/or multiple types of processing elements. It can be seen that it may include.　 For example, a processing device may include a plurality of processors or one processor and one controller.　 Additionally, other processing configurations, such as parallel processors, are possible.

Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device.　 Software and/or data may be used on any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. It can be embodied in .　 Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium.　 The computer-readable medium may include program instructions, data files, data structures, etc., singly or in combination.　 Program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and available to those skilled in the art of computer software.　 Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -Includes optical media (magneto-optical media) and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc.　 Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.　

As described above, although the embodiments have been described with limited examples and drawings, various modifications and variations can be made by those skilled in the art from the above description.　 For example, the described techniques are performed in a different order than the described method, and/or components of the described system, structure, device, circuit, etc. are combined or combined in a different form than the described method, or other components are used. Alternatively, appropriate results may be achieved even if substituted or substituted by an equivalent.

Therefore, other implementations, other embodiments, and equivalents of the claims also fall within the scope of the claims described below.

Claims (14)

When a new parcel is delivered to the unmanned delivery box, a situation determination module that determines an idle state in which the user is not using the delivery box based on images around the delivery box collected from the delivery box imaging device; and
When in an idle state, the loading optimization module updates the parcel loading information according to the loading space, weight of the parcel, and parcel movement path, instructs loading movement, performs loading movement, and then changes to standby state.
An unmanned parcel storage system including. According to paragraph 1,
The situation determination module is,
A behavior recognition module that detects the presence of objects and their locations in the images through object detection in the collected images surrounding the delivery box, and recognizes the behavior of each detected object; and
An idle state determination module that determines the idle state based on the object's behavior recognition information obtained through the behavior recognition module.
An unmanned parcel storage system including. According to paragraph 2,
The behavior recognition module is,
It is designed based on deep learning, and the internal layers include Fully Connected Network (FCN), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Attention Module.
It recognizes the behavior of the detected object and uses it as learning data for transfer learning to resolve data dependency by labeling the object's behavior recognition information and image data.
Unmanned parcel storage system. According to paragraph 2,
The idle state determination module,
If it is determined that the user is using the delivery box, the idle state is set to False and switched to the standby state.
If it is determined that the user is not using the delivery box, the idle state is set to True and the service is provided.
Unmanned parcel storage system. According to paragraph 1,
The loading optimization module is,
an optimization algorithm performing unit that optimizes internal parcel placement based on pre-stored parcel loading information when the determination result through the situation determination module is in an idle state; and
After the parcel placement is optimized according to the optimization algorithm, the parcel movement path designation unit specifies the parcel movement path for the parcels to be moved and instructs the automatic transport device to move the parcels, thereby optimizing loading.
Unmanned parcel storage system including. According to clause 5,
The optimization algorithm performing unit,
Currently, map information about the delivery arrangement inside the unmanned delivery box is input, the map information digitizes the actual delivery arrangement information, the map structure is designed in 1D, 2D, or 3D according to the shape of the actual delivery box, and from the arrangement information Bin Packing algorithm is used to remove empty space between two parcels.
Unmanned parcel storage system. According to clause 5,
The courier movement route designation unit,
In order to minimize the power consumption of the automatic transport device used to move parcels, the movement path of each parcel is minimized, the location is selected according to the weight of the parcel, and the movement route of the parcel is used as a weight to specify the parcel movement path, and the position of each parcel is changed. Outputs map information that can give movement instructions according to
Unmanned parcel storage system. When a new parcel is delivered to the unmanned delivery box, determining an idle state in which the user does not use the delivery box through a situation determination module based on images surrounding the delivery box collected from the delivery box imaging device; and
When in an idle state, the package loading information is updated according to the loading space, weight of the package, and package movement path through the loading optimization module to instruct loading movement, perform the loading movement, and then change to standby state.
Method of storing unmanned parcels including. According to clause 8,
When a new parcel is received in the unmanned delivery box, the step of determining an idle state in which the user does not use the delivery box through a situation determination module based on the image around the delivery box collected from the delivery box imaging device is,
A step where the behavior recognition module of the situation determination module detects objects in the collected images around the delivery box to determine whether an object exists and its location on the image, and recognizes the behavior of each detected object; and
A step in which the idle state determination module of the situation determination module determines the idle state based on the behavior recognition information of the object obtained through the behavior recognition module.
Method of storing unmanned parcels including. According to clause 9,
The step of the behavior recognition module of the situation determination module detecting the presence of an object and its location on the image through object detection in the collected images around the delivery box, and recognizing the behavior of each detected object,
It operates based on deep learning, and the internal layers include FCN (Fully Connected Network), CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), and Attention Module.
It recognizes the behavior of the detected object and uses it as learning data for transfer learning to resolve data dependency by labeling the object's behavior recognition information and image data.
How to store unmanned parcels. According to clause 9,
The step of the idle state determination module of the situation determination module determining the idle state based on the behavior recognition information of the object obtained through the behavior recognition module,
If it is determined that the user is using the delivery box, the idle state is set to False and switched to the standby state.
If it is determined that the user is not using the delivery box, the idle state is set to True and the service is provided.
How to store unmanned parcels. According to clause 8,
In the case of the idle state, the step of updating the parcel loading information according to the loading space, weight of the parcel, and parcel movement path through the loading optimization module to instruct loading movement and changing to the standby state after performing the loading movement is,
A step of optimizing internal parcel placement based on pre-stored parcel loading information when the optimization algorithm execution unit of the loading optimization module determines that the situation determination module is in an idle state; and
A step in which the parcel movement path designation unit of the loading optimization module optimizes the parcel placement according to an optimization algorithm, then specifies the parcel movement path for the parcels to be moved and instructs the automatic transport device to move the parcels, thus performing loading optimization.
Method of storing unmanned parcels including. According to clause 12,
The step of optimizing internal parcel placement based on pre-stored parcel loading information when the optimization algorithm execution unit of the loading optimization module determines that the situation determination module is in an idle state includes:
Currently, map information about the delivery arrangement inside the unmanned delivery box is input, the map information digitizes the actual delivery arrangement information, and the map structure is designed in 1D, 2D or 3D according to the shape of the actual delivery box, and from the arrangement information Bin Packing algorithm is used to remove empty space between two parcels.
How to store unmanned parcels. According to clause 12,
The step of performing loading optimization by the parcel movement path designation unit of the loading optimization module optimizing the parcel placement according to an optimization algorithm, then specifying the parcel movement path for the parcels to be moved and instructing the automatic transport device to move the parcels,
In order to minimize the power consumption of the automatic transport device used to move parcels, the movement path of each parcel is minimized, the location is selected according to the weight of the parcel, and the movement route of the parcel is used as a weight to specify the parcel movement path, and the position of each parcel is changed. Outputs map information that can give movement instructions according to
How to store unmanned parcels.