JP7165902B2 - Comprehensive container terminal system and operation method using artificial intelligence - Google Patents

Description

The present invention relates to a container terminal system and operation method, and in particular, a container terminal system capable of optimizing the operation of a container terminal by using information such as cargo information from factories and shippers, land transportation information, and terminal information. and an operation method.

At the container terminal, cargo is loaded into containers by shipping operations from shippers and factories, and the containers are transported to the container yard one by one by trailers. At this time, in the conventional container terminal, since there is no information such as the congestion situation of the container terminal, people come early and wait for their turn at the entrance, resulting in traffic congestion. Furthermore, it took a long time at the entrance of the container yard to check the paper loading/unloading slips and check the container for damage. In the container yard, it was often difficult to predict when a container would come to pick it up, so there were many fluctuations in cargo handling and work volume. In addition, container yards needed to handle a large amount of cargo in a short period of time as ships became larger. Even in import operations, it is not clear which position in the yard is best for containers, and the setting of temporary installation locations for unloading is done only by the know-how of skilled workers. I had a problem with waiting. This is especially noticeable in large ships.

In response to these problems, in order to cope with the increase in size of container ships and to load and unload containers efficiently even during peak hours, advanced container terminals are equipped with remote-controlled cargo handling equipment, automatic guided vehicles (AGVs), etc. A vertical hangar has been introduced.

However, these efforts are limited to improving the efficiency of individual elemental technologies (Patent Documents 1 to 4), and information inside and outside the container terminal is not sufficiently captured, and overall optimization has not been achieved.

JP-A-2005-239377 Japanese Unexamined Patent Application Publication No. 2010-189128 JP 2013-67491 A JP-A-05-105241

The purpose of the present invention is to realize an autonomously evolving AI container terminal with the world's highest level of productivity by combining IoT (Internet of Things), AI (Artificial Intelligence), and automation. Another object of the present invention is to provide a system and an operation method for optimizing the container terminal system by transmitting and controlling information from the AI container terminal.

In order to achieve the above objectives, various types of information such as container information from factories and shippers, land transportation information of containers, information on container handling at container terminals and information on the operation of cargo handling machines are collected and captured as big data. From this big data, the necessary data for each terminal is read, and by autonomously learning the work of collection, accumulation, planning, and control, the key performance indicators are improved and the container terminal system is optimized ( deep learning), the results are centrally controlled and output to each container terminal.

Specifically, the container terminal system according to the present invention comprises a database storing data related to carrying in and out of containers, and collecting, accumulating, planning, and controlling data input/output between the database and the container terminal. It consists of an AI system that performs overall optimization and a terminal system that receives control values formulated by the AI system, and at least the AI system can autonomously learn past container terminal operation results.

In this case, in the terminal system, the loading/unloading data can be confirmed by the driver of the trailer, and can be used for transportation at the time of loading/unloading. Also, the output of control values from the AI system to the container terminal is the container storage plan.

Also, the output of the control values from the AI system to the container terminal is characterized in that it is a calculated value of the optimum timing of the trailer in the premises and an instruction to the container storage location in the yard.

A method of operating a container terminal system according to the present invention inputs and outputs data related to container loading and unloading from a database storing big data, and autonomously repeats collection and accumulation based on the input and output data for learning. By doing so, it is possible to plan and control and optimize the logistics, the terminal system can receive the control values formulated by the AI system, and at least the AI system evaluates and formulates the container terminal operation results, autonomously It makes it possible to learn The output optimized by the formulation/autonomous learning is a stored plan.

Therefore, by utilizing information technology and automating container damage checks, it is possible to greatly reduce the waiting time for the chassis when loading and unloading cargo. In addition, in the terminal, remote control and automation can achieve high efficiency with a minimum system and equipment, and minimize cargo handling. In addition, it is possible to reduce the cargo handling time of the vessel, and it is possible to greatly reduce the waiting time for cargo handling related to the vessel in the terminal.

BRIEF DESCRIPTION OF THE DRAWINGS It is drawing which shows the structural example of the container terminal system which concerns on embodiment of this invention. This is an application example of one terminal of the same system.

A specific embodiment of a container terminal system according to the present invention will be described in detail below with reference to the drawings.
It should be noted that the embodiment can be freely modified as long as the technical idea of the present invention is not changed.

FIG. 1 is an overall configuration diagram of a container terminal system according to an embodiment.
The configuration of this system is that a platform 10 is provided with a database 12 and an AI system 14, where external information (vehicle information, customs clearance information, vanning information, weather information, etc.) and terminal internal information (cargo handling data, cargo handling, etc.) machine data, cargo data, etc.). A database 12 and an AI system 14 are provided for each terminal 16 (suffixes A, B, C, . . . ) to enable individual operations. The database 12 may be configured across a plurality of units and may be shared. Each terminal 16 (16A, 16B, 16C, ...) is provided with a terminal system 18 (18A, 18B, 18C, ...) constituting a terminal operation system (TOS). 20 (20A, 20B, 20C, . . . ), outpatient trailers 22 (22A, 22B, 22C, . . In addition, the terminal system 18 is instructed to create a yard storage plan, a ship plan, a gate system, and other plans constructed by the AI system 14 .

Information input to the database 12 as external information includes vehicle information, customs clearance information, vanning information, weather information, and the like, and is big data including pre-input calendar information and event information. For example, VICS (registered trademark) information or the like is received to input road traffic conditions (traffic jam information, etc.), and further, information distributed by the Meteorological Agency is received and input as weather information. In addition, as carry-in and carry-out information, it is electronically registered (conventionally, paper is sent to the terminal by FAX), but this carry-in and carry-out information includes vehicle number, driver information, container information (container number, B/L number, ETC in-vehicle device number, booking number, etc.) are included. Vehicle registration information (specification of position) grasped by ETC is received and used as vehicle position information, and container number, item, weight, etc. are electronically registered as vanning information.

On the other hand, as internal information from the terminal 16, the terminal system 18 that constitutes each terminal 16 selects cargo handling data, cargo handling machine data, cargo data, etc. inside the terminal 16 for the AI system 14, which is stored in the database. sent to 12.

The AI system 14 reads from the database 12 that stores big data, collects and accumulates the necessary data, analyzes and plans it with a neural network, learns repeatedly and autonomously, and determines key performance indicators. is improved, and the yard storage plan, vessel plan, etc. are centrally controlled as control values and output to the terminal system 18 constituting the TOS.

One container terminal 16, the AI system 14 that controls it, and the database 12 that reads big data will now be described with reference to FIG.

A trailer 30 transports a container 32 from a factory 26 or a shipper 28 to the container terminal 16 and is about to enter the terminal 16 through a gate 34 placed at the entrance of the container terminal 16 .

Here, the database 12 is automatically entered in advance with location information, cargo information, etc., including the vehicle body number, and the acquired carry-in/carry-out information is processed by time zone in the TOS of the terminal system 18 of each container terminal 16. The information is updated in real time for trailer drivers to see and be able to.

The trailer driver confirms the content of the trailers, and the schedule for arriving at the site is set in consideration of the concentration of trailers at the container terminal 16, thereby greatly reducing the waiting time at the gate.

In addition, on the terminal system 18 side of the terminal 16, the number of trailers staying near the gate 34 is grasped by collating the trailers that have passed near the terminal by ETC with the trailers that have actually passed through the gate, and the information In the same way as above, the inflow is expected to be leveled by making it public through the Internet.
Therefore, at the gate 34, gate waiting time is greatly reduced by leveling the inflowing vehicles.

In addition, by comparing the loading/unloading information registered in advance with the customs clearance completion information grasped by the terminal 16 side, it is instantly determined at the gate 34 whether the container 32 can be loaded/unloaded. , which greatly reduces the gate processing time.

As for the containers 32 which have been collated in advance, the storage location is determined when they pass through the gate 34, which reduces the time required to determine the storage location when they arrive at the gate. , the gate processing time is greatly reduced.
As a result, the use of information technology can greatly reduce the waiting time for the chassis when loading and unloading cargo.

Furthermore, at the gate 34, the chassis waiting time at the time of cargo loading and unloading can be greatly reduced by automating the container damage check. This will change the work of checking damage when carrying out containers, which is often done by direct visual inspection by workers, to the work of taking pictures of containers with cameras or laser irradiation and analyzing the images with AI. to reduce work time. However, in the confirmation by image, the container is screened as to whether or not confirmation by a person is necessary. This screening greatly reduces the time required to pass through the gate when unloading the container.

Next, within the terminal 16, remote control and automation will be used to achieve high efficiency and minimize cargo handling by optimizing the organization and arrangement of equipment.
First, necessary data are extracted from the database 12 within the container terminal 16 . This includes, for example, data related to the characteristics of the container 32 (number, size, type, actual/empty, export/import, shipper, item, weight, etc.) and information such as the destination of the container 32 (loading/loading port, departure/arrival of the container). port, etc.), vessel information related to the container (ship name, flight number, etc.), scheduled delivery date and time of the import container 32 (including unknown), past delivery results (previously known delivery date, actual delivery date, etc.) , Vessel stowage plan (including sequence, special cargo, etc.), yard plan, etc. are used. It includes the cargo handling information of the vessel, gate information, and yard cargo handling information, which are updated as needed. In addition, external factors such as date, day of the week, weather, external events, road traffic conditions (traffic information, etc.), gantry crane (G/C), RTG (Rubber Tired Gantry Crane) operator status, crane position information, or status Including such.

It extracts such data and forwards this information to the AI system 14 . The transferred AI system 14 formulates an optimization scheme. For example, a storage plan for the container 32 is formulated so as to minimize the following. At this time, the storage plan and container carry-in/out information in past container terminal operations are taken in as actual values. For example, based on the loading and unloading patterns of each shipper for each day of the week and each type of cargo, we will devise ways to store them in places where there is less cargo handling. For example, cargo that is likely to be picked up soon is placed in the upper part of the container, and cargo that is unlikely to be picked up for a while is placed in the lower part. Save it to 12. These are the marshalling rate, cargo handling machine work load, cargo handling rate, mooring time of the main ship 36, waiting time of the foreign trailer 38 at the premises, maintenance and management costs of the cargo handling machines, and the like. These reflect the actual cargo handling situation (vessel, outpatient, terminal) and are revised as needed.

Then, based on the past results, we can predict the loading/unloading time by shipper and item, predict the trend/uncertainty of the vessel plan by route, and predict the loading/unloading speed of G/C, RTG, etc. In addition, it predicts the concentration of foreign chassis, and evaluates and predicts those with unknown carry-in/out schedules based on past performance.

When this work is completed, the following output data is sent from the AI system 14 to the terminal system 18. Yard plan, instructions to on-premises chassis (storage location (row, column, height) of import containers in the yard, route to storage location and route from storage location to gate), import container with unknown shipping date and time storage location in the yard (row, column, height), instructions to foreign trailers (storage location in the yard of export containers, route to storage location, and route from storage location to gate).

As a result, a storage plan for the entire yard is formulated that allows cargo handling on the main ship 36 and cargo handling on the foreign trailer 38 to be compatible and minimizes cargo handling.
At the time of import, when the import container 32 is stored in the yard from directly under the GC, AI is used based on the scheduled date and time of subsequent carry-out to improve the cargo handling capacity (cargo handling machines operate without waste), and the entire yard storage location (row, column, height) of imported containers in the yard, work plan/layout plan of cargo handling equipment, and yard An overall storage plan can be determined.

In addition, when there is an import container with an unknown shipping date and time during cargo handling on the ship 36, AI is used based on data related to the characteristics of the container (shipper, item, etc.) and big data on past shipping records. Then, after estimating the date and time of the trailer's arrival at the terminal, based on the estimated arrival date and time, the storage location of the imported container in the yard is determined so as to minimize the number of cargo handling (minimize the number of cargo handling). can.

As a result, the following effects are expected.
Utilize the entire container yard by leveling the layout of the cargo handling machines, which eliminates interference between them and improves the cargo handling capacity of the cargo handling machines, as well as by improving the cargo handling capacity for external chassis during cargo handling onboard ships. By doing so, storage capacity can be improved. In addition, costs can be reduced by reducing the time required for marshalling work and moving cargo handling equipment.

Import containers unloaded from the ship can be loaded and placed in a reasonable storage location according to the order of the trailers that come to carry them out. Therefore, wasteful work of cargo handling machines such as RTGs in the yard and waiting time in the terminal for trailers can be reduced.

Since instructions can be given accurately to cargo handling machine operators without relying on experience or intuition, the burden required for decisions on the operators can be reduced.
Based on the booking list held at the time of importing export containers, the loading plan of the ship is estimated, and based on the estimated loading plan, AI is utilized to maximize the work efficiency of each cargo handling machine. storage locations (rows, columns, heights) of export containers in the yard, as well as cargo-handling equipment Work plans, layout plans, and storage plans for the entire yard can be determined.

As a result, by loading export containers in a reasonable storage location within the yard according to the loading location on the ship, the same effect as in the case of export containers can be expected.
In addition, it will be possible to significantly reduce the cargo handling time of the ship, and the waiting time for cargo handling in the terminal will be close to zero.

The processing capacity (working time per container) of the gantry crane (operator) is registered in the TOS of the terminal system 18 . Hatch cover handling data, G/C current work status, work log (timing of operator change, which container is being held, whether it is being moved), G/C operator information, G/C cargo handling speed, ship container position, RTG processing capacity, position, status information, RTG operator information, RTG cargo handling speed, vessel cargo handling plan (created in advance and location), premises trailer location (real-time GPS location, premises trailer status information (whether or not it is being worked on), premises trailer driver information, consideration of the impact of weather conditions on the movement speed to the chassis , container loading and unloading information (vanning registration information, bonded transportation approval information, CLS information), etc. are registered in the TOS, etc. Location information of foreign trailers, gate entry and exit times, gate allocation status, vehicle information at the gate (full transportation or not) Empty transportation?), and the line-up situation in front of the gate (the situation in front of the gate and around the terminal by camera analysis).

These data are extracted from the database 12 or IoT, and the data are transferred to the AI system 14 at regular intervals to formulate an optimization processing scheme.
Based on the ship's cargo handling status (G/C operation status, ship cargo handling plan), the order of G/C cargo handling tasks is predicted (assuming past data is deep-learned) and calculated.

Based on the above calculation, from the arrival time to the G / C from the on-premises chassis situation (position, status) and congestion avoidance, to minimize the "moving distance + waiting time" of the on-premises chassis, to the on-premises chassis for each cargo handling task instruct. The movement distance can be determined by calculating the shortest route (distance) based on the yard map from the position grasped by GPS or the like and the position of the G/C. Also, the waiting time can be calculated from the operation status of the G/C based on the expected arrival time and the task order of the G/C.

Transfer the above data to the TOS of the terminal system 18 . As a result, the optimal waiting timing of the on-premises trailer is calculated, the on-premises trailer is instructed to the storage location in the yard, and the RTG is instructed on the timing of the loading and unloading work for the on-premises trailer and the foreign trailer. In addition, it is applied to the calculation of the number of personnel input to the container terminal on a daily basis and the gate layout.
Therefore, it is possible to realize an equipment/personnel arrangement instruction and prediction system that makes the most of the processing capacity of the G/C using AI.

After fully understanding the processing capacity of each terminal and each GC, the number of trailers to be placed, the number of GCs in operation, and the processing capacity of operators (whether new or Calculation of optimal waiting timing for yard trailers that change depending on conditions such as container stowage position, ship position, weather conditions (tidal effects, etc.), container storage position, yard trailer It is possible to indicate the storage location in the yard.

Regarding the handling of foreign trailers, comprehensively determine the timing for foreign trailers to go to the loading/unloading place in the yard based on the work situation in the yard during cargo handling (timing for the on-site trailer to go to the loading/unloading place in the yard). Then, the RTG is instructed on the timing of loading and unloading work to the on-site trailer and outsourced trailer (in the case of export, instructions are also given to the outsourced trailer about the storage location in the yard). Thereby, the handling of in-house and out-of-town trailers can be balanced.

Based on past data such as the loading and unloading information of foreign trailers, the cargo handling information of the ship, and the operation status of trailers in the premises, we calculate the results on a daily basis. By making it possible to predict the number of trailers, it is possible to change the number of personnel entering the terminal and the gate layout (the number of lanes for loading and unloading empty containers and for loading and unloading full containers) on a daily basis.

All of the above work results will be shared and centralized in a database. By evaluating this with AI and continuously optimizing each judgment index at each terminal, the overall efficiency of operations at terminals connected to the system can be realized.

In this way, according to this embodiment, IoT utilization at a container terminal will improve the efficiency of gate processing and reduce driver waiting time by automatically identifying container and vehicle information. We aim to visualize and unify port information such as improving logistics efficiency and expanding handling capacity by consolidating cargo handling etc. on the platform. AI is used to optimize the movement of tens of thousands of containers and cargo handling machines, while accumulating and inheriting the skills of experts to formulate tacit knowledge, and to automatically detect and check container damage. Try to improve quality and work environment. In addition, by adopting automated transportation of containers and platooning of trailers, it is possible to improve the efficiency of container transportation and alleviate road congestion. effect can be achieved by

10...platform, 12...database, 14...AI system, 16...container terminal, 18...terminal system, 20...cargo handling machine, 22...outside trailer, 24...inside chassis.

Claims (3)

Collecting and accumulating necessary terminal internal data input/output between a database in which big data including vehicle information, customs clearance information, vanning information, and weather information related to container loading and unloading are stored in advance, and the data input/output between the database and the necessary terminal internal data. It consists of an AI system that optimizes logistics by planning and controlling each task with a neural network, and a terminal system that receives the control values formulated by the AI system,
At least, the AI system autonomously learns past container terminal operation results and reflects the actual cargo handling situation to make corrections as needed to calculate the optimal cargo handling timing for trailers in the premises and the container storage location in the yard. As an instruction, formulate a storage plan for the entire yard that minimizes the cargo handling that consists of the above-mentioned work while balancing the cargo handling of the ship and the cargo handling of the external trailer, and output this to the terminal as a container storage plan control value . A container terminal system that becomes possible. 2. The container terminal system according to claim 1, wherein said terminal system allows a driver of a trailer to confirm the loading/unloading data so that it can be used for transportation at the time of loading/unloading. We input and output big data including vehicle information, customs clearance information, vanning information, and weather information related to container loading and unloading from the database that stores big data in advance, and collect and store the necessary data inside the terminal including cargo handling data. The neural network plans and controls for each task to optimize logistics, and the terminal system can receive the control values formulated by the AI system, and at least the AI system evaluates the container terminal operation results and formulates By making autonomous learning possible and reflecting the actual cargo handling situation, it can be corrected as needed to calculate the optimum timing of cargo handling work for trailers in the premises and instruct the location of container storage in the yard. An operation method of a container terminal system for formulating a storage plan for the entire yard that minimizes cargo handling including the above-mentioned work while simultaneously handling foreign trailers, and outputting it to the terminal as a container storage plan control value .

Images