JP7461908B2 - Information processing method, information processing system, information processing device, and work vehicle - Google Patents

Description

The present invention relates to an information processing method, an information processing system, an information processing device, and a work vehicle.

For example, one example of a work vehicle that transports cargo is a garbage collection vehicle equipped with a loading device that compresses garbage using a compression plate while loading it into a garbage storage box (see, for example, Patent Document 1).

JP 2009-234738 A

In conventional refuse collection vehicles, the operation of the loading device is controlled according to a predefined operation dictionary, and the operation control cannot be flexibly changed depending on the usage situation, etc.

The present invention aims to provide an information processing method, an information processing system, an information processing device, and a work vehicle that can update an action dictionary that is referenced when controlling the operation of an on-board work device.

An information processing method according to one aspect of the present invention is a method for processing data acquired from a work vehicle having a storage unit, a working unit that loads or unloads a load from the storage unit, a memory unit that stores an action dictionary that is referenced when controlling the operation of the working unit, and a control unit that controls the operation of the working unit by referring to the action dictionary, and the method includes the steps of: acquiring work data relating to the loading or unloading operation from the work vehicle; generating update data for updating the action dictionary based on the acquired work data; and outputting the generated update data to the work vehicle in order to update the action dictionary provided in the work vehicle, by a computer.

An information processing system according to one aspect of the present invention is an information processing system including a work vehicle that transports a load and an information processing device communicatively connected to the work vehicle, the work vehicle including a storage unit that stores a load, a working unit that loads or unloads the load into the storage unit, a memory unit that stores an action dictionary that is referenced when controlling the operation of the working unit, and a control unit that controls the operation of the working unit by referring to the action dictionary, the information processing unit including an acquisition unit that acquires work data related to the loading or unloading work from the work vehicle, a generation unit that generates update data for updating the action dictionary based on the acquired work data, and an output unit that outputs the generated update data to the work vehicle to update the action dictionary of the work vehicle, and the work vehicle includes an update unit that updates the action dictionary stored in the memory unit based on the update data received from the information processing device.

An information processing device according to one aspect of the present invention includes an acquisition unit that acquires work data relating to the loading or unloading work from a work vehicle that includes a storage unit, a work unit that loads or unloads a load into or from the storage unit, a memory unit that stores an action dictionary that is referenced when controlling the operation of the work unit, and a control unit that controls the operation of the work unit by referring to the action dictionary, a generation unit that generates update data for updating the action dictionary based on the acquired work data, and an output unit that outputs the generated update data to the work vehicle in order to update the action dictionary included in the work vehicle.

A work vehicle according to one aspect of the present invention includes a storage unit for storing a load, a working unit for loading or unloading the load into or from the storage unit, a memory unit for storing an action dictionary referenced when controlling the operation of the working unit, a control unit for controlling the operation of the working unit by referring to the action dictionary, an output unit for outputting action data relating to the loading or unloading operation to an external device, an acquisition unit for acquiring update data for the action dictionary from the external device configured to generate update data for the action dictionary based on the action data, and an update unit for updating the action dictionary stored in the memory unit based on the acquired update data.

According to this application, it is possible to update the action dictionary that is referenced when controlling the operation of a work device.

1 is a schematic diagram showing a configuration of an information processing system according to a first embodiment; FIG. 1 is a side view showing a refuse collection vehicle as an example of a work vehicle. FIG. 2 is a side cross-sectional view showing the configuration of a working device provided on a refuse collection vehicle. FIG. 1 is a hydraulic circuit diagram of a garbage collection vehicle. FIG. 2 is a block diagram showing the configuration of a control system of a garbage collection vehicle. FIG. 13 is a conceptual diagram illustrating an example of an action dictionary. FIG. 2 is a block diagram showing the internal configuration of the business server. 2 is a block diagram showing an internal configuration of a management server. FIG. 13 is a conceptual diagram illustrating an example of a customer management table. FIG. 13 is a conceptual diagram illustrating an example of a work history table. FIG. 2 is a schematic diagram showing a configuration of an update learning model. 13 is a flowchart illustrating a procedure for generating an update learning model. 11 is a flowchart illustrating a procedure of a process executed by a management server according to the first embodiment. FIG. 13 is a schematic diagram showing an example of an action dictionary in the second embodiment. 13 is a flowchart illustrating a procedure of a process executed by a management server according to the second embodiment. FIG. 13 is a conceptual diagram illustrating an example of an update data management table. 13 is a flowchart illustrating a procedure for updating the action dictionary in the third embodiment. 13 is a flowchart illustrating a procedure of a proposal process. 13 is a flowchart illustrating a procedure for updating an action dictionary in the fourth embodiment. FIG. 13 is a schematic diagram showing an example of an action dictionary provided in a sewage collection vehicle in embodiment 5.

Hereinafter, the present invention will be described in detail with reference to the drawings showing embodiments thereof.
(Embodiment 1)
1 is a schematic diagram showing the configuration of an information processing system according to embodiment 1. The information processing system according to embodiment 1 includes a sewage collection vehicle 1, a business site server 200, and a management server 300.

The refuse collection truck 1, an example of a work vehicle that transports cargo, is a vehicle that travels around garbage collection stations set up in various locations and collects (transports) the garbage collected at the collection stations as cargo. The refuse collection truck 1 loads garbage by referring to an action dictionary described below, and collects and sequentially stores data related to the loading work (work data). The refuse collection truck 1 is equipped with a communication means for communicating with the business server 200. After discharging the collected garbage at a garbage treatment facility, the refuse collection truck 1 returns to the business and transmits the collected work data to the business server 200.

The business site server 200 is a server installed at the business site of the waste collection company. The business site server 200 stores the work data received from the refuse collection vehicle 1 in association with the identification information of the refuse collection vehicle 1 and the identification information of the worker. The business site server 200 may automatically create a daily work report DR for the refuse collection vehicle 1 based on the work data received from the refuse collection vehicle 1. The business site server 200 has a communication means for communicating with the management server 300 via a communication network NW such as the Internet network NW2. The business site server 200 transmits the work data received from the refuse collection vehicle 1 to the management server 300 together with the identification information of the refuse collection vehicle 1 and the identification information of the worker.

The management server 300 is a server installed in the management center. The management center is an institution that manages the sewage collection vehicle 1 used by each business establishment. The management center may lend the sewage collection vehicle 1 to a user (business establishment or worker) by subscription or lease. The management server 300 stores the work data received from the business establishment server 200 in association with the identification information of the sewage collection vehicle 1 and the identification information of the worker. The management server 300 generates update data for updating the action dictionary equipped in the sewage collection vehicle 1 based on the stored work data, and transmits the generated update data to the sewage collection vehicle 1. The update data transmitted from the management server 300 reaches the sewage collection vehicle 1 via the business establishment server 200.

The refuse collection vehicle 1 receives update data sent from the management server 300 via the business server 200. The refuse collection vehicle 1 updates its action dictionary based on the received update data. That is, the refuse collection vehicle 1 can obtain an action dictionary suitable for the work situation from the management server 300. The refuse collection vehicle 1 can optimize various actions during loading work by using the updated action dictionary.

Figure 2 is a side view showing a refuse collection vehicle 1 as an example of a work vehicle, and Figure 3 is a side cross-sectional view showing the configuration of a working device equipped on the refuse collection vehicle 1. The refuse collection vehicle 1 comprises a chassis 2 and a working device 3 provided on the chassis 2. The working device 3 comprises a loading device 10 equipped on a refuse input box 4 and an unloading device 20 equipped on a refuse storage box 5. In the following description, the front-rear, left-right, and up-down directions refer to the front-rear, left-right, and up-down directions as seen by a worker seated in the driver's seat of the refuse collection vehicle 1.

An opening 51 is provided on the rear surface of the trash storage box 5. The front surface of the trash input box 4 is partially open and the trash input box 4 is arranged so as to communicate with the trash storage box 5. An input port 41 is opened on the rear surface of the trash input box 4, and a trash storage chamber 42 is formed at the bottom inside the trash input box 4. Inside the trash input box 4, a loading device 10 is installed to crush and compress the trash input from the input port 41 inside the storage chamber 42 and load it into the trash storage box 5.

The loading device 10 includes a deflector 11 that is laid at an angle from the upper front end of the trash bin 4 toward the lower rear end, a sliding plate 12 that is provided so as to be able to move up and down along the deflector 11, and a compression plate 13 that is connected to the lower end of the sliding plate 12 and can rotate in the front-rear direction. The loading device 10 includes a pair of left and right lifting cylinders 12C that are connected to the side wall of the trash bin 4 and the sliding plate 12 as actuators for lifting and lowering the sliding plate 12 (see FIG. 4). The loading device 10 also includes a pair of left and right rotating cylinders 13C that are connected to the sliding plate 12 and the compression plate 13 as actuators for rotating the compression plate 13 (see FIG. 4).

The loading device 10 performs a cycle of operations including a reversal process (indicated by arrow A in FIG. 3) in which the rotating cylinder 13C is driven so that the compression plate 13 rotates backward, a lowering process (indicated by arrow B in FIG. 3) in which the lifting cylinder is driven so that the sliding plate 12 descends, a compression process (indicated by arrow C in FIG. 3) in which the rotating cylinder 13C is driven so that the compression plate 13 rotates forward, and a lifting process (indicated by arrow D in FIG. 3) in which the lifting cylinder 12C is driven so that the sliding plate 12 ascends, thereby compressing the garbage introduced into the storage chamber 42 and loading it into the garbage storage box 5.

In the compression and lifting processes, if the pressure acting on the loading device 10 is greater than a threshold value, an inching operation may be performed. In the inching operation, the sliding plate 12 and the compression plate 13 are operated individually or simultaneously to apply an appropriate pressure to trash that does not easily deform, such as furniture and electrical appliances, and the trash can be compressed and loaded into the trash storage box 5.

In addition, a backflow prevention operation may be performed during loading operations. In normal loading operations, the garbage thrown into the garbage dump box 4 is compressed and loaded into the garbage storage box 5 through a reversing process, a lowering process, a compression process, and a rising process from a predetermined position after one cycle is completed. In contrast, in the backflow prevention operation, the timing of the reversing process is delayed, and the reversing process is performed while the lowering process is being performed. This backflow prevention operation causes part of the opening 51 of the garbage storage box 5 to be blocked by the sliding plate 12, preventing the garbage from flowing back from the garbage storage box 5 to the garbage dump box 4.

The discharge device 20 includes a discharge plate 21 that is provided inside the garbage storage box 5 so as to be able to move back and forth in the front-rear direction, and a discharge cylinder 21C that drives the discharge plate 21. The discharge plate 21 is formed so that its width is approximately equal to the inner width of the garbage storage box 5, and its height is approximately equal to the height from the bottom wall to the top wall of the garbage storage box 5. The internal space of the garbage storage box 5 is divided by the discharge plate 21 into a garbage storage space 5A and a non-storage space 5B. The discharge cylinder 21C is disposed in the non-storage space 5B of the garbage storage box 5. The front end of the discharge cylinder 21C is connected to a bracket 22 fixed to the front of the garbage storage box 5, and the rear end is connected to the lower part of the discharge plate 21. The discharge device 20 can move the discharge plate 21 back and forth in the front-rear direction inside the garbage storage box 5 by the extension and contraction of the discharge cylinder 21C. The discharge cylinder 21C shown in FIG. 3 is a three-stage cylinder that incorporates three cylinders with different diameters.

The discharge plate 21 is configured to advance to near the front of the garbage storage box 5 when the discharge cylinder 21C is in the most contracted state, and to retreat to a position near the opening 51 when the discharge cylinder 21C is in the most initial extended state. When the garbage storage box 5 is empty, the discharge device 20 controls the discharge cylinder 21C to be in the most extended state, and retreats the discharge plate 21 to a position near the opening 51. By performing loading work in this state, the garbage is stored in the storage space 5A of the garbage storage box 5 in a compressed state. As the amount of garbage stored in the storage space 5A increases, the discharge device 20 controls the discharge cylinder 21C to shorten little by little, and moves the discharge plate 21 to the front of the garbage storage box 5. As a result, the volume of the storage space 5A expands. In this way, the working device 3 performs loading work using the loading device 10 while adjusting the position of the discharge plate 21 using the discharge device 20, and can store the loaded garbage in a compressed state in the storage space 5A of the garbage storage box 5. The front-rear position of the discharge plate 21 is detected by the position sensor POS. When performing a lowering work to discharge the garbage stored in the garbage storage box 5, the discharge device 20 controls the discharge cylinder 21C to be in the most extended state, and moves the discharge plate 21 to the rear end position of the garbage storage box 5, thereby discharging the garbage stored in the storage space 5A to the outside.

The trash bin 4 is arranged to be tiltable around a pivot 52 behind the trash storage bin 5, and is configured to be able to be opened and closed by the action of a tilting cylinder 52C (see FIG. 4). The trash storage bin 5 also has a locking mechanism 53 for locking the trash bin 4, and is configured to be able to lock the trash bin 4 by the action of a locking cylinder 53C (see FIG. 4).

Next, the hydraulic circuit of the garbage collection vehicle 1 will be described.
4 is a hydraulic circuit diagram of the refuse collection vehicle 1. The hydraulic circuit of the refuse collection vehicle 1 includes a hydraulic pump P to which the driving force of the vehicle engine is transmitted via a PTO (Power Take Off) (not shown), and a hydraulic tank T for storing hydraulic oil. The hydraulic oil in the hydraulic tank T is pumped up by the hydraulic pump P and supplied to a control valve CV. The control valve CV includes a plurality of electromagnetic control valves. The electromagnetic control valves included in the control valve CV include an electromagnetic control valve V1 that controls the extension and retraction of the lift cylinder 12C, an electromagnetic control valve V2 that controls the extension and retraction of the rotation cylinder 13C, an electromagnetic control valve V3 that controls the extension and retraction of the discharge cylinder 21C, and an electromagnetic control valve V4 that controls the extension and retraction of the tilt cylinder 52C and the lock cylinder 53C.

The electromagnetic control valves V1 to V4 of the control valve CV are, for example, 6-port, 3-position electromagnetic directional control valves. The control valve CV is configured to switch the open/close ports of the electromagnetic control valves V1 to V4 in response to a control signal from the operation control unit 140 (see FIG. 5), which will be described later, and supply the hydraulic oil discharged from the hydraulic pump P to the desired cylinder (lift cylinder 12C, rotating cylinder 13C, discharge cylinder 21C, tilt cylinder 52C, lock cylinder 53C). This controls the switching of the extension/retraction operation of each cylinder, or stops the operation. The return hydraulic oil that passes through the control valve CV flows through the hydraulic piping RP on the recovery side, is filtered by the return filter RF, and is then recovered in the hydraulic tank T.

A hydraulic sensor PS1 for detecting the hydraulic pressure (hydraulic pressure of the loading device 10) is connected to the oil passage connected to the back pressure side oil chamber of the rotating cylinder 13C. A solenoid check valve V, a relief valve R, and a hydraulic sensor PS2 are connected to the oil passage connected to the back pressure side oil chamber of the discharge cylinder 21C. The solenoid check valve V has a solenoid and is normally biased to a position (blocking position) that prevents the back pressure side oil chamber of the discharge cylinder 21C from communicating with the hydraulic tank T. When the solenoid is energized, the back pressure side oil chamber of the discharge cylinder 21C is switched to a position (communication position) that communicates with the hydraulic tank T. When the solenoid check valve V is open, the back pressure side oil chamber of the discharge cylinder 21C is connected to the hydraulic tank T, and when the solenoid check valve V is closed, the back pressure side oil chamber of the discharge cylinder 21C is blocked from the hydraulic tank T. When the hydraulic pressure in the back pressure side oil chamber of the discharge cylinder 21C reaches or exceeds a set value, the relief valve R returns the hydraulic oil to the hydraulic tank T. The hydraulic pressure sensor PS2 is provided in an oil passage connected to the back pressure side oil chamber of the discharge cylinder 21C and is configured to detect the hydraulic pressure (hydraulic pressure of the discharge device 20).

Next, the configuration of a control system for the garbage collection vehicle 1 will be described.
5 is a block diagram showing the configuration of the control system of the sewage collection vehicle 1. The control system of the sewage collection vehicle 1 comprises an operation setting unit 120, an operation control unit 140, and an image processing unit 160. The operation setting unit 120, the operation control unit 140, and the image processing unit 160 are configured to be able to transmit and receive information to and from each other.

The operation setting unit 120 is a processing circuit for performing various operation settings based on information input from the outside or information stored in the memory unit 121, and includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The operation setting unit 120 expands a computer program previously stored in the ROM onto the RAM, and the CPU executes the computer program while referring to the input information or the information stored in the memory unit 121, thereby performing operation settings for the working device 3, etc. Here, the operation setting refers to determining a control setting value used when controlling the operation of the working device 3. The control setting value does not need to be a single value, and may include a control setting value (control pattern) determined according to the operating situation (conditions) of the sewage collection vehicle 1. The information set by the operation setting unit 120 is output to the operation control unit 140.

The operation setting unit 120 is not limited to the above configuration, and may be one or more processing circuits including a microcomputer, memory, etc. The operation setting unit 120 may also include a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, a counter that counts numbers, etc.

The operation setting unit 120 is connected to a memory unit 121, a garbage type selection operation unit 122, a display unit 123, a sound detection unit 124, an on-board weighing unit 125, a position acquisition unit 126, a communication unit 127, and the like.

The storage unit 121 is configured, for example, with a non-volatile memory such as an EEPROM (Electronically Erasable Programmable Read Only Memory). The storage unit 121 stores an action dictionary for the operation device 3, data collected during operation (operation data), and the like.

FIG. 6 is a conceptual diagram showing an example of an action dictionary. The action dictionary includes control setting values that are referenced when controlling the action of the working device 3. In the action dictionary, control setting values for the working device 3 are stored for each action type of the working device 3. For example, the action dictionary stores the movement restriction set pressure (P1, P2, P3, ...) of the discharge plate 21 and the opening time (T1, T2, T3, ...) of the solenoid check valve V that controls the expansion and contraction action of the discharge cylinder 21C as control setting values corresponding to the position (X1, X2, X3, ...) of the discharge plate 21 for the discharge plate action during loading work. In addition, the action dictionary stores the operation timing (Ta1, Ta2, Ta3, ...) of the sliding plate 12 and the operation timing (Tb1, Tb2, Tb3, ...) of the compression plate 13 as control setting values corresponding to the hydraulic value (Y1, Y2, Y3, ...) of the hydraulic sensor PS1 for the inching action during loading work. Furthermore, the action dictionary stores the timing for delaying the reversal of the compression plate 13 (i.e., the time Tc1, Tc2, Tc3, ... for lowering the sliding plate 12 before starting the reversal of the compression plate 13) as a control setting value corresponding to the ratio of falling garbage (Z1, Z2, Z3, ...) for the backflow prevention action during loading work. Note that while FIG. 6 shows an action dictionary in table format, the action dictionary is not limited to the table format, and may be in function format or database format.

When loading work is performed by the work device 3, the operation setting unit 120 reads out the movement restriction set pressure of the discharge plate 21 under each condition and the opening time of the solenoid check valve V that controls the extension and retraction operation of the discharge cylinder 21C as control settings and outputs them to the operation control unit 140. The operation control unit 140 controls the movement of the discharge plate 21 during loading work by controlling the movement restriction set pressure of the discharge plate 21 and the extension and retraction operation of the discharge cylinder 21C with reference to the operation setting value input from the operation setting unit 120. The same is true when performing the inching operation and backflow prevention operation, and the operation control unit 140 only needs to control the operation of the sliding plate 12 and the compression plate 13 with reference to the operation setting value input from the operation setting unit 120.

The work data stored in the memory unit 121 includes at least one of the work location, type of garbage, image data obtained from the imaging unit 161, hydraulic pressure value of the loading device 10 during loading work, hydraulic pressure value of the discharge device 20 when moving the discharge plate 21 during loading work, position of the discharge plate 21 provided on the discharge device 20, weight of the loaded garbage, and compression sound of the garbage during loading work.

The garbage type selection operation unit 122 includes a switch for accepting the type of garbage (garbage) to be collected. When garbage is put into the garbage dump box 4, or when the put-in garbage is loaded into the garbage storage box 5, the garbage type selection operation unit 122 is operated by the worker to select the type of garbage to be collected. The garbage type selection operation unit 122 includes a physical switch for each type of garbage, such as "food waste," "PET bottles," "wood," and "cardboard." Alternatively, the garbage type selection operation unit 122 may be a switch that switches the type each time it is pressed, a rotary switch that switches the type by rotating, or a software switch in which selection items are arranged on a display panel. In addition, instead of mounting the garbage type selection operation unit 122, a configuration may be used in which the selection of the type is accepted using a mobile terminal CP such as a smartphone carried by the worker, and information on the selected type is acquired by communicating with the mobile terminal CP of the worker. At this time, the communication unit 127 of the refuse collection vehicle 1 and the mobile terminal CP of the worker may be paired using a short-range wireless communication technology such as Bluetooth (registered trademark). The result of the selection of the waste type by the waste type selection operation unit 122 is transmitted to the business server 200 as part of the work data after the refuse collection vehicle 1 returns to the business.

The display unit 123 includes a display panel for displaying information to be notified to the operator. The display panel may be a liquid crystal panel or an organic EL panel. The display unit 123 can display, for example, information on the type of garbage selected by the garbage type selection operation unit 122. In addition, by arranging a touch panel on the display unit 123, it is possible to accept various selection operations and setting operations, including the type of garbage.

The sound detection unit 124 includes an input device such as a microphone that is placed inside the waste bin 4. For example, the sound detection unit 124 is provided near the inner bottom plate of the waste bin 4, between one metal side wall of the waste bin 4 and a side cover that covers that metal side wall. The sound detection unit 124 detects sounds such as compression sounds that occur during loading work through the input device. The sounds detected by the sound detection unit 124 are recorded as digital data and output to the operation setting unit 120. The sound data detected by the sound detection unit 124 is transmitted to the business server 200 as part of the work data after the waste collection vehicle 1 returns to the business.

The on-board weighing unit 125 is equipped with a measuring device for measuring the weight of the garbage contained in the garbage collection box 5. The on-board weighing unit 125 is equipped with, for example, a strain sensor that measures the distortion of the part where the load of the contents acts, and converts the amount of distortion measured by the strain sensor into the weight of the garbage and outputs it. The weight of the garbage weighed by the on-board weighing unit 125 is output to the operation setting unit 120. The weight data of the garbage weighed by the on-board weighing unit 125 is transmitted to the business server 200 as part of the work data after the garbage collection vehicle 1 returns to the business.

The position acquisition unit 126 includes, for example, a GPS receiver (Global Positioning System). The position acquisition unit 126 receives radio waves transmitted from a GPS satellite (not shown) using the GPS receiver and measures the current position of the sewage collection vehicle 1. The information on the current position of the sewage collection vehicle 1 measured by the position acquisition unit 126 is output to the operation setting unit 120. Instead of mounting the position acquisition unit 126, the position information of the sewage collection vehicle 1 may be measured using a mobile terminal CP such as a smartphone carried by a worker, and the measured position information of the sewage collection vehicle 1 may be acquired by communicating with the mobile terminal CP of the worker. The position information of the sewage collection vehicle 1 acquired by the position acquisition unit 126 is transmitted to the business server 200 as part of the work data after the sewage collection vehicle 1 returns to the business.

The communication unit 127 has a communication interface for wireless communication with the business server 200. The communication interface of the communication unit 127 is, for example, a communication interface conforming to a short-range wireless communication standard such as WiFi (registered trademark), Bluetooth (registered trademark), ZigBee (registered trademark), ANT (registered trademark), or ANT+ (registered trademark). The communication unit 127 wirelessly communicates with the business server 200 to obtain various data from the business server 200 and transmit work data related to loading work to the business server 200. The communication unit 127 may not only perform short-range wireless communication with the business server 200, but also perform short-range wireless communication with a mobile terminal CP carried by a worker. In this case, the mobile terminal CP may be equipped with a communication interface conforming to mobile communication standards such as 3G, 4G, 5G, and LTE (Long Term Evolution), and may be configured to transmit work data of the garbage collection vehicle 1 from the mobile terminal CP to the business server 200 or management server 300 by connecting the mobile terminal CP to the business server 200 or management server 300 via the mobile communication network NW1 and the Internet network NW2 (see FIG. 1).

The operation control unit 140 includes a processing circuit for controlling the operation of the work device 3 by referring to the operation setting values input from the operation setting unit 120. The operation control unit 140 is connected to a PTO switch 141, a loading switch 142, a discharge switch 143, various electromagnetic control valves V1 to V4, a solenoid check valve V, hydraulic sensors PS1 and PS2, a position sensor POS for the discharge plate 21, and the like.

When an instruction to start loading or unloading work is given by inputting various switches, the operation control unit 140 refers to the operation setting values input from the operation setting unit 120 and controls the operation of the various solenoid control valves V1 to V4 and the solenoid check valve V to carry out the loading or unloading work of the garbage.

The operation control unit 140 also acquires data (operation data) at any time, such as the timing of operation of the sliding plate 12, compression plate 13, discharge plate 21, etc. during loading or discharge work, the hydraulic value of the loading device 10 during loading work (hydraulic value measured by hydraulic sensor PS1), the hydraulic value of the discharge device 20 during loading work (hydraulic value measured by hydraulic sensor PS2), and the position of the discharge plate 21 measured by the position sensor POS, and outputs the acquired operation data to the operation setting unit 120. The operation data output to the operation setting unit 120 is stored in the memory unit 121, and is transmitted to the business server 200 after the refuse collection vehicle 1 returns to the business.

The image processing unit 160 includes a processing circuit for executing a process for recognizing a person (head) near the trash bin 4, gloves worn by a worker, and the like. The image processing unit 160 is connected to an imaging unit 161 and a recognition dictionary storage unit 162. The imaging unit 161 includes an imaging element for imaging the vicinity of the trash bin 4 where loading work is performed, and outputs image data obtained by imaging to the image processing unit 160. The imaging unit 161 may include an imaging element for imaging the outside of the trash bin 4, and an imaging element for imaging the inside of the trash bin 4.

The recognition dictionary storage unit 162 includes a learning model LM2 that has been trained to output information about the situation near the waste bin 4 when image data obtained from the imaging unit 161 is input as a recognition dictionary. The learning model LM2 is constructed by CNN (Convolutional Neural Networks). Alternatively, the learning model LM2 may be a learning model based on R-CNN (Region-based CNN), YOLO (You Only Look Once), SSD (Single Shot Detector), GAN (Generative Adversarial Network), SVM (Support Vector Machine), decision tree, etc. The learning model LM2 is generated by performing learning using a predetermined learning algorithm using a data set of image data obtained from the imaging unit 161 and label data indicating whether or not the image shown by the image data includes a person (head) to be recognized or gloves worn by a worker, etc., as training data.

When the image processing unit 160 acquires image data from the imaging unit 161, it inputs the acquired image data into the learning model LM2 and executes calculations using the learning model LM2 to acquire information about the situation near the waste bin 4. Based on the information obtained from the learning model LM2, the image processing unit 160 recognizes people (heads) and gloves worn by workers that are near the waste bin 4. The image data acquired from the imaging unit 161 may be transmitted to the business server 200 as part of the work data after the waste collection vehicle 1 returns to the business.

The image processing unit 160 is configured to recognize a person (head) present near the trash bin 4, gloves worn by a worker, etc., but may also be configured to recognize the amount of trash thrown into the trash bin 4, the type of trash, the movement of trash falling downward from the tip of the compression plate 13 while the loading device 10 is in the inversion process (the movement of trash flowing backward), etc. In this case, a learning model that has been trained to output information about the situation inside the trash bin 4 when image data obtained from the imaging unit 161 is input as the recognition dictionary is used. Such a learning model is prepared separately from the above-mentioned learning model LM2, and is generated by, for example, using a data set of image data obtained by imaging the inside of the trash bin 4 by the imaging unit 161 and correct answer data as training data and learning with a predetermined learning algorithm. This correct data includes the amount of increase in garbage calculated based on the weighing results of the on-board weighing unit 125, the garbage type selection result by the garbage type selection operation unit 122, and the relationship between the position of the compression plate 13 and the garbage falling downward from the tip of the compression plate 13.
The image processing unit 160 inputs image data obtained by imaging the inside of the garbage dump box 4 using the imaging unit 161 into the above-mentioned learning model, and recognizes the amount and type of garbage dumped into the garbage dump box 4, the amount of garbage backflow, etc., based on the information obtained from the learning model.

Next, the configuration of the establishment server 200 will be described.
7 is a block diagram showing the internal configuration of the establishment server 200. The establishment server 200 is a dedicated or general-purpose server device, and includes a control unit 201, a storage unit 202, a first communication unit 203, a second communication unit 204, an operation unit 205, and a display unit 206.

The control unit 201 includes a CPU, a ROM, a RAM, etc. The ROM included in the control unit 201 stores control programs and the like for controlling the operation of each of the above hardware components. The CPU in the control unit 201 executes the control programs stored in the ROM and the various programs stored in the storage unit 202, and controls the operation of each of the above hardware components. The RAM included in the control unit 201 stores data that is temporarily used while the various programs are being executed.

The control unit 201 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, and a counter that counts numbers.

The memory unit 202 includes storage devices such as a hard disk drive (HDD) and a solid state drive (SSD). The memory unit 202 stores various programs executed by the control unit 201, information on the refuse collection vehicle 1 managed by the establishment, and the like. As information on the refuse collection vehicle 1, information on the worker who uses the refuse collection vehicle 1, information on the patrol route that the refuse collection vehicle 1 travels to collect garbage, work data received from the refuse collection vehicle 1, and the like are stored in the memory unit 202. The memory unit 202 may also store a daily work report DR for the refuse collection vehicle 1 that is automatically created based on the work data received from the refuse collection vehicle 1.

The first communication unit 203 has a communication interface for wireless communication with the refuse collection vehicle 1. The communication interface of the first communication unit 203 is, for example, a communication interface conforming to a short-range wireless communication standard such as WiFi (registered trademark), Bluetooth (registered trademark), ZigBee (registered trademark), ANT (registered trademark), or ANT+ (registered trademark).

The second communication unit 204 has a communication interface for communicating with the management server 300. The communication interface provided by the second communication unit 204 is a communication interface for connecting to a communication network NW such as the Internet network NW2.

The operation unit 205 has an interface for connecting input devices such as a keyboard and a touch panel. The control unit 201 accepts various operations and settings by an administrator, etc., through the operation unit 205.

The display unit 206 is equipped with a display device such as a liquid crystal panel or an organic EL panel, and displays information to be notified to an administrator, etc., based on information output from the control unit 201.

The business server 200 stores the work data received from the garbage collection vehicle 1 in the memory unit 202 and transmits it to the management server 300 via the second communication unit 204.

Next, the configuration of the management server 300 will be described.
8 is a block diagram showing the internal configuration of the management server 300. The management server 300 is a dedicated or general-purpose server device, and includes a control unit 301, a storage unit 302, a communication unit 303, an operation unit 304, and a display unit 305.

The control unit 301 includes a CPU, a ROM, a RAM, etc. The ROM included in the control unit 301 stores control programs and the like for controlling the operation of each of the above hardware components. The CPU in the control unit 301 executes the control programs stored in the ROM and the various programs stored in the storage unit 302, and controls the operation of each of the above hardware components. The RAM included in the control unit 301 stores data that is temporarily used while the various programs are being executed.

The control unit 301 is not limited to the above configuration, and may be one or more processing circuits including a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), a DSP (Digital Signal Processor), a quantum processor, volatile or non-volatile memory, etc. The control unit 301 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when an instruction to start measurement is given to when an instruction to end measurement is given, and a counter that counts numbers.

The memory unit 302 includes storage devices such as an HDD and an SSD. The memory unit 302 stores various data and computer programs. The memory unit 302 includes a customer management table TB1 that stores, for example, identification information of a refuse collection vehicle 1 rented out by a subscription or lease system and identification information of a customer (business or worker) to whom the refuse collection vehicle 1 is rented in association with each other. FIG. 9 is a conceptual diagram showing an example of the customer management table TB1. In addition to identification information of the refuse collection vehicle 1 and identification information of the customer, the customer management table TB1 stores information such as vehicle type information, rental period, customer name, address, and email address. The vehicle type information is distinguished, for example, by the total weight of refuse that the refuse collection vehicle 1 can carry.

The memory unit 302 also includes a work history table TB2 that stores the work data of the garbage collection vehicle 1 received through the business server 200. FIG. 10 is a conceptual diagram showing an example of the work history table TB2. The work history table TB2 stores, in association with each other, for example, the date and time, the position information (collection location) of the garbage collection vehicle 1, the garbage type, the operation of the discharge plate 21 during loading work, the inching operation, the control setting value for the backflow prevention operation, etc., the hydraulic value of the hydraulic sensors PS1 and PS2, the number of loading cycles, the loading weight, the compression sound, the image data (frame image at each date and time) of the inside and outside of the garbage input box 4 captured by the imaging unit 161, etc. Here, the garbage type represents the type of "food waste", "PET bottles", "wood", "cardboard", etc. selected by the worker using the garbage type selection operation unit 122. The number of loading cycles is the number of times the loading operation is performed, with the inversion process, the lowering process, the compression process, and the rising process of the compression plate 13 being one cycle. The loaded weight is the weight of the garbage measured by the on-board weighing unit 125. The compression sound is the volume and prominent frequency of the sound detected by the sound detection unit 124 when the garbage is compressed. The image data is image data (frame images corresponding to each date and time) of the inside and outside of the garbage bin 4 captured by the imaging unit 161.

Such a work history table TB2 is created for each refuse collection vehicle 1. That is, the work history table TB2 is created by recording the work data collected when the refuse collection vehicle 1 travels around various locations to collect refuse and returns to the business premises, in association with the identification information of the refuse collection vehicle 1. Note that the work history table TB2 may record work data for each work day, or may record work data for multiple work days together.

The memory unit 302 also includes an update learning model LM11 for generating update data for the action dictionary. FIG. 11 is a schematic diagram showing the configuration of the update learning model LM11. The update learning model LM11 is trained to output a control setting value for the operation device 3 when operation data is input. The learning model LM11 is a learning model constructed using a recurrent neural network (RNN), a long short-term memory (LSTM), or the like. Alternatively, the learning model LM11 may be a learning model based on a CNN, an R-CNN, YOLO, an SSD, a GAN, an SVM, a decision tree, or the like.

The learning model LM11 is, for example, a machine learning learning model including deep learning, and is configured by a neural network. The learning model LM11 includes an input layer LY1, intermediate layers LY2 and LY3, and an output layer LY4. In the example of FIG. 11, two intermediate layers LY2 and LY3 are shown, but the number of intermediate layers is not limited to two and may be three or more.

The input layer LY1, the intermediate layers LY2 and LY3, and the output layer LY4 each have one or more nodes, and the nodes in each layer are connected in one direction to the nodes in the previous and next layers with desired weights and biases. The input layer LY1 of the learning model LM11 receives the same number of data as the number of nodes in the input layer LY1. In this embodiment, the data input to the nodes in the input layer LY1 includes work data such as the type of garbage, image data near the loading device 10, the hydraulic pressure values of the hydraulic sensors PS1 and PS2, the position of the discharge plate 21, the amount of garbage loaded (measured value), and the sound of garbage compression.

The input working data is output to the nodes in the first hidden layer LY2 through the nodes in the input layer LY1. The data input to the first hidden layer LY2 is output to the nodes in the next hidden layer LY3 through the nodes in the hidden layer LY2. At this time, the output is calculated using an activation function including the weights and biases set between the nodes. In the same manner, calculations are performed using activation functions including the weights and biases set between the nodes, and are transmitted to subsequent layers one after another until the calculation result is obtained by the output layer LY4.

Here, parameters such as weights and biases connecting the nodes are learned by a predetermined learning algorithm. For example, a deep learning learning algorithm including backpropagation is used as the learning algorithm for learning various parameters. In this embodiment, a data set including the above-mentioned work data and the correct answer data to be output by the learning model LM11 is used as training data, and various parameters including weights and biases between nodes are learned by a predetermined learning algorithm so that when work data is input, a control setting value is output. The data set including the work data and the correct answer data to be output by the learning model LM11 may be manually selected by the worker. For example, if the compression sound when compressing the garbage by the loading device 10 is small, the worker may determine that the garbage was smoothly compressed and stored in the garbage storage box 5, and may provide the control setting value at that time as the correct answer data. The control unit 301 of the management server 300 may also refer to the work history table TB2 and select the work data and control setting values (correct answer data) to be used for the training data from data such as the number of loading cycles, the loaded weight, and the compression sound. For example, the control unit 301 may calculate the weight of the garbage loaded in one cycle, and determine that the larger the weight, the smoother the garbage can be compressed and stored in the garbage storage box 5, and provide the control setting value at that time as the correct answer data. Also, if the compression sound when compressing the garbage is small, the control unit 301 may determine that the garbage can be smoothly compressed and stored in the garbage storage box 5, and provide the control setting value at that time as the correct answer data. A particularly important point is that when the type of garbage is identified, each control setting value that allows the maximum loading amount to be achieved with as few loading cycles as possible is set as the correct answer data.

Furthermore, the learning model LM11 may be a learning model based on reinforcement learning. In this embodiment, a reward is given based on the compression sound and the load weight included in the work data, and the value in reinforcement learning is learned so as to maximize the total reward obtained in the future. For example, in Q-learning, which is one type of reinforcement learning, a value Q for selecting an action (control setting value) under a state (work data) in a certain environment is learned. At the time when Q-learning is started, the correct value of value Q for a combination of a state (work data) and an action (control setting value) is not known at all. Therefore, various control setting values are selected under certain work data, the total reward is calculated based on the reward given for the action (control setting value) at that time, and the correct value Q is learned by selecting a better control setting value.

The output layer LY4 outputs control setting values as the calculation results. For example, the output layer LY4 may be configured to be composed of three nodes, from the first node to the third node, and configured to output the control setting value of the discharge plate 21 during loading operation from the first node, the control setting value for inching operation from the second node, and the control setting value for backflow prevention operation from the third node.

The input and output of the learning model LM11 can be set appropriately. For example, the input to the learning model LM11 may be one or more operation data selected from the collection location, type of garbage, image data near the loading device 10, hydraulic pressure values of the hydraulic sensors PS1 and PS2, the position of the discharge plate 21, the amount of garbage loaded (measured value), and the sound of garbage compression. The output of the learning model LM11 may be one or two control setting values selected from the control setting value of the discharge plate 21 during loading work, the control setting value for the inching operation, and the control setting value for the backflow prevention operation.

The communication unit 303 of the management server 300 has a communication interface for communicating with the business server 200. The communication interface provided in the communication unit 303 is a communication interface for connecting to a communication network NW such as the Internet network NW2.

The operation unit 304 has an interface for connecting input devices such as a keyboard and a touch panel. The control unit 301 accepts various operations and settings by an administrator or the like through the operation unit 304.

The display unit 305 is equipped with a display device such as a liquid crystal panel or an organic EL panel, and displays information to be notified to an administrator, etc., based on information output from the control unit 301.

The management server 300 does not have to be a single computer, but may be a computer system consisting of multiple computers and peripheral devices. Furthermore, the management server 300 may be a virtual machine that is virtually constructed by software.

Figure 12 is a flowchart explaining the procedure for generating the update learning model LM11. The control unit 301 of the management server 300 selects training data from the data stored in the work history table TB2 (step S101). For example, the control unit 301 may select work data that can be determined to have smoothly compressed and loaded the rubbish and the control settings at that time as the training data. Also, if there is training data selected by a worker, the control unit 301 may select that training data.

The control unit 301 inputs the work data contained in the selected training data into the updated learning model LM11 (step S102), and executes calculations using the updated learning model LM11 (step S103). It is assumed that initial values are set in advance for the parameters that characterize each node constituting the updated learning model LM11. As a result of the calculations of the updated learning model LM11, the control setting value of the discharge plate 21 during loading work, the control setting value for inching operation, and the control setting value for backflow prevention operation are obtained.

The control unit 301 evaluates the calculation result of the update learning model LM11 (step S104) and determines whether learning of the update learning model LM11 is complete (step S105). The control unit 301 can evaluate the calculation result using an error function (also called an objective function, loss function, or cost function) based on the calculation result obtained by the calculation of step S103 and the control setting value included as the correct answer data. For example, in the process of optimizing (minimizing or maximizing) the error function by a gradient descent method such as the steepest descent method, the control unit 301 determines that learning of the update learning model LM11 is complete when the error function becomes equal to or less than a threshold (or equal to or greater than a threshold).

If it is determined that learning is not complete (S105: NO), the control unit 301 updates the weights and biases between the nodes used in each layer (step S106) and returns the process to step S101. The control unit 301 can update the weights and biases between each node using the backpropagation method, which sequentially updates the weights and biases between nodes from the output layer LY4 to the input layer LY1.

If it is determined that learning is complete (S105: YES), the control unit 301 obtains the learned update learning model LM11, and stores this in the memory unit 302 (step S107).

The procedure for updating the action dictionary will be described below.
13 is a flow chart for explaining the procedure of the process executed by the management server 300 according to the first embodiment. The management server 300 acquires the work data of the sewage collection vehicle 1 (step S121). The management server 300 may acquire the work data of the sewage collection vehicle 1 through the business site server 200. The acquired work data is registered in the work history table TB2.

The control unit 301 inputs work data under certain conditions to the updated learning model LM11 (step S122) and executes calculations using the updated learning model LM11 (step S123). As a result of the calculations, the control setting value of the discharge plate 21 during loading work, the control setting value for inching operation, and the control setting value for backflow prevention operation are obtained. The control unit 301 can obtain the control setting value under each condition by inputting work data under various conditions to the updated learning model LM11.

The control unit 301 transmits the control setting values obtained from the update learning model LM11 as update data to the garbage collection vehicle 1 via the communication unit 303 (step S124). The update data transmitted from the management server 300 reaches the garbage collection vehicle 1 via the business server 200. When the garbage collection vehicle 1 receives the update data (control setting values) transmitted from the management server 300, it updates the operation dictionary by overwriting the control setting values stored as the operation dictionary with the newly received control setting values.

In the flowchart of FIG. 13, the control setting values to be stored in the action dictionary of the memory unit 121 of the garbage collection vehicle 1 are transmitted as update data, but the action dictionary may be created in the management server 300 and the created action dictionary may be transmitted as update data. In this case, the action dictionary stored in the memory unit 121 is updated by rewriting it with the newly received action dictionary.

In addition, the management server 300 is configured to transmit update data to the garbage collection vehicle 1 via the business server 200, but it may also be configured to transmit the update data directly to the garbage collection vehicle 1. Furthermore, in this embodiment, the management server 300 is configured to generate update data, but the update data may also be generated in the business server 200. In this case, the action dictionary of the garbage collection vehicle 1 can be updated by communicating between the business server 200 and the garbage collection vehicle 1.

The management server 300 is configured to include a communication unit 303 for transmitting update data to the garbage collection vehicle 1, but may also include a recording unit for recording the update data on a recording medium such as a USB memory. The recording medium on which the update data is recorded is provided to the business establishment. The business establishment server 200 may read the update data recorded on the recording medium using a reading device not shown in the figure, and transmit the read update data to the garbage collection vehicle 1. Alternatively, a reading device may be provided in the garbage collection vehicle 1, and the update data recorded on the recording medium may be read using this reading device, and the action dictionary may be updated based on the read update data.

As described above, in the first embodiment, update data is generated using the update learning model LM11, and the generated update data is used to update the operation dictionary of the sewage collection vehicle 1, so that appropriate operation control can be expected according to the usage status of the sewage collection vehicle 1.

(Embodiment 2)
In the second embodiment, a form in which the action dictionary is configured by a learning model will be described.

Figure 14 is a schematic diagram showing an example of an action dictionary in the second embodiment. The action dictionary in the second embodiment is composed of a learning model LM1. In this embodiment, the learning model LM1 generated in advance in the management server 300 is stored as an action dictionary in the memory unit 121 of the garbage collection vehicle 1. The generation procedure of the learning model LM1 is the same as the generation procedure of the update learning model LM11 described above. The learning model LM1 is trained to output the control setting value of the discharge plate 21 during loading work, the control setting value for inching operation, and the control setting value for backflow prevention operation in response to the input of work data under certain conditions. The operation setting unit 120 outputs the control setting value obtained from the learning model LM1 to the operation control unit 140. The operation control unit 140 can control the operation of the work device 3 (loading device 10) by referring to the control setting value input from the operation setting unit 120.

The procedure for updating the action dictionary (learning model LM1) will be described below.
15 is a flow chart for explaining the procedure of the process executed by the management server 300 according to the second embodiment. The control unit 301 of the management server 300 selects training data from the data stored in the work history table TB2 (step S201). For example, the control unit 301 may select, as the training data, work data that can be judged to have smoothly compressed and loaded the rubbish and the control setting values at that time. Also, if there is training data selected by an operator, the control unit 301 may select that training data.

To retrain the update learning model LM11, the control unit 301 inputs the work data included in the selected training data into the update learning model LM11 (step S202) and executes calculations using the update learning model LM11 (step S203). As a result of the calculations of the update learning model LM11, the control setting value of the discharge plate 21 during loading work, the control setting value for inching operation, and the control setting value for backflow prevention operation are obtained.

The control unit 301 evaluates the calculation results of the update learning model LM11 (step S204) and determines whether or not re-learning of the update learning model LM11 is complete (step S205). The control unit 301 can evaluate the calculation results using an error function (also referred to as an objective function, loss function, or cost function) based on the calculation results obtained from the update learning model LM11 and the control setting values included as the correct answer data. For example, in the process of optimizing (minimizing or maximizing) the error function using a gradient descent method such as the steepest descent method, the control unit 301 determines that re-learning of the update learning model LM11 is complete when the error function becomes equal to or less than a threshold (or equal to or greater than a threshold).

If it is determined that re-learning is not complete (S205: NO), the control unit 301 updates the weights and biases between the nodes used in each layer (step S206) and returns the process to step S201. The control unit 301 can update the weights and biases between each node using the backpropagation method, which sequentially updates the weights and biases between nodes from the output layer LY4 to the input layer LY1.

If it is determined that the re-learning is complete (S205: YES), the control unit 301 stores the re-learned updated learning model LM11 in the memory unit 302 (step S207).

When re-learning is complete, the control unit 301 transmits the re-learned updated learning model LM11 to the garbage collection truck 1 as update data (step S208). The update data transmitted from the management server 300 reaches the garbage collection truck 1 via the business server 200. When the garbage collection truck 1 receives the update data (update learning model LM11) transmitted from the management server 300, it updates the action dictionary by replacing the learning model LM1 stored as the action dictionary with the update learning model LM11.

In this embodiment, the management server 300 is configured to transmit the updated learning model LM11 after re-learning to the garbage collection vehicle 1, but it may also be configured to transmit only the parameters that define the learning model LM11.

As described above, in the second embodiment, update data is generated by re-learning the update learning model LM11, and the operation dictionary of the sewage collection vehicle 1 is updated by the generated update data, so that appropriate operation control can be expected according to the usage situation of the sewage collection vehicle 1.

In the second embodiment, the procedure for updating the action dictionary (learning model LM1) has been described, but the recognition dictionary (learning model LM2) can also be updated in exactly the same manner. That is, in the management server 300, an update learning model for generating update data for the recognition dictionary is prepared, and this update learning model is re-learned in the same manner as above, thereby generating update data (re-learned update learning model). The management server 300 transmits the re-learned update learning model to the garbage collection vehicle 1 as update data for the recognition dictionary. When the garbage collection vehicle 1 receives update data (update learning model) transmitted from the management server 300, it updates the recognition dictionary by replacing the learning model LM2 stored as the recognition dictionary with the re-learned update learning model. The same applies to the case where the recognition dictionary includes a learning model for recognizing the amount and type of garbage inside the garbage input box 4 and the amount of garbage backflow.

(Embodiment 3)
In the third embodiment, a configuration will be described in which, when a user changes sanitation vehicle 1, an action dictionary is provided for the new sanitation vehicle.

The management server 300 can generate update data for each customer and each vehicle model. The management server 300 stores the generated update data for each customer and each vehicle model in the storage unit 302 in association with the customer's identification information and vehicle model. FIG. 16 is a conceptual diagram showing an example of an update data management table. The update data management table stores update data in association with the customer's identification information and vehicle model. The update data may include update data for the action dictionary and update data for the recognition dictionary. The update data may also be a table or function that specifies the control setting values for each action, or may be an update learning model that has been trained to output the control setting values when work data is input. Note that for new customers, a default update learning model may be prepared, and update data for new customers may be generated using a method such as transfer learning.

Figure 17 is a flowchart explaining the procedure for updating the action dictionary in embodiment 3. The control unit 301 of the management server 300 determines whether the user has changed sewage collection vehicles 1 (step S301). The control unit 301 can determine whether the user has changed sewage collection vehicles 1, for example, by referring to the customer management table TB1 and determining whether the vehicle identification information associated with the same customer identification information has changed. If no change has occurred (S301: NO), the control unit 301 ends the processing according to this flowchart without performing the following processing.

If it is determined that a vehicle change has occurred (S301: YES), the control unit 301 searches the update data management table using the user's customer identification information and the vehicle model identified based on the vehicle identification information after the change as search keys, and identifies the update data to be sent (step S302).

The control unit 301 transmits the identified update data to the destination refuse collection vehicle (step S303). The update data transmitted from the management server 300 reaches the destination refuse collection vehicle via the business server 200. When the refuse collection vehicle receives the update data transmitted from the management server 300, it updates the operation dictionary by replacing the control setting values or learning model stored as the operation dictionary with the update data.

As described above, when a user switches between sewage collection vehicles 1, the action dictionary stored in the new sewage collection vehicle can be updated with update data generated based on the work data obtained from the original sewage collection vehicle 1, thereby reducing the sense of discomfort that accompanies switching between sewage collection vehicles 1.

The management server 300 in the third embodiment is configured to refer to the customer management table TB1 to determine whether or not to change the sewage collection vehicle 1, but may also refer to the work history table TB2 to recommend changing the sewage collection vehicle 1 or to suggest to the business establishment to increase or decrease the number of vehicles. FIG. 18 is a flowchart explaining the procedure of the proposal process. The control unit 301 of the management server 300 refers to the work history table TB2 to calculate the average load weight per day, for example, on a monthly basis (step S321). The control unit 301 identifies the type of sewage collection vehicle to be recommended based on the calculated average load weight (step S322). For example, if the calculated average load weight is close to the maximum load weight (rated) of the sewage collection vehicle currently being used, a vehicle with a larger maximum load weight is identified as the vehicle to be recommended. Conversely, if the calculated average load weight is sufficiently smaller than the maximum load weight (rated) of the sewage collection vehicle currently being used and a vehicle with a smaller maximum load weight can be used, the vehicle is identified as the vehicle to be recommended.

If a change of vehicle is recommended, the control unit 301 transmits information about the recommended refuse collection vehicle to the business server 200 (step S323). Note that if the vehicle type identified in step S322 is the same as the vehicle type of the refuse collection vehicle currently being used, the processing of step S323 can be omitted.

Next, the control unit 301 refers to the work history table TB2 to calculate the monthly operating time (step S324). The control unit 301 determines whether the calculated operating time exceeds a first threshold (step S325), and if it determines that it exceeds the first threshold (S325: YES), it transmits information to the business server 200 recommending an increase in the number of garbage collection trucks (step S326).

On the other hand, if the calculated operating time is equal to or less than the first threshold (S325: NO), the control unit 301 determines whether the calculated operating time is less than the second threshold (step S327). Here, the second threshold is less than the first threshold. If it is determined that the calculated operating time is less than the second threshold (S327: YES), the control unit 301 transmits information to the business server 200 recommending a reduction in the number of garbage collection trucks (step S328).

If it is determined in step S327 that the calculated operating time is equal to or greater than the second threshold value (S327: NO), the operating time can be estimated to be appropriate, and the control unit 301 ends the processing according to this flowchart without making any proposal to increase or decrease the number of garbage collection vehicles.

With the above configuration, information on recommended garbage collection vehicles can be sent to each business, and suggestions can be made on switching garbage collection vehicles or increasing or decreasing the number of vehicles.

(Embodiment 4)
In embodiment 4, a configuration is described in which, when a garbage collection vehicle 1 used by one user (user A) is lent to another user (user B), the action dictionary is updated for the new user (user B).

Figure 19 is a flowchart explaining the procedure for updating the action dictionary in embodiment 4. The control unit 301 of the management server 300 determines whether the rental destination of the sewage collection vehicle 1 has changed (step S401). For example, the control unit 301 refers to the customer management table TB1, and can determine that the rental destination has changed if the customer identification information associated with the same vehicle identification information has switched from the identification information of user A to the identification information of user B. If the rental destination has not changed (S401: NO), the control unit 301 ends the processing according to this flowchart without performing the following processing.

If it is determined that a change in the rental destination has occurred (S401: YES), the update data management table is searched using the customer identification information of user B and the vehicle type identified based on the vehicle identification information of the sewage collection vehicle 1 that was rented to user A as search keys, and the update data to be transmitted is generated (step S402).

The control unit 301 transmits the generated update data to the garbage collection vehicle 1 that user B plans to use (step S403). The update data transmitted from the management server 300 reaches the garbage collection vehicle 1 via the business server 200. When the garbage collection vehicle 1 receives the update data transmitted from the management server 300, it updates the operation dictionary by replacing the control setting values or learning model stored as the operation dictionary with the update data.

As described above, even if the rental destination of the refuse collection vehicle 1 is changed, the action dictionary stored in the refuse collection vehicle 1 can be updated with update data generated for the new user, thereby reducing the sense of discomfort that comes with changing the refuse collection vehicle.

(Embodiment 5)
In the fifth embodiment, a configuration will be described in which the action dictionary to be used is changed depending on the type of dust.

Figure 20 is a schematic diagram showing an example of an action dictionary provided in the sewage collection truck 1 in embodiment 5. The example in Figure 20 shows that the sewage collection truck 1 has a learning model LM1a for food waste, a learning model LM1b for PET bottles, a learning model LM1c for wood, and a learning model LM1d for cardboard as its action dictionary.

The operation setting unit 120 selects one of the learning models LM1a to LMd according to the type of dust selected by the dust type selection operation unit 122. The operation setting unit 120 can obtain appropriate control setting values according to the type of dust by inputting work data into the selected learning model LM1a (LM1b to LM1d).

As described above, in the fifth embodiment, the operation dictionary can be changed according to the type of garbage selected by the worker, so that the operation of the work device 3 (loading device 10) can be controlled in a way that is appropriate for the type of garbage.

In this embodiment, an example of application to a refuse collection vehicle 1 has been described, but the present invention is not limited to refuse collection vehicles and can be applied to various work vehicles equipped with a work section for loading or unloading cargo from a storage section. For example, the present invention can be applied to any work vehicle (specially equipped vehicle) such as a dump truck, a detachable vehicle, a refrigerated/freezer vehicle, a liquid transport vehicle, a powder transport vehicle, a water tanker, a sprinkler vehicle, a mixer vehicle, a tank truck, a suction vehicle, a high-pressure washer vehicle, a vehicle transport vehicle, a road maintenance work vehicle, an airport work vehicle, or a transport vehicle with a transport device in the luggage compartment.

In addition, in this embodiment, a refuse collection truck 1 is used as an example of a work vehicle, and the action dictionary provided on the refuse collection truck 1 is applied only when performing loading work, but the present invention is not limited to this, and the action dictionary may be applied to unloading work using an unloading device.

The embodiments disclosed herein are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is indicated by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims.

1. Refuse collection vehicle (work vehicle)
2 Chassis 3 Working device 4 Garbage drop box 5 Garbage storage box 10 Loading device 11 Deflector 12 Sliding plate 12C Lifting cylinder 13 Compression plate 13C Rotating cylinder 20 Discharge device 21 Discharge plate 21C Discharge cylinder 120 Operation setting unit 140 Operation control unit 160 Image processing unit 200 Business establishment server 201 Control unit 202 Memory unit 203 First communication unit 204 Second communication unit 205 Operation unit 206 Display unit 300 Management server 301 Control unit 302 Memory unit 303 Communication unit 304 Operation unit 305 Display unit

Claims (20)

acquiring work data relating to the loading or unloading work from a work vehicle including a storage unit, a working unit that performs loading or unloading work of a cargo into or from the storage unit, a memory unit that stores an action dictionary that is referenced when controlling the operation of the working unit, and a control unit that controls the operation of the working unit by referring to the action dictionary;
generating update data for updating the action dictionary based on the acquired task data;
An information processing method comprising: a process of outputting generated update data to the work vehicle in order to update an action dictionary provided in the work vehicle, the process being performed by a computer. The work vehicle is a vehicle loaned to a user,
The information processing method according to claim 1 , wherein, when the vehicle loaned to the user is switched from one work vehicle to another work vehicle, the update data generated for the other work vehicle is output to the other work vehicle. The work vehicle is a vehicle loaned to a user,
The information processing method according to claim 1 or 2, wherein, when a user to whom a work vehicle is rented is switched from one user to another user, update data generated for the other user is output to the one work vehicle. Acquire the identification information of the user to whom the device is to be loaned,
The information processing method according to claim 2 or 3, further comprising storing identification information of the loaned work vehicle and identification information of the user to whom the work vehicle is loaned in association with each other. Identifying a work vehicle to be loaned to a user from among a plurality of types of work vehicles based on the acquired work data;
The information processing method according to any one of claims 1 to 4, further comprising the step of outputting information about the identified work vehicle. The information processing method according to claim 1 , further comprising: outputting proposal information for proposing an increase or decrease in the number of work vehicles or a change in load capacity based on the acquired work data. The information processing method according to claim 1 , wherein the action dictionary includes control setting values for the working unit. the computer includes an update learning model that is trained to output a control setting value for the working unit when work data related to the loading work or the unloading work is input;
The information processing method according to claim 7 , further comprising the step of outputting a control setting value output from the learning model for update as update data to the work vehicle. 9. The information processing method according to claim 1, wherein the action dictionary is constituted by a learning model that is trained to output a control setting value for the working unit when the work data is input. the computer includes an update learning model that is trained to output a control setting value for the working unit when work data related to the loading work or the unloading work is input;
Acquire work data from the work vehicle and correct answer data that should be output by the learning model for update when the work data is input;
Re-learning the updated learning model using a data set including the acquired task data and correct answer data as training data;
The information processing method according to claim 9 , further comprising the step of outputting the learning model obtained by relearning as update data to the work vehicle. The working unit includes a discharge device including a discharge cylinder and a discharge plate that moves in the front-rear direction of the vehicle by the extension and contraction of the discharge cylinder,
The operation dictionary defines a set pressure for restricting the movement of the discharge plate and an opening time of a control valve for controlling the extension and retraction of the discharge cylinder,
The information processing method according to claim 1 , further comprising generating the update data in order to update at least one of the movement restriction set pressure and the release time. The working unit includes a loading device including a sliding cylinder, a sliding plate that slides in the vertical direction of the vehicle by the extension and contraction of the sliding cylinder, a swinging cylinder, and a compression plate that swings in the front-rear direction of the vehicle by the extension and contraction of the swinging cylinder,
The action dictionary defines a control pattern of the sliding plate and the compression plate,
The information processing method according to claim 1 , further comprising generating the update data in order to update the control pattern. The working unit includes a loading device including a sliding cylinder, a sliding plate that slides in the vertical direction of the vehicle by the extension and contraction of the sliding cylinder, a swinging cylinder, and a compression plate that swings in the front-rear direction of the vehicle by the extension and contraction of the swinging cylinder,
The action dictionary defines a timing for reversing the compression plate when the sliding plate is lowered from a predetermined stop position,
The information processing method according to claim 1 , further comprising generating the update data in order to update the timing. the work vehicle further includes an imaging unit that images the inside and outside of the working unit, and a recognition dictionary that is referenced when recognizing the inside and outside of the working unit based on image data obtained by the imaging unit,
The computer further includes an update learning model that is trained to output information regarding the inside and outside of the work unit when image data obtained by the imaging unit is input,
Acquiring image data obtained from the imaging unit of the work vehicle and correct answer data that should be output by the learning model for update when the image data is input;
Re-learning the updated learning model using a data set including the acquired task data and correct answer data as training data;
The information processing method according to any one of claims 1 to 13, further comprising the step of outputting a learning model obtained by relearning to the work vehicle as update data for the recognition dictionary. The information processing method according to any one of claims 1 to 14, wherein the work data includes at least one of the work location, the type of garbage to be loaded, images from an imaging unit that images the inside and outside of the work section, the hydraulic value of a loading device that loads garbage into the storage section, the hydraulic value of an ejection device when loading garbage into the storage section, the position of an ejection plate provided on the ejection device, the weight of the loaded garbage, and the compression sound of the garbage during loading work. The information processing method according to any one of claims 1 to 15, wherein the work vehicle is provided with an action dictionary that is selected according to the type of cargo, and when a selection for the type of cargo is accepted, the action dictionary corresponding to the selected type is referenced to control the operation of the working unit. An information processing system including a work vehicle that transports a load and an information processing device that is communicatively connected to the work vehicle,
The work vehicle is
A storage section for storing cargo;
a working section for loading or unloading cargo into or from the storage section;
a storage unit that stores an action dictionary to be referenced when controlling the action of the working unit;
a control unit that controls an operation of the working unit by referring to the operation dictionary,
The information processing device includes:
an acquisition unit that acquires work data related to the loading work or the unloading work from the work vehicle;
a generation unit that generates update data for updating the action dictionary based on the acquired task data;
an output unit that outputs generated update data to the work vehicle in order to update the action dictionary provided in the work vehicle;
The work vehicle is
an update unit that updates the action dictionary stored in the storage unit based on update data received from the information processing device. The work vehicle is
A communication unit that communicates with a terminal device carried by a user of the work vehicle,
Accepting the type of cargo through the communication unit;
The information processing system according to claim 17 , wherein the control unit controls the operation of the working unit by referring to an action dictionary prepared for each type of loaded object. an acquisition unit that acquires operation data relating to the loading operation or the unloading operation from a work vehicle that includes a storage unit, a working unit that performs loading or unloading of a cargo into or from the storage unit, a memory unit that stores an action dictionary that is referenced when controlling the operation of the working unit, and a control unit that controls the operation of the working unit by referring to the action dictionary;
a generation unit that generates update data for updating the action dictionary based on the acquired task data;
an output unit that outputs generated update data to the work vehicle in order to update an action dictionary provided in the work vehicle. A storage section for storing cargo;
a working section for loading or unloading cargo into or from the storage section;
a storage unit that stores an action dictionary to be referenced when controlling the action of the working unit;
A control unit that controls the operation of the working unit by referring to the operation dictionary;
an output unit that outputs operation data related to the loading operation or the unloading operation to an external device;
an acquisition unit that acquires update data for the action dictionary from the external device that is configured to generate update data for the action dictionary based on the work data;
and an update unit that updates the action dictionary stored in the storage unit based on the acquired update data.

Images