Control devices for industrial vehicles, control systems for industrial vehicles, and programs for industrial vehicle control devices.

Abstract

To provide an industrial vehicle control device, an industrial vehicle control system and a program for the industrial vehicle control device capable of accurately acquiring an estimation index for estimating a work state of an industrial vehicle with a few man-hours.SOLUTION: A work state estimation section 30 performs machine learning with input of operation information related to an operation state of a forklift and input of correct answer data of a work state. The work state estimation section 30 is capable of performing the machine learning on the basis of the input of the operation information on actual work conducted in the past and the correct answer data showing a work state as a result of the operation information. Thus, a control device is capable of acquiring an estimation index for estimating the work state of the forklift through a process to perform the machine learning using the operation information in the past and the correct answer data. Since the estimation index is based on actual data in the past, the control device is capable of accurately estimating the work state.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a control device for an industrial vehicle, a control system for an industrial vehicle, and a program for the control device for an industrial vehicle. [Background technology]

[0002] A known example of a control device for an industrial vehicle is described in Patent Document 1. The control device for an industrial vehicle described in Patent Document 1 estimates the working state of the industrial vehicle based on operation information for the industrial vehicle, and controls the industrial vehicle based on the estimated working state. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2019-189435 A Summary of the Invention [Problem to be solved by the invention]

[0004] In order for the control device to estimate the work state, a work state estimation model is constructed that sets estimation indices such as transition conditions and transition thresholds for each work state. However, designing estimation indices for estimating the work state is difficult, making it hard to improve accuracy, and requiring a huge number of design man-hours.

[0005] An object of the present invention is to provide an industrial vehicle control device, an industrial vehicle control system, and an industrial vehicle control device program that can accurately obtain estimated indicators for estimating the working status of an industrial vehicle with minimal labor. [Means for solving the problem]

[0006] A control device for an industrial vehicle according to one aspect of the present invention is a control device for estimating the working state of an industrial vehicle, and has a working state estimation unit that estimates the working state of the industrial vehicle using machine learning, and the working state estimation unit performs machine learning by inputting operation information regarding the operating state of the industrial vehicle and inputting correct answer data for the working state.

[0007] The control device for an industrial vehicle has a work state estimation unit that estimates the work state of the industrial vehicle using machine learning. Therefore, the control device can design an estimation index for estimating the work state of the industrial vehicle through machine learning. Here, the work state estimation unit performs machine learning by inputting operation information regarding the operation state of the industrial vehicle and correct answer data for the work state. In this way, the work state estimation unit can perform machine learning based on input of operation information actually performed in the past and correct answer data indicating the work state resulting from the operation information. This allows the control device to acquire an estimation index for estimating the work state of the industrial vehicle simply by performing machine learning using past operation information and correct answer data. Furthermore, because the estimation index is based on actual past data, it is possible to accurately estimate the work state. As described above, an estimation index for estimating the work state of an industrial vehicle can be acquired with little effort.

[0008] The work state estimation unit may receive current operation information and output the current work state based on the learning result, thereby enabling the work state estimation unit to easily estimate the current work state simply by inputting the current operation information.

[0009] The control device for the industrial vehicle may further include a luggage state estimating unit that outputs a luggage state related to the state of luggage handled by the industrial vehicle based on the work state output by the work state estimating unit, and a work state correcting unit that corrects the work state based on the work state output by the work state estimating unit and the luggage state output by the luggage state estimating unit. In this case, the control device can improve the accuracy of estimating the work state by making corrections taking into account the luggage state of the industrial vehicle.

[0010] The operation information may include at least one of an accelerator operation amount, a tire angle, a lift operation amount, a reach operation amount, and a tilt operation amount. These parameters reflect the intention of the worker in the industrial vehicle. The work state estimation unit can perform appropriate machine learning by using such parameters as the operation information.

[0011] The task state estimation unit may use LSTM as a machine learning model and perform machine learning using operation information acquired for each time series as learning data. When performing machine learning using operation information acquired for each time series, LSTM can perform learning that takes into account not only the most recent data but also data going back in time. This allows the task state estimation unit to obtain more accurate estimation indicators.

[0012] An industrial vehicle control system according to one aspect of the present invention includes the above-described industrial vehicle control device.

[0013] A program for a control device for an industrial vehicle according to one aspect of the present invention is used in the control device for the industrial vehicle described above.

[0014] These industrial vehicle control systems and industrial vehicle control device programs can achieve the same effects and aims as the above-mentioned industrial vehicle control device. [Effects of the Invention]

[0015] According to the present invention, it is possible to provide an industrial vehicle control device, an industrial vehicle control system, and an industrial vehicle control device program that can accurately obtain a work state estimation model for estimating the work state of an industrial vehicle with little labor. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a block diagram showing a driving assistance system according to an embodiment of the present invention; [Figure 2]FIG. 1(a) is a perspective view showing how the photographing unit is attached to the forklift, and FIG. 1(b) is a schematic diagram for explaining the angle of the photographing unit. [Figure 3] FIG. 1 is a diagram illustrating a working state of a forklift. [Figure 4] FIG. 4 is a block diagram of a work status acquisition unit. [Figure 5] FIG. 10 is a diagram illustrating a state during machine learning of the task state estimation unit. [Figure 6] 3 is a flowchart showing the processing contents of the driving assistance system. [Figure 7] 10 is a graph illustrating the effects of the baggage state estimation unit and the work state correction unit. [Figure 8] 10 is a graph showing experimental results of the control device. DETAILED DESCRIPTION OF THE INVENTION

[0017] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings. In the drawings, identical or equivalent elements are designated by the same reference numerals, and redundant description will be omitted.

[0018] Fig. 1 is a block diagram showing a driving assistance system 100 (a control system for an industrial vehicle) including a control device 20 according to an embodiment of the present invention. The driving assistance system 100 is a system for remotely controlling an industrial vehicle. As shown in Fig. 1, the driving assistance system 100 includes a forklift 1 (industrial vehicle) and a remote control device 2.

[0019] The forklift 1 includes a driving control unit 11 and multiple image capturing units 12. The driving control unit 11 receives command signals from the remote control device 2 and performs driving and steering control based on the command signals. The multiple image capturing units 12 are provided at various locations on the forklift 1 and capture images of the surrounding environment of the forklift 1. The image capturing units 12 acquire the captured images as support information used for work support and transmit them to the display control unit 22 (described later). Figure 2(a) shows an example of the installation locations of the multiple image capturing units 12. The image capturing units 12 are installed at the front end, widthwise ends, ceiling, etc. of the forklift 1. When an XYZ coordinate system is set as shown in Figure 2(b), the image capturing units 12 at each location may be installed tilted around the X-axis, Y-axis, and Z-axis. In Figure 2(a), eight image capturing units 12 are provided, each capturing an image of a different location.

[0020] As shown in Fig. 1, the remote control device 2 includes a memory unit 15, an operation unit 16, a display unit 17, an operation information acquisition unit 18, and a control device 20. The memory unit 15 is a device for storing various types of information. The operation unit 16 is a user interface through which an operator inputs operations to remotely operate the forklift 1. The display unit 17 is a user interface for outputting images. The display unit 17 has a first area D1, a second area D2, and a third area D3 that can output different images from each other.

[0021] The operation information acquisition unit 18 acquires operation information when the operator is operating an operation target (here, the operation unit 16). The operation information acquisition unit 18 is configured by, for example, a sensor provided on an operation lever of the operation unit 16, or a means for detecting the operation content based on a signal indicating the operation content of the operation unit 16.

[0022] The control device 20 is a control unit that controls the entire remote operation device 2. The control device 20 includes an ECU (Electronic Control Unit) that manages the remote operation device 2. The ECU is an electronic control unit that includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a CAN (Controller Area Network), a communication circuit, etc. The ECU, for example, loads a program stored in the ROM into the RAM and executes the program loaded into the RAM with the CPU, thereby realizing various functions. The control device 20 includes a driving command unit 21, a display control unit 22, a work state acquisition unit 23, and a viewpoint information acquisition unit 26.

[0023] The operation command unit 21 is a unit that generates a command signal based on an operation input through the operation unit 16 and transmits it to the operation control unit 11.

[0024] The display control unit 22 is a unit that controls the display content of the display unit 17. The display control unit 22 displays information for supporting the operator in remote operation work on the display unit 17. The display control unit 22 controls the display content of the first area D1, the second area D2, and the third area D3 of the display unit 17 using information in the database of the storage unit 15.

[0025] The display control unit 22 selects an image to be displayed on the display unit 17 based on the viewpoint information acquired by the viewpoint information acquisition unit 26. The first area D1 and the second area D2 are large screen areas that display a specific image in a large size, and display an image of an image capturing unit 12 selected by the display control unit 22 from among images captured by the multiple image capturing units 12. The third area D3 corresponds to a small screen area that displays multiple images in a small size, and displays images from all of the image capturing units 12 on the forklift 1 as environmental information for checking the surroundings.

[0026] For example, FIG. 3(a) is a table that defines the working states of the forklift 8. Here, examples of working states include seven states related to the driving of the forklift 3 and the work of picking up / unloading cargo: "Stop," "Move forward," "Approach to the shelf," "Adjust heading," "Load / Unload," "Retreat," and "Move reverse." These states are divided into a state with a load (Load) and a state without a load (NoLoad), thereby defining a total of 14 working states. The driving of the forklift 3 corresponds to "Stop," "Move forward," "Approach to the shelf," "Retreat," and "Move reverse," while the work of picking up / unloading cargo corresponds to "Adjust heading" and "Load / Unload." The definition and number of each working state are not limited to those described above and can be changed as appropriate. Alternatively, the working state may be defined solely based on the driving of the forklift 3 (industrial vehicle). FIG. 3(b) is a schematic diagram showing each working state when the forklift 1 retrieves and unloads cargo from the shelf SF. In the database of the storage unit 15, each working state is associated with the image capturing unit 12 that can capture areas that require attention during the working state. Therefore, the display control unit 22 can display images of areas that require particular attention on a large screen in the first area D1 and the second area D2 according to the working state of the forklift 1 being operated by the operator. Note that the image displayed on the display unit 17 under the control of the display control unit 22 for driving assistance may be referred to as a "driving assistance image."

[0027] The appropriate timing for switching the driving assistance video varies from operator to operator. In addition, when the state of the work being performed by the forklift 1 changes, the video to be displayed on the display unit 17 changes. In consideration of this situation, the work state acquisition unit 23 and the viewpoint information acquisition unit 26 adapt the driving assistance video to the individual operator and perform processing corresponding to the state transition of the work state.

[0028] The work status acquisition unit 23 acquires the work status being performed by the forklift 1. The work status acquisition unit 23 acquires the work status based on the operation information acquired by the operation information acquisition unit 18. The work status acquisition unit 23 acquires the work status using a work status estimation model in which estimation indicators such as transition conditions and transition thresholds for transitioning to each work status are set. The work status acquisition unit 23 may acquire the work status estimation model stored in the storage unit 15. Details of the work status acquisition unit 23 will be described later. The work status acquisition unit 23 transmits the acquired work status to the viewpoint information acquisition unit 26.

[0029] The viewpoint information acquisition unit 26 acquires the operator's viewpoint based on the work status acquired by the work status acquisition unit 23. Here, the storage unit 15 has a database in which the work status of the forklift 1 and viewpoint information based on the operator's gaze are linked. Therefore, the viewpoint information acquisition unit 26 acquires viewpoint information corresponding to the work status by comparing the work status of the forklift 1 with the database. The viewpoint information acquisition unit 26 transmits the acquired viewpoint information to the display control unit 22. As a result, the display control unit 22 controls the image displayed on the display unit 17 based on the viewpoint information, and can display the optimal image for the work status of the forklift 1 and the individual operator.

[0030] Next, the detailed configuration of the work status acquisition unit 23 will be described with reference to Figure 4. Figure 4 is a block diagram showing the configuration of the work status acquisition unit 23. As shown in Figure 4, the work status acquisition unit 23 comprises a work status estimation unit 30, a cargo status estimation unit 31, and a work status correction unit 32.

[0031] The work state estimation unit 30 estimates the work state of the forklift 1 using machine learning. The work state estimation unit 30 receives current operation information u(t) as input and outputs the current work state x(t) based on the learning results. The work state estimation unit 30 has a machine learning model M1.

[0032] The working states are classified into categories "1" to "14" as shown in FIG. 3(a). The operation information includes at least one of the accelerator operation amount, tire angle, lift operation amount, reach operation amount, and tilt operation amount. The accelerator operation amount is a parameter indicating the accelerator operation amount of the forklift 1, and the larger the accelerator operation amount, the faster the vehicle speed of the forklift 1. The tire angle is a parameter indicating the turning angle of the tires of the forklift 1, and the larger the tire angle, the larger the turning angle of the forklift 1. The lift operation amount is a parameter indicating the operation amount of the operation to lift the forks of the forklift 1, and the larger the lift operation amount, the larger the forks are lifted. The reach operation amount is a parameter indicating the operation amount of the operation to extend the forks of the forklift 1 forward, and the larger the reach operation amount, the larger the forks are moved forward. The tilt operation amount is a parameter indicating the operation amount of the operation to tilt the forks of the forklift 1, and the larger the tilt operation amount, the larger the forks are tilted.

[0033] The work state estimation unit 30 estimates and outputs the current work state x(t) by comparing the input current operation information u(t) with a work state estimation model, which sets estimation indices such as transition conditions and transition thresholds for transitioning to each work state. The work state estimation model determines which parameters of the above-mentioned operation information must increase (decrease) to what extent to transition from one work state to another. For example, as shown in FIG. 3(b), in "Adjust heading" of "Work state 4," the forklift 1 is making a large turn, and in "Load" of "Work state 5," the forklift 1 is moving slowly. Therefore, the work state estimation unit 30 can estimate that the forklift 1 has transitioned from "Work state 4" to "Work state 5" when the accelerator operation amount is large and the tire angle is large, and each parameter falls below a predetermined threshold. This work state estimation model uses machine learning to set estimation indices that are optimal for the operator based on the operator's past driving performance. Note that an individual task state estimation model may be created for each operator, or a task state estimation model not limited to an individual operator may be created. The task state estimation models are stored in the storage unit 15, and the task state estimation unit 30 acquires an appropriate task state estimation model corresponding to an operator from the storage unit 15 at the necessary timing.

[0034] FIG. 5 is a block diagram showing the state of the work state estimating unit 30 during learning. As shown in FIG. 5, the work state estimating unit 30 performs machine learning by inputting operation information related to the operation state of the forklift 1 and correct answer data of the work state. The machine learning process shown in FIG. 5 may be performed in advance, before the forklift 1 is actually remotely operated. Data prepared in advance is used as the operation information and correct answer data. For example, the forklift 1 is actually operated in a laboratory or the like, and data is acquired in a state in which the operation information at a certain time is linked to the actual work state at that time. The actual work state at this time becomes the correct answer data.

[0035] The work state estimation unit 30 may use LSTM (Long Short Term Memory) as the machine learning model M1. In this case, the work state estimation unit 30 may perform machine learning using operation information acquired in time series as training data. The operation information acquired in time series consists of, for example, a data set of "operation information u(t) and correct work state y(t)" acquired at predetermined intervals from "time 0" to "time T". The work state estimation unit 30 may also use RNN (Recurrent Neural Network) as the machine learning model M1, which is capable of handling operation information acquired in time series.

[0036] Returning to Figure 4, the load state estimation unit 31 outputs a load state g(t) relating to the state of the load handled by the forklift 1, based on the work state x(t) output by the work state estimation unit 30. The load state estimation unit 31 has a pre-created load state estimation model (state transition model) M2. The load state estimation unit 31 obtains the load state estimation model M2 from the storage unit 15. The load state estimation unit 31 estimates the load state g(t) by comparing the estimated work state x(t) with the load state estimation model M2. The load state g(t) indicates either "NL: NoLoad (no load)" or "L: Load (load present)". The load state estimation unit 31 outputs the estimated load state g(t) to the work state correction unit 32.

[0037] The luggage state estimation model M2 prohibits luggage state transitions other than the specified working state. For example, in the example of FIG. 3, in "working state 5," the forklift 1 approaches the shelf SF without a luggage, and after receiving the luggage, transitions to "working state 6," a working state of preparation for travel. In this case, the luggage state estimation model M2 allows a transition from "NL: no luggage" to "L: luggage present" when transitioning from "working state 5" to "working state 6," but prohibits any other transitions. Also, in "working state 12," the forklift 1 approaches the shelf SF with a luggage, and after unloading the luggage onto the shelf SF, transitions to "working state 13," a working state of preparation for travel. In this case, the luggage state estimation model M2 allows a transition from "L: luggage present" to "NL: no luggage" when transitioning from "working state 12" to "working state 13," but prohibits any other transitions.

[0038] The work state correction unit 32 corrects the work state based on the work state x(t) output by the work state estimation unit 30 and the baggage state g(t) output by the baggage state estimation unit 31, and outputs a corrected work state s(t). The work state correction unit 32 determines whether the work state x(t) output by the work state estimation unit 30 is a work state with baggage or a work state without baggage. Next, the work state correction unit 32 compares the presence or absence of baggage in the work state x(t) with the baggage state g(t) output by the baggage state estimation unit 31. If the presence or absence of baggage in both cases matches, the work state correction unit 32 sets the corrected work state s(t) to be the same as the work state x(t). On the other hand, if the presence or absence of baggage in both cases is inconsistent, the work state correction unit 32 corrects the work state x(t) to match the baggage state g(t). For example, if the task state estimation unit 30 estimates a transition from a task state (t-1) with luggage to a task state (t) without luggage, while the task state estimation unit 31 estimates no transition between a task state g(t-1) with luggage and a task state g(t), the task state correction unit 32 corrects the task state x(t) without luggage to a task state s(t) with luggage. Note that, within the same operation category in the table of FIG. 3(a), erroneous estimations are likely to occur. For example, within the operation "Adjust heading," it is easy to confuse "task state 4" with "task state 11." Therefore, the task state correction unit 32 corrects the presence or absence of luggage within the same operation category. Therefore, if "task state 4" is incorrect, the task state correction unit 32 corrects it to "task state 1," which is the same operation category as "task state 4." If "Working State 11" is incorrect, the working state correction unit 32 corrects it to "Working State 4," which is in the same operation category as "Working State 11."

[0039] Next, an example of processing details showing a driving assistance method by the control device 20 will be described with reference to FIG. 6. The processing shown in FIG. 6 is performed when a driving assistance video is displayed on the display unit 17 and the operator is remotely operating the forklift 1. As shown in FIG. 6, the operation information acquisition unit 18 acquires operation information by the operator (step S10). Next, the work state estimation unit 30 estimates a work state based on the operation information acquired in step S (step S20). Next, the baggage state estimation unit 31 acquires a baggage state based on the work state estimated in step S20 (step S30). Next, the work state correction unit 32 corrects the work state based on the work state estimated in step S20 and the baggage state estimated in step S30 (step S40). Next, the viewpoint information acquisition unit 50 acquires viewpoint information based on the work state corrected in step S50 (step S50). Next, the display control unit 22 selects a driving assistance video to be displayed on the display unit 17 based on the viewpoint information acquired in step S50 (step S60).

[0040] Next, the operation and effect of the control device 20 according to this embodiment will be described.

[0041] The control device 20 has a work state estimation unit 30 that estimates the work state of the forklift 1 using machine learning. Therefore, the control device 20 can design estimation indicators for estimating the work state of the forklift 1 using machine learning. Here, the work state estimation unit 30 performs machine learning by inputting operation information related to the operation state of the forklift 1 and correct data on the work state. In this way, the work state estimation unit 30 can perform machine learning based on input of operation information that was actually performed in the past and correct data that shows what kind of work state resulted from that operation information. As a result, the control device 20 can obtain estimation indicators for estimating the work state of the forklift 1 by simply performing machine learning using past operation information and correct data. Furthermore, since these estimation indicators are based on actual past data, it is possible to estimate the work state with high accuracy. Thus, estimation indicators for estimating the work state of the forklift 1 can be obtained accurately with little effort.

[0042] The work state estimation unit 30 may receive current operation information and output the current work state based on the learning result. This allows the work state estimation unit 30 to easily estimate the current work state simply by receiving the current operation information.

[0043] The control device 20 may further include a load condition estimation unit 31 that outputs a load condition relating to the condition of the load handled by the forklift 1 based on the work condition output by the work condition estimation unit 30, and a work condition correction unit 32 that corrects the work condition based on the work condition output by the work condition estimation unit 30 and the load condition output by the load condition estimation unit 31. In this case, the control device 20 can improve the accuracy of work condition estimation by making corrections that take into account the load condition of the forklift 1.

[0044] Referring to Figure 7, the effects of the cargo status estimation unit 31 and the work status correction unit 32 will be explained. Figure 7(a) is a graph showing the estimated work status results estimated by the work status estimation unit 30 when operation information at a certain time is shown chronologically on the horizontal axis. Figure 7(b) is a graph showing the estimated results when the work status estimated by the work status estimation unit 30 is corrected by the work status correction unit 32. The dashed lines in Figures 7(a) and 7(b) are graphs that show the presence or absence of cargo. If the dashed line is in the upper part of the graph, it indicates the state of "cargo present," and if it is in the lower part, it indicates the state of "no cargo."

[0045] For example, in Figure 7(a), it is estimated that the state transitioned from "working state 4," where the forks are aligned without a load, to "working state 11," where the forks are aligned with a load, and then to "working state 5," which is the loading / unloading state with a load. Thus, as shown by "FD" in Figure 7(a), the location where the estimation result fluctuates sharply indicates that the working state estimation unit 30 initially estimated an incorrect working state, and then subsequently estimated the correct working state.

[0046] In response to this, the cargo state estimation unit 31 prohibits the transition from "NL: No cargo" to "L: Cargo present" when transitioning from "Work state 4" to "Work state 11". Therefore, the cargo state estimation unit 31 outputs the cargo state "NL: No cargo" to the work state correction unit 32 based on the work state estimation unit 30's estimation result of "Work state 11" which indicates the presence of cargo. As a result, the work state correction unit 32 corrects "Work state 11" which indicates the presence of cargo to "Work state 4" which indicates the absence of cargo, in accordance with the estimation result of the cargo state estimation unit 31. This makes it possible to estimate that the transition to "Work state 5" occurred at the correct timing, i.e., the timing when the transition actually occurred from "Cargo present" to "Cargo absent", without going through "Work state 11" from "Work state 4". As a result, in Figure 7(b), the occurrence of an incorrect work state, such as "FD" in Figure 7(a), can be suppressed.

[0047] The operation information may include at least one of an accelerator operation amount, a tire angle, a lift operation amount, a reach operation amount, and a tilt operation amount. These parameters reflect the operator's intention in the forklift 1. The work state estimation unit 30 can perform appropriate machine learning by using such parameters as the operation information.

[0048] The task state estimation unit 30 may use LSTM as the machine learning model M1 and perform machine learning using operation information acquired for each time series as learning data. When performing machine learning using operation information acquired for each time series, the LSTM can perform learning that takes into account not only the most recent data but also data going back in time. This allows the task state estimation unit 30 to acquire more accurate estimation indices.

[0049] The driving support system 100 (industrial vehicle control system) according to this embodiment has the industrial vehicle control device 1 described above.

[0050] The program for the control device of the industrial vehicle according to this embodiment is used in the industrial vehicle control device 20 described above.

[0051] Specifically, the program for the control device of the industrial vehicle causes the computer system to execute a work state estimation step of estimating the work state of the industrial vehicle using machine learning. In the work state estimation step, machine learning is performed by inputting operation information regarding the operation state of the industrial vehicle and correct answer data of the work state.

[0052] These industrial vehicle control systems and the programs of the industrial vehicle control device 20 can achieve the same functions and effects as the industrial vehicle control device 20 described above.

[0053] Referring to FIG. 8, experimental results for estimating task states using the control device 20 according to the present embodiment will be described. Here, the percentage of match between the actual task state and the estimated result (accuracy rate) was calculated for a specific task pattern. In FIG. 8, the solid line graph G1 indicates the estimated result, and the dashed line graph G2 indicates the actual task state. In the graphs shown in FIG. 8, where graphs G1 and G2 overlap, the estimated result matches the actual task state, while where graphs G1 and G2 do not overlap, the estimated result does not match the actual task state. Multiple operators were prepared, 44 task data sets were acquired, and the average accuracy rate for all 44 tasks was calculated and evaluated. The accuracy rate for the estimation results of the control device 20 according to the present embodiment was improved by approximately 10% compared to the estimation results using a task state model designed without machine learning.

[0054] Although several preferred embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments.

[0055] In the above-described embodiment, the driving assistance system assists the operator in the work performed during remote operation, but the driving assistance system may also assist when an industrial vehicle is operated by a human operator, or when the operator performs a simulation operation of the industrial vehicle.

[0056] The industrial vehicle is not limited to a forklift, but may be a towing tractor, a skid steer loader, or the like.

[0057] [Form 1] A control device for estimating a working state of an industrial vehicle, a work state estimation unit that estimates the work state of the industrial vehicle using machine learning, The work state estimation unit is a control device for an industrial vehicle that performs machine learning by inputting operation information regarding an operation state of the industrial vehicle and inputting correct answer data of the work state.

[0058] [Form 2] 2. The control device for an industrial vehicle according to aspect 1, wherein the work state estimation unit receives the current operation information as input and outputs the current work state based on a learning result.

[0059] [Form 3] A cargo condition estimation unit outputs a cargo condition relating to the condition of the cargo handled by the industrial vehicle, based on the work condition output by the work condition estimation unit. 3. The control device for an industrial vehicle according to claim 1, further comprising: a work state correction unit that corrects the work state based on the work state output by the work state estimation unit and the luggage state output by the luggage state estimation unit.

[0060] [Form 4] 4. The control device for an industrial vehicle according to any one of aspects 1 to 3, wherein the operation information includes at least one of an accelerator operation amount, a tire angle, a lift operation amount, a reach operation amount, and a tilt operation amount.

[0061] [Form 5] The control device for an industrial vehicle according to any one of aspects 1 to 4, wherein the work state estimation unit uses LSTM as a machine learning model and performs machine learning using the operation information acquired for each time series as learning data.

[0062] [Form 6] A control system for an industrial vehicle having a control device for an industrial vehicle as described in any one of the embodiments 1 to 5.

[0063] [Form 7] A program for an industrial vehicle control device used in an industrial vehicle control device according to any one of claims 1 to 5. [Explanation of symbols]

[0064] 1...Forklift, 30...Working condition estimation unit, 31...Load condition estimation unit, 32...Working condition correction unit, 100...Driving support system (industrial vehicle control system).

Claims

1. A control device for estimating a working state of an industrial vehicle, a work state estimation unit that estimates a work state of the industrial vehicle using machine learning; a luggage state estimating unit that outputs a luggage state regarding a state of luggage handled by the industrial vehicle based on the work state output by the work state estimating unit; a work state correction unit that corrects the work state output by the work state estimating unit based on the work state output by the work state estimating unit and the luggage state output by the luggage state estimating unit, the work state estimation unit performs machine learning by inputting operation information relating to an operation state of the industrial vehicle and inputting correct answer data of the work state of the industrial vehicle; The luggage status indicates a status of no luggage or a status of luggage, the luggage state estimation unit uses a luggage state estimation model to prohibit the luggage state from transitioning except for a designated task; The work state correction unit outputs the work state output by the work state estimation unit as a corrected work state when the work state output by the work state estimation unit and the presence or absence of luggage in the luggage state match, and corrects the work state output by the work state estimation unit to match the luggage state when the work state output by the work state estimation unit and the presence or absence of luggage in the luggage state do not match.

2. The control device for an industrial vehicle according to claim 1 , wherein the work state estimation unit receives the current operation information as an input and outputs the current work state based on a learning result.

3. A control device for an industrial vehicle as described in claim 1, wherein the operation information includes at least one of an accelerator operation amount, a tire angle, a lift operation amount, a reach operation amount, and a tilt operation amount.

4. The control device for an industrial vehicle described in Claim 1, wherein the work state estimation unit uses LSTM as a machine learning model and performs machine learning using the operation information acquired for each time series as learning data.

5. An industrial vehicle control system having an industrial vehicle control device described in any one of claims 1 to 4.

6. A program for an industrial vehicle control device used in an industrial vehicle control device described in any one of claims 1 to 4.

Images