Loading weight estimation system of specially-equipped vehicle

Abstract

To provide a loading weight estimation system for specially-equipped vehicles, capable of notifying a user of an estimation result by highly accurately estimating a loading weight.SOLUTION: A loading weight estimation system includes: acquiring, for specially-equipped vehicles equipped with a hydraulic actuator for lifting / lowering cargo boxes, measurement data relating to a magnitude of hydraulic pressure from a pressure gauge that measures a magnitude of hydraulic pressure acting on the hydraulic actuator; specifying a first stop state where a cargo box stops after ascending and a second stop state where the cargo box stops after descending, with their distinguished from each other; estimating a loading weight in the cargo box by selectively using, according to the specified stop state, any one of a first learning model configured to output an arithmetic result of the loading weight when measurement data including the magnitude of hydraulic pressure measured in the first stop state is input and a second learning model configured to output an arithmetic result of the loading weight when measurement data including a magnitude of hydraulic pressure measured in a second stopped state is input; and notifying a user of information of the estimated loading weight.SELECTED DRAWING: Figure 8

Description

[Technical field]

[0001] The present invention relates to a load weight estimation system for a special purpose vehicle. [Background technology]

[0002] 2. Description of the Related Art In recent years, large vehicles such as trucks have become known that are equipped with a weight meter that measures the vehicle's weight (load weight).

[0003] Conventional weight meters use load sensors attached to both ends of the front and rear axles to measure the load on each of the front, rear, left and right tires, and calculate the loaded weight from the sum of the outputs of each load sensor. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2004-132871 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, conventional weight scales use a method of converting the sum of the outputs of each load sensor into the loaded weight, so the correct loaded weight may not be measured when the vehicle is on a slope or when the load on the loading platform is not evenly balanced.

[0006] An object of the present invention is to provide a load weight estimation system for a special purpose vehicle that can accurately estimate the load weight and notify the estimation result. [Means for solving the problem]

[0007] A load weight estimation system for special purpose vehicles according to one embodiment of the present invention includes an acquisition unit that acquires measurement data relating to the amount of displacement from a plurality of displacement sensors that respectively measure the amount of displacement of a plurality of parts of the special purpose vehicle on which the load of a load acts, an estimation unit that estimates the load weight by inputting the measurement data acquired by the acquisition unit into a learning model that is configured to output a calculation result regarding the load weight in response to the input of the measurement data relating to the amount of displacement, and a notification unit that notifies information regarding the load weight estimated by the estimation unit. The displacement amount refers to at least one of "the magnitude of displacement occurring at the part itself" or "the relative magnitude of displacement of the part compared to other parts" when a load is applied to multiple parts of a specially equipped vehicle. The displacement sensor refers not only to sensors directly attached to the multiple parts, but also to sensors built into measuring members (e.g., load cells) attached to the parts.

[0008] It is preferable that the acquisition unit acquires measurement data related to the inclination from an inclinometer that measures the inclination of the vehicle, and the estimation unit estimates the load weight by inputting the measurement data of the displacement and the inclination acquired by the acquisition unit into a learning model configured to output a calculation result regarding the load weight in response to the input of the measurement data related to the displacement and the inclination.

[0009] It is also preferable that the displacement sensors are attached to a plurality of locations of a structure fixed to a chassis frame, spaced apart in the longitudinal or lateral direction of the vehicle.

[0010] It is also preferable that the displacement sensors are attached to a plurality of locations on the axle, spaced apart in the longitudinal or lateral direction of the vehicle.

[0011] A load weight estimation system for a special purpose vehicle according to one embodiment of the present invention includes, for a special purpose vehicle equipped with a hydraulic actuator for raising and lowering a cargo box, an acquisition unit that acquires measurement data related to the magnitude of hydraulic pressure from a pressure gauge that measures the magnitude of the hydraulic pressure acting on the hydraulic actuator, an identification unit that distinguishes between a first stop state in which the cargo box stops after being raised and a second stop state in which the cargo box stops after being lowered, an estimation unit that estimates the load weight in the cargo box by selectively using either a first learning model that is configured to output a calculation result for the load weight when measurement data including the magnitude of hydraulic pressure measured in the first stop state is input, or a second learning model that is configured to output a calculation result for the load weight when measurement data including the magnitude of hydraulic pressure measured in the second stop state is input, depending on the stop state identified by the identification unit, and an alarm unit that alarms information related to the load weight estimated by the estimation unit.

[0012] In addition, it is preferable that, when an instruction to estimate the load weight is given, the device is provided with a control unit that controls the operation of the hydraulic actuator so that the loading box stops after being lifted, and when the loading box stops after being lifted under the control of the control unit, the estimation unit estimates the loading weight of the loading box by inputting the measurement data acquired by the acquisition unit into the first learning model.

[0013] In addition, it is preferable that the input to the first learning model and the second learning model further includes at least one of the inclination angle, distortion, and temperature of the special purpose vehicle.

[0014] It is also preferable to include a calibration unit that calibrates the estimation result by the estimation unit according to each specially equipped vehicle.

[0015] It is also preferable that the vehicle is provided with a judgment unit that judges the loading state of the cargo in accordance with the loading weight estimated by the estimation unit, and the notification unit notifies the loading state in a manner corresponding to the judgment result of the judgment unit.

[0016] It is also preferable that the notification unit includes a display device provided in a visible portion of a packing box in which the cargo is loaded, and displays the loading status on the display device in a display mode corresponding to the determination result.

[0017] It is also preferable that the vehicle further comprises a state detection unit that detects a state of the host vehicle, and the notification unit notifies information relating to the state detected by the state detection unit.

[0018] In addition, it is preferable that the state detected by the state detection unit includes at least one of a state of a packing box on which the load is loaded and a loading state of the load. Effect of the Invention

[0019] According to the present application, the loaded weight can be estimated with high accuracy and the estimation result can be notified. [Brief description of the drawings]

[0020] [Figure 1] FIG. 1 is a side view showing the overall configuration of a specially equipped vehicle according to a first embodiment. [Diagram 2] FIG. 1 is a plan view showing an overall configuration of a specially equipped vehicle according to a first embodiment. [Diagram 3] FIG. 2 is a side view of the packing box in an upright position. [Figure 4] FIG. 1 is a block diagram illustrating a configuration of a loaded weight estimation system. [Diagram 5] FIG. 13 is a conceptual diagram illustrating an example of a measurement value table. [Figure 6] FIG. 2 is a schematic diagram illustrating the configuration of a learning model. [Figure 7] 11 is a flowchart illustrating a procedure for generating a learning model. [Figure 8] 11 is a flowchart illustrating a procedure for estimating a loaded weight using a learning model. [Figure 9] FIG. 13 is a schematic diagram showing an example of a load weight display. [Figure 10] 13 is a graph showing the measurement results. [Figure 11] 13 is a graph showing the estimation results of a learning model. [Figure 12] FIG. 11 is a side view showing the overall configuration of a specially equipped vehicle according to a second embodiment. [Figure 13] FIG. 4 is an explanatory diagram illustrating a mechanism for lifting and lowering a cargo box in a specially equipped vehicle. [Figure 14] 4 is a graph showing the relationship between the pitch angle of the truck chassis and the hydraulic pressure value. [Figure 15] FIG. 11 is a block diagram illustrating the configuration of a loaded weight estimation system in a third embodiment. [Figure 16] FIG. 2 is a schematic diagram illustrating the configuration of a learning model. [Figure 17] 13 is a flowchart illustrating a procedure for estimating a loaded weight in the third embodiment. [Figure 18] FIG. 13 is a block diagram illustrating the configuration of a loaded weight estimation system in a fourth embodiment. [Figure 19] 13 is a flowchart illustrating a procedure for estimating a loaded weight in the fourth embodiment. [Figure 20] FIG. 11 is an explanatory diagram illustrating a calibration method. [Figure 21] 10 is a flowchart illustrating a procedure for notifying a loading state. [Figure 22] FIG. 13 is a schematic diagram showing an example of notification of a loading state. [Figure 23] FIG. 13 is a side view showing the overall configuration of a specially equipped vehicle according to a seventh embodiment. [Figure 24] 23 is a flowchart illustrating a procedure of a process executed by an estimation device in a seventh embodiment. [Diagram 25] 13 is a schematic diagram showing an example of notification of the state of a packing box. FIG. [Figure 26] FIG. 23 is a schematic diagram illustrating the configuration of a learning model in embodiment 7. [Figure 27] 13 is a schematic diagram showing an example of notification of the loading state of a packing box; FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] Hereinafter, the present invention will be described in detail with reference to the drawings showing embodiments thereof. (Embodiment 1) Fig. 1 is a side view showing the overall configuration of a specially equipped vehicle 1 according to the first embodiment, Fig. 2 is a plan view thereof, and Fig. 3 is a side view in a state where a cargo box is raised. The specially equipped vehicle 1 illustrated in Figs. 1 to 3 is a dump truck equipped with a truck chassis 2 which is a traveling section, and a dump device 3 which is an example of a mounting device mounted on the traveling section. 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 from a driver sitting in the driver's seat of the truck chassis 2. Note that Fig. 2 shows a state in which the dump device 3 has been removed for the sake of explanation.

[0022] The truck chassis 2 includes a cab 20 in which a driver's seat is provided, and a chassis frame 21 that supports the cab 20. The chassis frame 21 includes a pair of left and right main frames (vertical joists) 21A, 21A extending in the front-rear direction, and cross members (horizontal joists) 21B, 21B, ..., 21B that connect the pair of left and right main frames 21A, 21A (see FIG. 2). The front wheels 22F and rear wheels 22R of the truck chassis 2 are rotatably attached to the main frames via a suspension device (not shown). The truck chassis 2 includes an engine (prime mover) not shown in the figure, and a transmission connected to the engine via a clutch, and is configured to travel by transmitting the driving force of the engine to the drive system of the drive wheels (for example, the front wheels 22F) via the transmission.

[0023] The dumping apparatus 3 comprises a subframe 30 fixed onto the chassis frame 21, and a cargo box 4 supported by the subframe 30 and in which cargo such as soil and sand is loaded. The cargo box 4 is supported rotatably about a hinge shaft 31 extending in the left-right direction at the rear end of the subframe 30. The cargo box 4 is a box body with an open top, and comprises a front panel 41 arranged to surround a rectangular bottom 40, a pair of left and right side panels 42, and a rear panel (rear flap) 43. The rear panel 43 is configured to be openable and closable.

[0024] The dumping device 3 is equipped with a hoist mechanism 5 for tilting the cargo box 4. The hoist mechanism 5 includes, for example, a lift arm 51, a hydraulic cylinder 52, and a tension link 53. When the hydraulic cylinder 52 is contracted, the cargo box 4 is kept in a horizontal position. On the other hand, when the hydraulic cylinder 52 is extended, the front part of the cargo box 4 is lifted and rotates around the hinge shaft 31. As a result, the cargo box 4 is tilted downward toward the rear as shown in FIG. 3.

[0025] The specially equipped vehicle 1 is equipped with various sensors for detecting the vehicle state. The specially equipped vehicle 1 is equipped with strain sensors 81A to 81D as an example of a displacement sensor for measuring the amount of displacement at a plurality of locations on which the load of the loaded object acts. The strain sensors 81A to 81D are configured, for example, with strain gauges. Here, the strain sensor 81A is attached near the right end of the axle 23F of the front wheel 22F, and the strain sensor 81B is attached near the left end of the axle 23F of the front wheel 22F. The strain sensor 81C is attached near the right end of the axle 23R of the rear wheel 22R, and the strain sensor 81D is attached near the left end of the axle 23R of the rear wheel 22R. The strain sensors 81A, 81B and the strain sensors 81C, 81D are attached symmetrically with respect to the center line of the specially equipped vehicle 1 in the front-rear direction, for example. The strain sensors 81A-81D measure the amount of strain of the axles 23F, 23R in response to the load in a time series manner and output measurement data related to the measured amount of strain. In the following description, when there is no need to explain each of the strain sensors 81A-81D individually, they will also be simply referred to as strain sensor 81 (see FIG. 4).

[0026] The specially equipped vehicle 1 may be equipped with an inclinometer 82 that measures the inclination of the truck chassis 2. The inclinometer 82 is attached to an appropriate location of the chassis frame 21 (for example, near the center in the front-rear and left-right directions). The inclinometer 82 measures the front-rear inclination (pitch) and left-right inclination (roll) of the truck chassis 2 in chronological order, and outputs measurement data related to the measured inclination. The specially equipped vehicle 1 may be equipped with an inclinometer (not shown) that measures the inclination of the cargo box 4, and may acquire measurement data related to the front-rear inclination (pitch) and left-right inclination (roll) of the cargo box 4 in chronological order.

[0027] The specially equipped vehicle 1 may be equipped with a thermometer 83 that measures the temperature (ambient temperature) at and near the mounting position of the strain sensor 81. The ambient temperature measured by the thermometer 83 can be used to calibrate the value of the strain sensor 81, so the thermometer 83 is mounted in a location suitable for measuring the ambient temperature of the strain sensor 81. For example, the thermometer 83 is mounted in at least one location, such as the chassis frame 21, the subframe 30, or the axles 23F, 23R. The thermometer 83 measures the ambient temperature in chronological order, and outputs measurement data related to the measured temperature.

[0028] The specially equipped vehicle 1 may be equipped with a pressure gauge 84 that measures the cylinder pressure of the hydraulic cylinder 52. The pressure gauge 84 measures the cylinder pressure of the hydraulic cylinder 52 in time series, and outputs measurement data related to the measured cylinder pressure.

[0029] The specially equipped vehicle 1 is equipped with an estimation device 100 that estimates the weight of the load (load weight) based on the measurement data including the amount of distortion measured by the distortion sensor 81. In this embodiment, the load weight represents the total weight of the load loaded in the packing box 4, the passengers riding in the specially equipped vehicle 1, the fuel loaded in the specially equipped vehicle 1, and the like, other than the running part and the mounting device constituting the specially equipped vehicle 1. Note that the total weight of the running part and the mounting device when no load is loaded (hereinafter referred to as the vehicle weight) is assumed to be known. The internal configuration of the estimation device 100 and the contents of the processing executed by the estimation device 100 will be described in detail later, but in this embodiment, the load weight of the specially equipped vehicle 1 is estimated using a learning model LM1 that has learned the relationship between the measurement data including the amount of distortion and the load weight of the specially equipped vehicle 1. The estimation device 100 is attached to the chassis frame 21, for example. Alternatively, the estimation device 100 may be provided inside the cab 20.

[0030] In this embodiment, a dump truck equipped with a dump device 3 will be described as an example of a specially equipped vehicle 1, but the specially equipped vehicle 1 is not limited to a dump truck and may be any specially equipped vehicle whose load weight can change depending on the cargo, such as a dry van, a refrigerated vehicle, a liquid transport vehicle, a powder transport vehicle, a water truck, a sprinkler truck, or a garbage collection vehicle.

[0031] The configuration of the loaded weight estimation system according to this embodiment will be described below. 4 is a block diagram illustrating the configuration of the loaded weight estimation system. The loaded weight display system includes an estimation device 100 that acquires measurement data including the amount of distortion measured by a distortion sensor 81 and estimates the loaded weight of the specially equipped vehicle 1 based on the acquired measurement data, and a display device 120 that notifies information related to the loaded weight estimated by the estimation device 100.

[0032] The estimation device 100 is a dedicated or general-purpose computer, and includes a control unit 101, a storage unit 102, an operation unit 103, an input unit 104, an output unit 105, and a communication unit 106.

[0033] The control unit 101 includes, for example, a central processing unit (CPU), a read only memory (ROM), and a random access memory (RAM). The ROM included in the control unit 101 stores a control program for controlling the operation of each hardware unit included in the estimation device 100. The CPU in the control unit 101 executes the control program stored in the ROM and various computer programs stored in a storage unit 102 described below, and controls the operation of each hardware unit, thereby realizing the function of the estimation device 100 in this embodiment. The RAM included in the control unit 101 temporarily stores data used during the execution of calculations, etc.

[0034] The control unit 101 is configured to include a CPU, a ROM, and a RAM, but may alternatively be one or more arithmetic circuits or control circuits including a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), a DSP (Digital Signal Processor), a quantum processor, a volatile or non-volatile memory, etc. The control unit 101 may also include functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when a measurement start instruction is given until when a measurement end instruction is given, a counter that counts numbers, etc.

[0035] The storage unit 102 includes a storage device using a hard disk, a flash memory, etc. The storage unit 102 stores a computer program executed by the control unit 101, various data acquired from the outside, various data generated inside the estimation device 100, etc.

[0036] The computer programs stored in the memory unit 102 include a learning program PG1 for generating a learning model LM1, and an estimation program PG2 for estimating the load weight of the specially equipped vehicle 1 using the learning model LM1.

[0037] These computer programs may be provided by a non-transitory recording medium M on which the computer programs are readably recorded. The recording medium M is, for example, a portable memory such as a CD-ROM, a USB memory, or a Secure Digital (SD) card. The control unit 101 reads the various programs from the recording medium M using a reading device (not shown) and stores the various programs thus read in the storage unit 102.

[0038] The storage unit 102 may include a measurement value table TB1 that stores measurement data obtained from the strain sensor 81, the inclinometer 82, the thermometer 83, and the pressure gauge 84 in a chronological order. FIG. 5 is a conceptual diagram showing an example of the measurement value table TB1. The measurement value table TB1 is a table that stores the measurement data of the amount of strain measured by the strain sensor 81, the inclination angle (pitch and roll) of the specially equipped vehicle 1 measured by the inclinometer 82, the environmental temperature measured by the thermometer 83, and the cylinder pressure of the hydraulic cylinder 52 measured by the pressure gauge 84 in association with time (measurement time). The measurement value table TB1 is prepared for each load weight, such as a table that stores the measurement values ​​measured with the load weight fixed at 1 ton, a table that stores the measurement values ​​measured with the load weight fixed at 2 tons, and a table that stores the measurement values ​​measured with the load weight fixed at 3 tons. The values ​​stored in the measurement value table TB1 may be the output values ​​of the sensors, or may be physical quantities converted from the output values ​​of the sensors.

[0039] The memory unit 102 may include a learning model LM1 for estimating the load weight of the special purpose vehicle 1 from measurement data including the amount of distortion. The learning model LM1 is configured to output a calculation result related to the load weight when measurement data including the amount of distortion is input. The learning model LM1 is defined by its definition information. The definition information of the learning model LM1 includes, for example, information that specifies the structure of the learning model LM1 (type and number of layers, number of nodes, etc.) and parameters such as connection weights determined by learning. The learning model LM1 will be described in detail later.

[0040] The operation unit 103 is composed of switches, buttons, etc., and accepts various operations. The control unit 101 executes appropriate processing based on the operation accepted through the operation unit 103. Note that, in this embodiment, the estimation device 100 is configured to include the operation unit 103, but the operation unit 103 is not essential, and the operation may be accepted via an externally connected device or the communication unit 106.

[0041] The input unit 104 includes an interface for connecting various sensors, and is connected with sensors such as the strain sensor 81, the inclinometer 82, the thermometer 83, and the pressure gauge 84. These sensors may be connected to the input unit 104 by wire or wirelessly. Measurement data relating to the amount of strain output from the strain sensor 81, measurement data relating to the inclination of the special purpose vehicle 1 output from the inclinometer 82, measurement data relating to the temperature output from the thermometer 83, measurement data relating to the cylinder pressure of the hydraulic cylinder 52 output from the pressure gauge 84, and the like are appropriately input to the input unit 104.

[0042] The output unit 105 includes an output interface for connecting a display device 120 such as a liquid crystal monitor. The display device 120 is provided, for example, near the driver's seat of the cab 20. Alternatively, the display device 120 may be provided on the rear side of the front panel 41. The output interface provided in the output unit 105 may be an output interface that outputs an analog video signal, or may be an output interface that outputs a digital video signal such as DVI (Digital Visual Interface) or HDMI (High-Definition Multimedia Interface, registered trademark). The output unit 105 outputs display data to the display device 120 so that the display device 120 displays, for example, an estimation result of the load weight using the learning model LM1.

[0043] In the present embodiment, the display device 120 is connected to the outside of the estimation device 100. However, the estimation device 100 may include the display device 120 mounted thereon.

[0044] The communication unit 106 includes a communication interface for transmitting and receiving various data to and from an external device. An example of a counterpart with which the estimation device 100 communicates via the communication unit 106 is various ECUs (Electronic Controller Units) and PLCs (Programmable Logic Controllers) mounted on the specially equipped vehicle 1. In this case, the communication unit 106 may include a communication port conforming to, for example, RS-485 in order to communicate with various ECUs and PLCs mounted on the specially equipped vehicle 1, or may include a communication interface conforming to a communication standard for in-vehicle communication such as CAN (Controller Area Network). Other examples of counterparts with which the estimation device 100 communicates via the communication unit 106 include a server device installed outside the specially equipped vehicle 1 and a mobile terminal carried by a user. In this case, the communication unit 106 may include a communication interface conforming to a wireless communication standard such as WiFi (registered trademark), 3G, 4G, 5G, or LTE (Long Term Evolution) in order to communicate with an external server device or the like.

[0045] The learning model LM1 will be described below. 6 is a schematic diagram for explaining the configuration of the learning model LM1. The learning model LM1 in this embodiment is, for example, a support vector regression model, and includes an input layer to which various measurement data are input, an intermediate layer including a kernel that performs a predetermined calculation based on the measurement data input to the input layer, and an output layer that combines outputs from the intermediate layer and outputs the calculation results.

[0046] The input layer, hidden layer, and output layer of the learning model LM1 each contain 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 connection weights. In a support vector machine that is nonlinearly extended using the kernel trick, the connection weights from the hidden layer to the output layer are adaptively determined by learning. On the other hand, the connection weights from the input layer to the hidden layer are fixed and are mechanically calculated from the training data.

[0047] Measurement data including the amount of distortion is input to the input layer of the learning model LM1. For example, as shown in Fig. 6, measurement data related to the amount of distortion of the right front axle (amount of distortion measured by the distortion sensor 81A), the amount of distortion of the left front axle (amount of distortion measured by the distortion sensor 81B), the amount of distortion of the right rear axle (amount of distortion measured by the distortion sensor 81C), the amount of distortion of the left rear axle (amount of distortion measured by the distortion sensor 81D), the inclination including roll and pitch (inclination measured by the inclinometer 82), and the environmental temperature (temperature measured by the thermometer 83) are input to the input layer of the learning model LM1.

[0048] The measurement data input to the input layer is weighted by a connection weight determined using training data and output to the intermediate layer. The intermediate layer executes calculations using kernels based on the data input from the input layer. The data calculated in each kernel of the intermediate layer is weighted by a connection weight determined by learning and output to the output layer. The output layer outputs a calculation result related to the load weight by combining the data input from the intermediate layer. Here, the calculation result output by the output layer may be an estimated value of the load weight, or may be a probability corresponding to a certain load weight. In the latter case, the output layer is composed of multiple nodes, and the first node outputs the probability that the load weight is 1 ton, the second node outputs the probability that the load weight is 2 tons, ..., and the Nth node (N is an integer of 2 or more) outputs the probability that the load weight is N tons.

[0049] The learning model LM1 shown in FIG. 6 is configured to output a calculation result related to the load weight in response to the input of measurement data related to the amount of distortion measured by the distortion sensor 81, the inclination measured by the inclinometer 82, and the environmental temperature measured by the thermometer 83, but the input / output relationship in the learning model LM1 is not limited to the above and can be set appropriately. For example, the learning model LM1 may be configured to output a calculation result related to the load weight in response to the input of measurement data further including the cylinder pressure of the hydraulic cylinder 52 measured by the pressure gauge 84. The learning model LM1 may also be configured to receive measurement data related to the amount of distortion measured by the distortion sensor 81 and output a calculation result related to the load weight. Furthermore, two or three selected pieces of measurement data from the measurement data of the distortion sensors 81A to 81D may be input to the learning model LM1 and output a calculation result related to the load weight. The learning model LM1 may also be configured to receive measurement data of the amount of distortion and either one of the measurement data of the inclination or the environmental temperature and output a calculation result related to the load weight.

[0050] Furthermore, the control unit 101 may perform pre-processing to correct the amount of distortion measured by the distortion sensor 81 using the environmental temperature measured by the thermometer 83, and input data including the corrected amount of distortion to the learning model LM1.

[0051] In a learning phase prior to starting operation, the estimation device 100 collects measurement data including the amount of distortion, and learns using the collected measurement data as training data to generate the learning model LM1 as described above.

[0052] FIG. 7 is a flowchart for explaining the procedure for generating the learning model LM1. Prior to learning, the control unit 101 of the estimation device 100 collects measurement data including the amount of distortion measured by the multiple distortion sensors 81 (step S101). At this time, the control unit 101 fixes the load weight and collects measurement data including the amount of distortion, and when sufficient measurement data is obtained for the load weight, changes the load weight and collects measurement data including the amount of distortion at the changed load weight. The control unit 101 may collect measurement data sequentially while changing the load weight. In this manner, measurement data including the amount of distortion when the load weight is changed in various ways can be obtained. Note that the posture of the special purpose vehicle 1 when measuring the amount of distortion and the like may be not only a horizontal posture, but also various inclined postures such as a front downward posture, a rear downward posture, a left downward posture, and a right downward posture. The load weight may be given by loading items of known weight into the shipping box 4, or may be actually measured using a measuring instrument such as a truck scale. The measurement data collected in step S101 is stored for each load weight in a measurement value table TB1 of the storage unit 102.

[0053] The type of measurement data collected in step S101 may be selected according to the configuration of the learning model LM1 to be generated. For example, when generating a learning model LM1 that outputs a calculation result related to a load weight in response to input of measurement data related to the amount of distortion, inclination, and environmental temperature, the control unit 101 may collect measurement data related to the amount of distortion measured by the distortion sensor 81, the inclination measured by the inclinometer 82, and the environmental temperature measured by the thermometer 83. The same applies to the case of generating a learning model LM1 having a different input / output relationship from the above. For example, when generating a learning model LM1 that outputs a calculation result related to a load weight in response to input of measurement data related to the amount of distortion, the control unit 101 may collect only measurement data related to the amount of distortion measured by the distortion sensor 81.

[0054] After collecting the measurement data, the control unit 101 reads out the learning program PG1 from the storage unit 102 and executes it, thereby carrying out the following processing.

[0055] The control unit 101 selects a set of training data from the measurement value table TB1 (step S102). The training data includes a series of measurement data measured at the same time and the value of the loaded weight when the measurement data was obtained.

[0056] Next, the control unit 101 inputs the selected training data to the learning model LM1 (step S103), and executes a calculation by the learning model LM1 (step S104). That is, the control unit 101 inputs measurement data such as the amount of distortion, tilt, and environmental temperature to the nodes constituting the input layer of the learning model LM1, executes a calculation using the kernel of the intermediate layer, and outputs the calculation result from the output layer. Note that, in the initial stage before learning begins, initial values ​​are given to the definition information describing the learning model LM1.

[0057] Next, the control unit 101 evaluates the calculation result obtained in step S104 (step S105) and determines whether learning is complete or not (step S106). Specifically, the control unit 101 can evaluate the calculation result using an error function (also referred to as an objective function, loss function, or cost function) based on the calculation result obtained in step S104 and the training data. For example, the control unit 101 may determine that learning is complete when the error function becomes equal to or smaller than a threshold (or equal to or larger than a threshold) in the process of optimizing (minimizing or maximizing) the error function by a gradient descent method such as the steepest descent method. In order to avoid the problem of overlearning, a method such as cross-validation or early termination may be adopted to terminate learning at an appropriate timing.

[0058] When it is determined that the learning is not completed (S106: NO), the control unit 101 updates the connection weights between the nodes of the learning model LM1 (step S107), returns the process to step S102, and continues the learning using other training data. The control unit 101 can update the connection weights between the nodes by using the backpropagation method that sequentially updates the connection weights between the nodes from the output layer to the input layer of the learning model LM1.

[0059] When it is determined that the learning is completed (S106: YES), the control unit 101 stores the learned learning model LM1 in the storage unit 102 (step S108), and ends the process according to this flowchart.

[0060] As described above, the estimation device 100 according to this embodiment can generate a learning model LM1 by collecting measurement data including the amount of distortion when the load weight is known and using the load weight and the measurement data as training data.

[0061] In this embodiment, the learning model LM1 is generated in the estimation device 100, but an external server (not shown) that generates the learning model LM1 may be provided, and the learning model LM1 may be generated by the external server. In this case, the external server may acquire training data collected by the special purpose vehicle 1 by communication or the like, and generate the learning model LM1 using the acquired training data. Furthermore, the estimation device 100 may acquire the trained learning model LM1 from the external server by communication or the like, and store the acquired learning model LM1 in the storage unit 102.

[0062] In the operation phase, the estimation device 100 can estimate the loaded weight by inputting measurement data including the amount of distortion to the trained learning model LM1.

[0063] 8 is a flowchart for explaining the procedure for estimating the load weight using the learning model LM1. The control unit 101 of the estimation device 100 reads out the estimation program PG2 from the storage unit 102 and executes it to perform the following processes.

[0064] When the control unit 101 acquires measurement data including the amount of distortion measured by the multiple distortion sensors 81 through the input unit 104, it inputs the acquired measurement data to the learning model LM1 (step S121) and executes a calculation by the learning model LM1 (step S122). At this time, the control unit 101 provides the acquired measurement data to the nodes constituting the input layer of the learning model LM1. The data provided to the input layer is weighted by the connection weights determined using the training data and output to the intermediate layer. In the intermediate layer, a calculation using a kernel is executed, and the data is weighted by the connection weights determined by learning and output to the output layer. The nodes in the output layer output the calculation results related to the load weight.

[0065] The control unit 101 estimates the load weight based on the calculation result of the learning model LM1 (step S123). If the learning model LM1 is configured to output an estimated value of the load weight, the control unit 101 may estimate the output value of the learning model LM1 as the load weight. Also, if the learning model LM1 is configured to output the probability of a certain load weight, the control unit 101 can estimate the load weight by selecting the value of the load weight with the highest probability.

[0066] Next, the control unit 101 notifies the estimated load weight (step S124). At this time, the control unit 101 outputs information on the estimated load weight from the output unit 105 and displays it on the display device 120. FIG. 9 is a schematic diagram showing an example of a display of the load weight. FIG. 9 shows an example in which information on the estimated value of the load weight, the load ratio, and the estimated date and time is displayed as text information on the display device 120. Here, the estimated value of the load weight is the value of the load weight estimated using the above-mentioned learning model LM1. The load ratio is a value calculated as the ratio of the load weight (estimated value) to the upper limit value. The estimated date and time is the date and time when the load weight is estimated using the learning model LM1, and is information obtained from the built-in clock of the control unit 101, for example. The control unit 101 generates display screen data based on the estimated load weight estimated using the learning model LM1, the loading rate calculated as a percentage of the upper limit, and date and time information obtained from the built-in clock, and outputs the generated display screen data to the display device 120, thereby causing the display device 120 to display a screen such as that shown in Figure 9.

[0067] In this embodiment, the estimated loaded weight is displayed as text information on the display device 120, but the estimated loaded weight may be displayed diagrammatically as a graph or a meter. In addition, in this embodiment, the estimated load weight is displayed on the display device 120, but the information on the estimated load weight may be transmitted from the communication unit 106 to notify a user terminal or the like. If an audio output device such as a speaker is connected to the estimation device 100, the information on the estimated load weight may be output as audio.

[0068] Below, an example of actual measured values ​​of the measurement data and an estimation result using the learning model LM1 is shown. Fig. 10 is a graph showing the measurement results. The graph shown in Fig. 10 shows the results of measuring the amount of distortion, the tilt angle, and the environmental temperature using a distortion sensor 81, an inclinometer 82, and a thermometer 83 while the specially equipped vehicle 1 is stopped while traveling on a general road with a load of a known weight (4000 kg in this example) loaded, the load is unloaded at the destination, and the vehicle continues traveling without the load. The horizontal axis of the graph represents the elapsed time (more precisely, the measurement timing), and the vertical axis represents the amount of distortion (μST), the tilt angle (degrees), or the environmental temperature (°C).

[0069] In the graph shown in FIG. 10, the distortion amount of the front right axle represents the distortion amount measured by the distortion sensor 81A attached near the right end of the axle 23F of the front wheel 22F. Similarly, the distortion amounts of the front left axle, the rear right axle, and the rear left axle represent the distortion amounts measured by the distortion sensors 81B to 81D attached near the left end of the axle 23F of the front wheel 22F, near the right end of the axle 23R of the rear wheel 22R, and near the left end of the axle 23R of the rear wheel 22R, respectively. In addition, the inclination (roll and pitch) represents the inclination angle (roll and pitch) of the special-purpose vehicle 1 measured by the inclinometer 82, and the environmental temperature represents the temperature measured by the thermometer 83. As shown in FIG. 10, the distortion amount measured by the distortion sensor 81 includes the measurement value when the special-purpose vehicle 1 is inclined in the range of ±5 degrees in the front-rear and left-right directions, and the measurement value when the environmental temperature changes variously. These measured values ​​are stored in a measured value table TB1 in association with the elapsed time.

[0070] FIG. 11 is a graph showing the estimation result of the learning model LM1. The horizontal axis of the graph represents the elapsed time, and the vertical axis represents the estimated value (kg) of the load weight. In the example of FIG. 11, a part of the measurement data when the load weight is 4000 kg (see FIG. 10) and a part of the measurement data when the load weight is 0 kg (not shown) are used as training data to generate the learning model LM1, and the remaining measurement data is input to the learned learning model LM1 to estimate the load weight of the special-purpose vehicle 1. More specifically, of the 4809 sets of measurement values ​​of the measured distortion amount, tilt angle, and environmental temperature, 70% is used as training data to generate the learning model LM1, and the remaining 30% is input to the learning model LM1 to estimate the load weight of the special-purpose vehicle 1. Note that the measurement data is not limited to the load weights of 4000 kg and 0 kg, and it is also possible to use measurement data based on other load weights.

[0071] The solid line graph (after correction) shows the estimation results using the trained learning model LM1. For reference, the dashed line graph (before correction) shows the measurement results of the loaded weight measured using a conventional weight scale. When the loaded weight of special purpose vehicle 1 is 4000 kg, it can be seen that the conventional weight scale indicates measurement values ​​of approximately 2200 kg to 4500 kg, with an error of ±1150 kg with a median value of 3350 kg. In contrast, when the loaded weight of special purpose vehicle 1 is estimated using learning model LM1, an estimated value of approximately 4000±50 kg is obtained, showing that the loaded weight can be estimated with greater accuracy than the conventional weight scale.

[0072] Note that the estimation results shown in Figure 11 are based on a learning model LM1 generated using a portion of the measurement data when the load weight is 4000 kg and a portion of the measurement data when the load weight is 0 kg as training data. However, it is possible to estimate various load weights by collecting measurement data while changing the load weight in various ways and generating a learning model LM1 using the collected measurement data.

[0073] 11 shows the result of estimating the load weight using the learning model LM1 configured to output a calculation result regarding the load weight in response to the input of measurement data related to the amount of distortion measured by the distortion sensor 81, the inclination measured by the inclinometer 82, and the environmental temperature measured by the thermometer 83, but the input / output relationship in the learning model LM1 is not limited to the above and can be set appropriately. As described above, the learning model LM1 may be configured to output a calculation result regarding the load weight in response to the input of measurement data further including the cylinder pressure of the hydraulic cylinder 52 measured by the pressure gauge 84. The learning model LM1 may also be configured to receive measurement data related to the amount of distortion measured by the distortion sensor 81 and output a calculation result regarding the load weight. Furthermore, two or three selected measurement data from the measurement data of the distortion sensors 81A to 81D may be input to the learning model LM1 and output a calculation result regarding the load weight. The learning model LM1 may also be configured to receive measurement data of the amount of distortion and either one of the measurement data of the inclination or the environmental temperature and output a calculation result regarding the load weight.

[0074] As described above, in this embodiment, by using the learning model LM1 which has learned the relationship between the measurement data including the distortion amount and the load weight, it is possible to accurately estimate the load weight regardless of whether the special purpose vehicle 1 is inclined or not, and to report the estimated load weight.

[0075] In this embodiment, the support vector regression model has been described as an example of the learning model LM1, but a regression analysis method such as linear regression or logistic regression may be used. In addition, a learning model learned by other methods including a method using a search tree such as a decision tree, a regression tree, a random forest, or a gradient boosting tree, a Bayesian estimation method including naive Bayes, a time series prediction method including AR (Auto Regressive), MA (Moving Average), a state space model, a clustering method including K-nearest neighbors, a method using ensemble learning including boosting and bagging, a clustering method including hierarchical clustering, non-hierarchical clustering, a topic model, association analysis, and emphasis filtering may be used. In addition, the learning model LM1 may be configured by a neural network by deep learning, a convolutional neural network, a recurrent neural network, or the like.

[0076] In the above embodiment, the strain sensor 81 is made of a strain gauge, but the present invention is not limited to this, and the strain sensor 81 may be a strain gauge type load cell. In that case, the shape of the load cell may be either a bar type or a pin type.

[0077] In the above embodiment, the strain sensor 81 is attached to the axles 23F, 23R or the subframe 30, but the present invention is not limited thereto, and the strain sensor 81 may be attached to other locations. For example, the strain sensor may be attached to the cargo box 4 in the dump truck 3, or to the hinge shaft 31 connecting the cargo box 4 and the subframe 30. The strain sensor 81 may also be attached to the chassis frame 21 or the suspension system (not shown). For example, when attaching to the suspension system, it is useful to measure the degree to which the mounting portion of the strain sensor 81 sinks compared to other portions. Therefore, in such a case, a displacement sensor (such as a laser distance sensor) that measures the amount of sinking relative to other portions rather than measuring the amount of distortion at the mounting portion may be used. Note that when measuring the amount of displacement such as the amount of sinking, it is also easily affected by the air pressure of the front wheels 22F or the rear wheels 22R of the specially equipped vehicle 1, so it is more preferable to add these measurement data to the measurement value table TB1 using a tire pressure monitor or the like.

[0078] In the above embodiment, the estimation device 100 is provided inside the chassis frame 21 or the cab 20 of the specially equipped vehicle 1, but the present invention is not limited to this, and the estimation device 100 may be provided at a location separate from the specially equipped vehicle 1, such as a management center. In this case, the specially equipped vehicle 1 may be provided with various sensors such as a strain sensor 81, an inclinometer 82, a thermometer 83, and a pressure gauge 84, a communication device, and a display device, and data from these various sensors may be transmitted to the estimation device 100 by the communication device. The estimation device 100 may process the data received from the various sensors and transmit the results to the communication device of the specially equipped vehicle 1, thereby causing the display device of the specially equipped vehicle 1 to display the estimated load weight.

[0079] Furthermore, in the above embodiment, the load weight is estimated based on the measurement data related to the amount of strain from the multiple strain sensors 81, but it may be estimated based on other indices, such as the measurement data related to the amount of hydraulic pressure change from multiple hydraulic sensors. For example, a hydraulic load cell may be provided near the hinge shaft 31 together with a pressure gauge 84 that measures the cylinder pressure of the hydraulic cylinder 52 in the hoist mechanism 5. In this case, the hydraulic cylinder 52 is slightly extended to slightly tilt the cargo box 4, and the load weight is estimated based on the measurement data of the hydraulic pressure from the pressure gauge 84 and the load cell. Such measurement data is not limited to the above-mentioned strain, hydraulic pressure, etc., and other displacement amounts may also be measured as appropriate.

[0080] (Embodiment 2) In the first embodiment, the measurement data relating to the amount of strain is acquired from the strain sensor 81 attached to the axles 23F, 23R, but the attachment location of the strain sensor 81 is not limited to the axles 23F, 23R. In the second embodiment, a configuration in which a strain sensor 81 is attached to a structure fixed to a chassis frame 21 will be described.

[0081] Fig. 12 is a side view showing the overall configuration of a specially equipped vehicle 1 according to embodiment 2. The specially equipped vehicle 1 shown in Fig. 12 differs from embodiment 1 only in the mounting location of the strain sensor 81. The strain sensor 81 in embodiment 2 is mounted at multiple locations on a structure fixed to the chassis frame 21. One example of a structure fixed to the chassis frame 21 is the subframe 30. The example in Fig. 12 shows a configuration in which the strain sensor 81 is mounted at two locations on the subframe 30, spaced apart in the fore-aft direction.

[0082] 12, the strain sensors 81 are attached to two locations spaced apart in the fore-aft direction of the subframe 30, but the strain sensors 81 may be attached to two locations symmetrical to the left and right with respect to the center line in the fore-aft direction of the specially equipped vehicle 1. The number of attached strain sensors 81 is not limited to two, and the strain sensors 81 may be attached to three or more locations. Furthermore, a dedicated bracket for attaching the strain sensors 81 may be prepared, and the strain sensors 81 may be attached to the subframe 30 via the dedicated bracket.

[0083] The estimation device 100 generates a learning model LM1 similar to that of embodiment 1 using measurement data including the amount of strain measured by a strain sensor 81 attached to the subframe 30, and can estimate the weight of the cargo loaded in the shipping box 4 using the generated learning model LM1.

[0084] As described above, in the second embodiment, the strain sensor 81 is attached to the subframe 30, so that the loaded weight can be estimated without making any modifications to the truck chassis 2.

[0085] (Embodiment 3) In embodiment 3, a configuration is described in which the magnitude of hydraulic pressure acting on a hydraulic actuator (in this embodiment, hydraulic cylinder 52) for raising and lowering the cargo box 4 is measured, and the load weight is estimated based on the measured magnitude of hydraulic pressure.

[0086] FIG. 13 is an explanatory diagram for explaining the lifting mechanism of the cargo box 4 in the specially equipped vehicle 1. The specially equipped vehicle 1 is equipped with a hoist mechanism 5 as a mechanism for lifting and lowering the cargo box 4. As described above, the hoist mechanism 5 is equipped with a lift arm 51, a hydraulic cylinder 52, and a tension link 53. When the hydraulic cylinder 52 of the hoist mechanism 5 is extended, the front part of the cargo box 4 is lifted and the cargo box 4 rotates in a direction in which the inclination angle increases. In this embodiment, the rotation of the cargo box 4 in a direction in which the inclination angle increases is also referred to as the lifting of the cargo box 4. On the other hand, when the hydraulic cylinder 52 of the hoist mechanism 5 is contracted, the front part of the cargo box 4 is lowered and the cargo box 4 rotates in a direction in which the inclination angle decreases. In this embodiment, the rotation of the cargo box 4 in a direction in which the inclination angle decreases is also referred to as the lowering of the cargo box 4.

[0087] The hydraulic mechanism that expands and contracts the hydraulic cylinder 52 includes a hydraulic pump 61, a hydraulic oil tank 62, and a control valve 63. The hydraulic pump 61, which is a hydraulic supply source, is driven by power from an engine 70 transmitted via a PTO (Power Take-Off) 71 to pump up hydraulic oil in the hydraulic oil tank 62 through hydraulic piping 64 and supply the hydraulic oil (pressurized oil) to the hydraulic cylinder 52 through a main pipe 65 connected to a discharge port. Note that a PTO switch 72 provided in the cab 20 switches between connecting and disconnecting the power transmission of the engine 70.

[0088] The supply direction of the hydraulic oil discharged from the hydraulic pump 61 is switched by a control valve 63 operated by a manual operation lever 67. For example, when the control valve 63 is in a neutral position by operation of the operation lever 67, hydraulic oil is not supplied from the hydraulic pump 61 to the hydraulic cylinder 52, and the loading box 4 is not tilted. When the operation lever 67 is operated to the raised position, the control valve 63 is switched and hydraulic oil (pressurized oil) is supplied from the hydraulic pump 61 to the hydraulic cylinder 52. The hydraulic cylinder 52 extends when hydraulic oil is supplied, and lifts the loading box 4. On the other hand, when the operation lever 67 is operated to the lowered position, the control valve 63 is switched and the hydraulic oil supplied to the hydraulic cylinder 52 flows back to the hydraulic oil tank 62. Accordingly, the hydraulic cylinder 52 contracts, and the loading box 4 is lowered.

[0089] The hydraulic mechanism is provided with a pressure gauge 84 for measuring the magnitude of hydraulic pressure acting on the hydraulic cylinder 52. The specially equipped vehicle 1 is also provided with an inclinometer 82 for measuring the inclination (pitch and roll) of the truck chassis 2, and an inclinometer 85 for measuring the inclination (pitch and roll) of the cargo box 4. Furthermore, the specially equipped vehicle 1 may be provided with the strain sensor 81 and thermometer 83 described in the first embodiment.

[0090] The inventors of the present application conducted a detailed investigation into the relationship between the pitch angle of the truck chassis 2 and the hydraulic pressure value and discovered that a gap occurs between the hydraulic pressure value when the cargo box 4 is raised and then stopped, and the hydraulic pressure value when the cargo box 4 is lowered and then stopped.

[0091] FIG. 14 is a graph showing the relationship between the pitch angle of the truck chassis 2 and the hydraulic pressure value. The horizontal axis of the graph represents the pitch angle (deg) of the truck chassis 2, and the vertical axis represents the hydraulic pressure value (MPa). The graph shown in FIG. 14 shows the results of measuring the hydraulic pressure value while variously changing the pitch angle of the truck chassis 2 and the known load weight of the cargo box 4. When the pitch angle is less than 0 degrees, the special purpose vehicle 1 is in a state where the front is lowered, and when the pitch angle is more than 0 degrees, the special purpose vehicle 1 is in a state where the front is raised.

[0092] In the graph of Fig. 14, the oil pressure values ​​measured when the cargo box 4 is stopped at a predetermined dump angle (e.g., 0.5 degrees) after being raised are indicated by black circles. The oil pressure values ​​measured when the cargo box 4 is stopped at a predetermined dump angle (e.g., 0.5 degrees) after being lowered are indicated by white circles. The dump angle is the relative angle of the cargo box 4 with respect to the truck chassis 2, and is a value calculated by subtracting the inclination angle of the truck chassis 2 (measurement value of the inclinometer 82) from the inclination angle of the cargo box 4 (measurement value of the inclinometer 85).

[0093] Moreover, each dashed line in the graph of Fig. 14 is a regression line of a plurality of black dots measured when the pitch angle of the truck chassis 2 is changed at a predetermined known load weight (e.g., 10 tons) among all the black dots in the graph. Moreover, each dashed dot line in the graph of Fig. 14 is a regression line of a plurality of white dots measured when the pitch angle of the truck chassis 2 is changed at a predetermined known load weight (e.g., 10 tons) among all the white dots in the graph.

[0094] From the graph shown in Fig. 14, it can be seen that even with the same load weight, the hydraulic pressure value increases as the pitch angle of the truck chassis 2 decreases (the more downward the front). It can also be seen that even with the same load weight, a gap occurs between the hydraulic pressure value (hydraulic pressure value shown by black circles) measured when the cargo box 4 is raised and then stopped at a specified dump angle, and the hydraulic pressure value (hydraulic pressure value shown by white circles) measured when the cargo box 4 is lowered and then stopped at a specified dump angle. It can also be seen that the gap increases as the load weight increases.

[0095] The load weight estimation system in the third embodiment estimates the weight (load weight) of the load loaded on the cargo box 4 using measurement data including the hydraulic pressure value. The inventors of the present application have found that a gap occurs between the hydraulic pressure value when the cargo box 4 is stopped after being raised and the hydraulic pressure value when the cargo box 4 is stopped after being lowered. Therefore, the load weight estimation system in the third embodiment estimates the load weight using different learning models for when the cargo box 4 is stopped after being raised and when the cargo box 4 is stopped after being lowered.

[0096] 15 is a block diagram for explaining the configuration of the loaded weight estimation system in the embodiment 3. The estimation device 100 includes a control unit 101, a storage unit 102, an operation unit 103, an input unit 104, an output unit 105, and a communication unit 106. The configurations of these hardware units are similar to those explained in the embodiment 1, and therefore the explanations thereof will be omitted.

[0097] In the third embodiment, a distinction is made between a state where the packing box 4 stops after rising (first stop state) and a state where the packing box 4 stops after falling (second stop state), and different learning models are used for the two states to estimate the load weight of the packing box 4. That is, in the third embodiment, a learning model LM10 for estimating the load weight using measurement data including the hydraulic value measured in the first stop state, and a learning model LM20 for estimating the load weight using measurement data including the hydraulic value measured in the second stop state are prepared. The learning models LM10 and LM20 are defined by their respective definition information, and are stored in, for example, the storage unit 102. The definition information of the learning models LM10 and LM20 includes parameters that define the model structure, such as the type and number of layers and the number of nodes, and parameters that are determined by learning, such as connection weights set between nodes.

[0098] When the control unit 101 of the estimation device 100 acquires measurement data including a hydraulic pressure value measured in the first stop state, the control unit 101 inputs the acquired measurement data to the learning model LM10 and estimates the load weight of the cargo box 4 by performing a calculation using the learning model LM10. When the control unit 101 acquires measurement data including a hydraulic pressure value measured in the second stop state, the control unit 101 inputs the acquired measurement data to the learning model LM20 and estimates the load weight of the cargo box 4 by performing a calculation using the learning model LM20. Note that a metering switch 89 provided inside the cab 20 for giving an instruction to start the estimation process may be connected to the input unit 104 of the estimation device 100.

[0099] 16 is a schematic diagram for explaining the configuration of learning models LM10 and LM20. Learning models LM10 and LM20 are similar to learning model LM1 explained in embodiment 1, and include an input layer to which various measurement data are input, an intermediate layer in which a predetermined calculation is performed based on the measurement data input to the input layer, and an output layer that combines outputs from the intermediate layer and outputs the calculation results.

[0100] FIG. 16A shows the configuration of the learning model LM10. The hydraulic pressure value measured by the pressure gauge 84 in the first stop state (state after the lifting of the cargo box 4 has stopped) is input to the input layer of the learning model LM10. The measurement data input to the input layer of the learning model LM1 may further include at least one of the tilt angle of the truck chassis 2 measured by the inclinometer 82, the distortion of the axles 23F, 23R measured by the distortion sensor 81, and the environmental temperature measured by the thermometer 83. The environmental temperature may be the temperature around the special-purpose vehicle 1 or the oil temperature. The middle layer has a plurality of layers, and each layer is composed of a plurality of nodes. When the above measurement data is input to the input layer of the learning model LM10, the control unit 101 of the estimation device 100 executes a predetermined calculation using parameters set in each node constituting the middle layer. The output layer combines the outputs from the middle layer and outputs an estimated value of the load weight in the cargo box 4 as a calculation result.

[0101] The learning model LM10 is trained to output information related to an estimated value of the load weight of the packing box 4 in response to the input of the various measurement data described above. When generating the learning model LM10, for example, the estimation device 100 collects the various measurement data described above while changing the weight (presumed to be known) of the load to be loaded on the packing box 4 in various ways. The estimation device 100 uses a data set including a large number of collected measurement data and known load weights as training data, and trains the learning model ML10 to output information related to an estimated value of the load weight in response to the input of the measurement data. The learning algorithm is the same as that in the first embodiment, and therefore the description thereof will be omitted. The learning model LM10 may be generated in an external server (not shown). In this case, the estimation device 100 may acquire the learning model LM10 by communication and store it in the storage unit 102.

[0102] Fig. 16B shows the configuration of the learning model LM20. The learning model LM20 differs from the learning model LM10 in that the hydraulic pressure value in the second stop state (the state after the descent of the box 4 has stopped) is input to the input layer instead of the hydraulic pressure value in the first stop state, but the other configurations are the same as those of the learning model LM10. The learning model LM20 is generated by learning using measurement data including the hydraulic pressure value measured in the second stop state and a data set including the known load weight as training data.

[0103] 17 is a flowchart explaining the procedure for estimating the loaded weight in the embodiment 3. The control unit 101 of the estimation device 100 judges whether or not the weighing switch 89 is turned on by monitoring a signal input through the input unit 104 (step S301). If the switch is not turned on (S301: NO), the control unit 101 waits until the weighing switch 89 is turned on.

[0104] When the weighing switch 89 is turned on (S301: YES), the control unit 101 starts the load weight estimation process. When the estimation process starts, an instruction may be given to the user to adjust the dump angle to a predetermined angle. The predetermined angle is, for example, an angle in the range of more than 0.1 degrees and less than 0.8 degrees. The control unit 101 may give the instruction by displaying text information on the display device 120, or may give the instruction by outputting sound from a speaker (not shown). In the third embodiment, the dump angle is manually adjusted using the operating lever 67.

[0105] The control unit 101 sequentially acquires measurement data of the tilt angle measured in chronological order by the inclinometers 82, 85 through the input unit 104, and detects the current dump angle based on the acquired measurement data (step S302). The control unit 101 can obtain the dump angle by subtracting the tilt angle of the truck chassis 2 obtained as the measurement value of the inclinometer 82 from the tilt angle of the cargo box 4 obtained as the measurement value of the inclinometer 85. The control unit 101 stores the detected dump angles in chronological order in the memory unit 102.

[0106] The control unit 101 determines whether the dump angle detected in step S302 is greater than the minimum angle θ1 (step S303). The minimum angle θ1 is a value that is set as the minimum value of an angle range suitable for measuring the load weight of the cargo box 4. An example of the minimum angle θ1 is 0.1 degrees.

[0107] When it is determined that the current dump angle is equal to or less than the minimum angle θ1 (S303: NO), the control unit 101 instructs the user to raise the packing box 4 (step S304). The control unit 101 instructs the user by displaying text information to the effect that the packing box 4 should be raised on the display device 120. Alternatively, the control unit 101 may instruct the user by outputting a sound to the effect that the packing box 4 should be raised from a speaker (not shown). After instructing the user, the control unit 101 returns the process to step S302.

[0108] When it is determined that the current dump angle is larger than the minimum angle θ1 (S303: YES), the control unit 101 determines whether the current dump angle is smaller than the maximum angle θ2 (step S305). The maximum angle θ2 is a value that is set as the maximum value of an angle range suitable for measuring the load weight of the cargo box 4. An example of the maximum angle θ2 is 0.8 degrees.

[0109] When it is determined that the current dump angle is equal to or greater than the maximum angle θ2 (S305: NO), the control unit 101 instructs the user to lower the packing box 4 (step S306). The control unit 101 instructs the user, for example, by displaying text information on the display device 120 to the effect that the packing box 4 should be lowered. Alternatively, the control unit 101 may instruct the user by outputting a sound from a speaker (not shown) to the effect that the packing box 4 should be lowered. After instructing the user, the control unit 101 returns the process to step S302.

[0110] If it is determined that the current dump angle is smaller than the maximum angle θ2 (S305: YES), the control unit 101 refers to the output of the built-in timer and determines whether or not a predetermined time has elapsed since the dump angle entered the set angle range (step S307). If the predetermined time has not elapsed (S307: NO), the control unit 101 waits until the predetermined time has elapsed.

[0111] When it is determined that the predetermined time has elapsed (S307: YES), the control unit 101 determines whether the cargo box 4 was in a stopped state until the predetermined time elapsed (step S308). The control unit 101 can determine whether the cargo box 4 was in a stopped state by determining whether there has been a change from the dump angle history data stored in the memory unit 102. When the cargo box 4 was not in a stopped state (S308: NO), the control unit 101 returns the process to step S302.

[0112] If the container 4 remains stationary until the predetermined time has elapsed (S308: YES), the control unit 101 notifies the user that the loading weight is ready to be weighed (step S309). For example, the control unit 101 causes the display device 120 to display text information indicating that the loading weight is ready to be weighed. Alternatively, the control unit 101 may cause a speaker (not shown) to output a sound indicating that the loading weight is ready to be weighed.

[0113] Next, the control unit 101 determines whether the container 4 has stopped after rising (step S310). The control unit 101 can determine whether the container 4 has stopped after rising from the history data of the dump angle stored in the memory unit 102.

[0114] When the control unit 101 determines that the container 4 has stopped after ascending (S310: YES), it sets the learning model used to estimate the load weight to the learning model LM10 for stopping ascent (step S311). On the other hand, when the control unit 101 determines that the container 4 has stopped after descending (S310: NO), it sets the learning model used to estimate the load weight to the learning model LM20 for stopping descent (step S312).

[0115] The control unit 101 obtains measurement data including hydraulic pressure values ​​from the input unit 104, and inputs the obtained measurement data to the learning model LM10 for ascent stop set in step S311 or the learning model LM20 for descent stop set in step S312, thereby estimating the load weight of the cargo box 4 (step S313). At this time, the control unit 101 provides the obtained measurement data to each node constituting the input layer of the learning model LM10 (or the learning model LM20), executes calculations by the intermediate layer, and obtains the calculation results output from the output layer, thereby estimating the load weight of the cargo box 4. Note that the measurement data provided to the nodes of the input layer may be only hydraulic pressure values, or may further include at least one of the tilt angle of the truck chassis 2, the distortion of the axles 23F, 23R, and the environmental temperature in addition to the hydraulic pressure values.

[0116] Next, the control unit 101 notifies the estimated load weight (step S314). At this time, the control unit 101 outputs information on the estimated load weight from the output unit 105 and displays it on the display device 120. The control unit 101 may display the information on the load weight as text information on the display device 120, or may use a schematic display method such as a graph display or a meter display. The control unit 101 may also be configured to output the information on the estimated load weight as sound from a speaker (not shown). After this series of processes is completed, the weighing switch 89 is turned off, and the container 4 is lowered until the dump angle becomes 0 degrees.

[0117] The inventors of the present application have found that the hydraulic pressure value measured by the pressure gauge 84 differs between a state in which the cargo box 4 stops after rising and a state in which the cargo box 4 stops after falling, even if the dump angle is the same. For this reason, if the same learning model is used to estimate the load weight for both, there is a possibility that the estimation accuracy will be low. However, in the third embodiment, the learning model LM10 for stopping the rise and the learning model LM20 for stopping the descent are selectively used to estimate the load weight, so that the decrease in the estimation accuracy can be suppressed.

[0118] (Embodiment 4) A configuration for automatically adjusting the dump angle will be described in the fourth embodiment. Note that the configuration of the specially equipped vehicle 1 is similar to that of the third embodiment, and therefore a description thereof will be omitted.

[0119] 18 is a block diagram illustrating the configuration of a loaded weight estimation system according to embodiment 4. The loaded weight estimation system according to embodiment 4 includes an estimation device 100 and a lifting control device 200 connected to the estimation device 100. The estimation device 100 is similar to that described in embodiment 3, and includes a control unit 101, a storage unit 102, an operation unit 103, an input unit 104, an output unit 105, and a communication unit 106.

[0120] The lifting control device 200 includes an input unit 201, a control unit 202, and an output unit 203, and controls the operation of a hydraulic mechanism included in the specially equipped vehicle 1 to control the lifting and lowering of the cargo box 4. The input unit 201 includes an input interface. Information output from the estimation device 100, operation information of the operating lever 67, operation information of the PTO switch 72, etc. are input to the input unit 201. The information input to the input unit 201 is output to the control unit 202.

[0121] The control unit 202 is configured, for example, by a PLC (Programmable Logic Controller). The control unit 202 generates a control signal for controlling the lifting and lowering of the packing box 4 based on information input through the input unit 201 according to programmed logic. The control unit 202 outputs the generated control signal to the control valve 63 from the output unit 203. The output unit 203 has an output interface, and is connected to the control valve 63, the estimation device 100, and the like. Note that the control valve 63 in the fourth embodiment is configured as an electrically controllable electromagnetic control valve.

[0122] In this embodiment, the loaded weight estimation system is configured to include the estimation device 100 and the lifting control device 200 as separate entities, but the two may be integrated into one body.

[0123] 19 is a flowchart explaining the procedure for estimating the loaded weight in the embodiment 4. The control unit 101 of the estimation device 100 determines whether or not the weighing switch 89 is turned on by monitoring a signal input through the input unit 104 (step S401). If the switch is not turned on (S401: NO), the control unit 101 waits until the weighing switch 89 is turned on.

[0124] When the metering switch 89 is turned on (S401: YES), the control unit 101 judges whether the PTO switch 72 is on or not (step S402). When the PTO switch 72 is not on (S402: NO), the control unit 101 instructs the user to turn on the PTO switch 72 (step S403) and switches the power transmission destination of the engine 70 to the hydraulic pump 61. The instruction to the user may be given by displaying text information on the display device 120, or may be given by outputting the instruction as sound from a speaker (not shown).

[0125] If the PTO switch 72 is on (S402: YES), the control unit 101 notifies the user that the dump angle will be adjusted (step S404). For example, the control unit 101 causes the display device 120 to display text information indicating that the dump angle will be adjusted. Alternatively, the control unit 101 may cause a speaker (not shown) to output a sound indicating that the dump angle will be adjusted.

[0126] Next, the control unit 101 instructs the lifting control device 200 to lift the packing box 4 (step S405). Specifically, the control unit 101 generates a control signal instructing the lifting control device 200 to lift the packing box 4, and issues the instruction to the lifting control device 200 by outputting the generated control signal to the lifting control device 200 from the output unit 105. In response to the instruction from the estimation device 100, the control unit 202 of the lifting control device 200 outputs a control signal for lifting the packing box 4 to the control valve 63, thereby lifting the packing box 4.

[0127] The control unit 101 sequentially acquires measurement data of the tilt angle measured in time series by the inclinometers 82 and 85 through the input unit 104, and detects the current dump angle based on the acquired measurement data (step S406).

[0128] The control unit 101 determines whether the dump angle detected in step S406 is greater than the minimum angle θ1 (step S407). The minimum angle θ1 is a value that is set as the minimum value of an angle range suitable for measuring the load weight of the cargo box 4. An example of the minimum angle θ1 is 0.1 degrees.

[0129] When it is determined that the current dump angle is equal to or smaller than the minimum angle θ1 (S407: NO), the control unit 101 returns the process to step S405 and continues the lift control of the container 4.

[0130] When it is determined that the current dump angle is larger than the minimum angle θ1 (S407: YES), the control unit 101 determines whether the current dump angle is smaller than the maximum angle θ2 (step S408). The maximum angle θ2 is a value that is set as the maximum value of an angle range suitable for measuring the load weight of the cargo box 4. An example of the maximum angle θ2 is 0.8 degrees.

[0131] When it is determined that the current dump angle is equal to or greater than the maximum angle θ2 (S408: NO), the control unit 101 notifies an error (step S409) because the inclination angle of the cargo box 4 is outside the angle range suitable for measuring the load weight, and ends the process according to this flowchart. After notifying the error, the control unit 101 may instruct the lifting control device 200 to lower the cargo box 4.

[0132] When it is determined that the current dump angle is smaller than the maximum angle θ2 (S408: YES), the control unit 101 instructs the lifting control device 200 to stop after the lifting of the packing box 4 (step S410). Specifically, the control unit 101 generates a control signal instructing the lifting control device 200 to stop the packing box 4, and outputs the generated control signal to the lifting control device 200 from the output unit 105, thereby instructing the lifting control device 200. In response to the instruction from the estimation device 100, the control unit 202 of the lifting control device 200 outputs a control signal for stopping the packing box 4 to the control valve 63, thereby stopping the packing box 4.

[0133] Next, the control unit 101 notifies the user that the load weight is ready to be weighed (step S411). For example, the control unit 101 causes the display device 120 to display text information indicating that the load weight is ready to be weighed. Alternatively, the control unit 101 may cause a speaker (not shown) to output audio information indicating that the load weight is ready to be weighed.

[0134] Next, the control unit 101 sets the learning model used to estimate the load weight to the learning model LM10 for stopping ascent (step S412). The control unit 101 acquires measurement data including hydraulic pressure values ​​from the input unit 104, and inputs the acquired measurement data to the learning model LM10 for stopping ascent set in step S412, thereby estimating the load weight in the container 4 (step S413).

[0135] Next, the control unit 101 notifies the estimated load weight (step S414). At this time, the control unit 101 outputs information on the estimated load weight from the output unit 105 and causes the display device 120 to display it. The control unit 101 may cause the display device 120 to display the information on the load weight as text information, or may use a schematic display method such as a graph display or a meter display. The control unit 101 may also be configured to output the information on the estimated load weight as audio information from a speaker (not shown).

[0136] The control unit 101 may estimate the loaded weight, notify the user, and then instruct the lifting control device 200 to lower the packing box 4. In response to the instruction from the estimation device 100, the control unit 202 of the lifting control device 200 may lower the packing box 4 by outputting a control signal for lowering the packing box 4 to the control valve 63.

[0137] When the load weight is measured manually, the dump angle needs to be adjusted to a predetermined angle (e.g., 0.5 degrees), which may be troublesome for the user. In contrast, in this embodiment, the load weight can be measured automatically by operating the weighing switch 89, which reduces the troublesomeness of the operation.

[0138] In addition, in embodiment 4, since the load weight is estimated after the ascent of the cargo box 4 stops, it is sufficient for the memory unit 102 to store the learning model LM10 for stopping the ascent, and it is not necessary to store the learning model LM20 for stopping the descent.

[0139] (Embodiment 5) In the fifth embodiment, a configuration for calibrating the estimation result of the learning model LM1 in accordance with each specially equipped vehicle 1 will be described.

[0140] FIG. 20 is an explanatory diagram for explaining the calibration method. In the fifth embodiment, a calibration layer CL is provided after the learning model LM1. The calibration layer CL calibrates the output (load weight) of the learning model LM1 according to each individual special purpose vehicle 1. The parameters for calibration are determined at the time of shipping the special purpose vehicle 1. For example, the output of the learning model LM1 when a load of a known weight is loaded is obtained, and parameters that convert the output of the learning model LM1 into the actual load weight can be derived in advance. For example, if the output of the learning model LM1 is X1 and the actual load weight is X2, parameters a and b that satisfy X2=aX1+b can be derived.

[0141] Calibration parameters (parameters a and b in the above example) derived in advance are stored in the memory unit 102 of the estimation device 100. When the control unit 101 inputs the measurement data to the learning model LM1 and obtains the calculation result of the learning model LM1, the control unit 101 reads out the calibration parameters stored in the memory unit 102 and calibrates the calculation result, thereby obtaining an estimated value of the calibrated loaded weight.

[0142] As described above, in embodiment 5, the calculation results of the learning model LM1 can be calibrated to accommodate multiple specially equipped vehicles 1 having different sizes, shapes, weights, etc. of the cargo boxes 4, so there is no need to generate a learning model LM1 that depends on each individual specially equipped vehicle 1; instead, a standard learning model can be generated in an external computer and installed in each estimation device 100.

[0143] In addition, the above-mentioned calibration technique can be applied not only to the learning model LM1 described in embodiment 1, but also to the learning models LM10 and LM20 described in embodiment 3, and the learning model LM2 described in embodiment 7 below.

[0144] (Embodiment 6) In the sixth embodiment, a configuration will be described in which the loading state is determined from an estimated loading weight, and the loading state is notified in a manner according to the determination result.

[0145] 21 is a flowchart for explaining the procedure for notifying the loading state. The control unit 101 of the estimation device 100 executes the following process at regular intervals after the loading operation into the packing box 4 has started, for example.

[0146] The control unit 101 estimates the load weight in a procedure similar to that shown in the flowchart of Fig. 8. That is, the control unit 101 inputs measurement data including the amount of distortion measured by the multiple distortion sensors 81 to the learning model LM1 (step S601), executes a calculation by the learning model LM1 (step S602), and estimates the load weight based on the calculation result of the learning model LM1 (step S603).

[0147] The control unit 101 judges the loading state based on the estimated loading weight (step S604). For example, the control unit 101 stores the estimated loading weight in an internal memory, compares the loading weight estimated in step S603 with a preset upper limit, and judges the loading state to be "loadable" if the estimated loading weight still has room (for example, if it is less than 90% of the upper limit). The control unit 101 may also compare the estimated loading weight with a preset upper limit, and judge the loading state to be "close to the upper limit" if the estimated loading weight is close to the upper limit (for example, if it is more than 90% and less than 100% of the upper limit). Furthermore, the control unit 101 may judge the loading state to be a "loading stop" state if the estimated loading weight reaches the upper limit. Furthermore, the control unit 101 may judge the loading state to be "overloaded" if the estimated loading weight exceeds a preset upper limit of the loading weight.

[0148] The control unit 101 notifies the loading state in a manner according to the result of the determination in step S604 (step S605).

[0149] Fig. 22 is a schematic diagram showing an example of notification of the loading state. Fig. 22 shows the specially equipped vehicle 1 as viewed from the rear. In the example of Fig. 22, the display device 120 is composed of an indicator lamp that can be turned on, blinking, or turned off in blue or red, and is provided at the top of the rear side of the front panel 41.

[0150] FIG. 22A shows an example of a notification when the loading status is determined to be "loading possible". When the control unit 101 determines that the current loading status is "loading possible" based on the loading weight estimated using the learning model LM1, it executes control to cause the display device 120 to flash, for example, in blue only once per second. As a result, the display device 120 transitions from an off state to a state in which it flashes slowly in blue. By checking the display mode on the display device 120 (in this case, the blue light flashing slowly), the worker can understand that the current loading status is "loading possible".

[0151] FIG. 22B shows an example of a notification when it is determined that the loading state is "close to the upper limit". When the control unit 101 determines that the current loading state is "close to the upper limit" based on the loading weight estimated using the learning model LM1, it executes control to cause the display device 120 to blink, for example, in blue five times per second. As a result, the display device 120 transitions from a slow blinking state to a fast blinking state. By checking the display mode on the display device 120 (in this case, fast blinking in blue), the worker can understand that the current loading state is "close to the upper limit".

[0152] FIG. 22C shows an example of a notification when the loading state is determined to be "loading stop." When the control unit 101 determines that the current loading state is "loading stop" based on the loading weight estimated using the learning model LM1, it executes control to light the display device 120, for example, in blue. As a result, the display device 120 transitions from a flashing blue state to a lit state. By checking the display state on the display device 120 (in this case, lit in blue), the worker can understand that the current loading state is "loading stop."

[0153] FIG. 22D shows an example of a notification when the loading state is determined to be "overloaded". When the control unit 101 determines that the current loading weight is "overloaded" based on the loading weight estimated using the learning model LM1, it executes control to cause the display device 120 to flash, for example, in red. As a result, the display device 120 transitions from a blue lit state to a flashing red state. The worker can recognize that the current loading state is "overloaded" by checking the display state on the display device 120 (in this case, flashing red).

[0154] As described above, in this embodiment, the display mode on the display device 120 is changed, so that the current loading state of the packing box 4 can be notified to the worker.

[0155] In the present embodiment, as an example, a configuration in which the display device 120 is provided on the upper part of the rear side of the front panel 41 has been described, but the installation location of the display device 120 is not limited to the location shown in FIG. 22, and the display device 120 may be installed in any installation location as long as it is a location where the worker can see it. For example, the display device 120 may be installed on a side portion of the front panel 41, or on a side panel 42 or a rear panel 43 other than the front panel 41. The display device 120 may also be installed near the driver's seat. Furthermore, a configuration in which a notification is sent by communication to a portable terminal carried by the worker may be used.

[0156] In addition, in this embodiment, the display mode on the display device 120 is changed according to the loading state by changing the color and the lighting / flashing state, but characters or figures according to the loading state may be displayed on the display device 120. Furthermore, the loading state may be notified not only by display on the display device 120 but also by sound or voice.

[0157] (Embodiment 7) In the seventh embodiment, a configuration will be described in which the state of the host vehicle is detected and information relating to the detected state of the host vehicle is notified.

[0158] 23 is a side view showing the overall configuration of the specially equipped vehicle 1 according to embodiment 7. In addition to the above-mentioned configuration, the specially equipped vehicle 1 according to embodiment 7 is equipped with an inclinometer 85 that measures the inclination of the loading platform, and an imaging device 86 that images the inside of the loading box 4. The specially equipped vehicle 1 may further be equipped with a GPS (Global Positioning System) receiver 87 that measures the current position of the vehicle.

[0159] The inclinometer 85 is attached to an appropriate position on the packing box 4, measures the inclination (pitch) of the packing box 4 in a time series manner in the fore-and-aft direction, and outputs measurement data related to the measured inclination to the estimation device 100. The inclinometer 85 may measure the inclination (roll) of the packing box in a time series manner in addition to the inclination (pitch) of the packing box in the fore-and-aft direction.

[0160] The imaging device 86 is installed, for example, so as to capture an image of a range from the top of the front panel 41 diagonally downward and rearward, and captures images of the inside of the packing box 4 in a time-series manner. The imaging device 86 includes a solid-state imaging element such as a CMOS (Complementary Metal Oxide Semiconductor), and outputs digital image data obtained from the solid-state imaging element to the estimation device 100. The imaging device 86 is preferably one capable of acquiring distance information, such as a stereo camera or a distance image sensor.

[0161] The GPS receiver 87 receives radio waves transmitted from a GPS satellite (not shown) and chronologically measures the current position of the specially equipped vehicle 1. The GPS receiver 87 outputs position information relating to the current position of the specially equipped vehicle 1 to the estimation device 100.

[0162] 24 is a flowchart explaining the procedure of the process executed by the estimation device 100 in the seventh embodiment. When the control unit 101 of the estimation device 100 acquires measurement data output from the inclinometer 85 via the input unit 104 (step S701), it judges the state of the packing box 4 based on the acquired measurement data (step S702). At this time, the control unit 101 may judge whether or not the packing box 4 is inclined (lifted up) with respect to the chassis frame 21. In addition, the control unit 101 may calculate the height of the top of the packing box 4 (vehicle height) from the angle of inclination as the state of the packing box 4.

[0163] The control unit 101 notifies the state of the packing box 4 determined in step S702 (step S703). FIG. 25 is a schematic diagram showing an example of the notification of the state of the packing box 4. The control unit 101 can notify the state of the dump truck by outputting characters or figures indicating the state of the dump truck from the output unit 105 to the display device 120. FIG. 25A shows a state in which the display device 120 displays, as character information, that dumping is in progress. FIG. 25B shows a state in which the display device 120 displays, in a schematic manner, that the height of the top of the packing box 4 has reached 4.8 m. Furthermore, the control unit 101 may output an alert when the calculated height of the top of the packing box 4 exceeds a set value.

[0164] When the control unit 101 acquires image data output from the imaging device 86 through the input unit 104 (step S704), it determines the loading state of the packing box 4 based on the acquired image data (step S705). The control unit 101 can determine the loading state by using, for example, a learning model LM2 (see FIG. 26) configured to output information related to the loading state in response to input of image data of the inside of the packing box 4.

[0165] 26 is a schematic diagram for explaining the configuration of the learning model LM2 in the seventh embodiment. The learning model LM2 in the seventh embodiment is, for example, a learning model based on CNN (Convolutional Neural Networks), and includes an input layer, an intermediate layer, and an output layer. The learning model LM2 is trained in advance to output, for example, information regarding the height of a load in response to input of image data obtained by imaging the inside of a packing box 4.

[0166] The input layer receives image data from an imaging device 86 that captures an image of the inside of the packing box 4. The image data input to the input layer is sent to the intermediate layer through the nodes that make up the input layer.

[0167] The intermediate layer is composed of, for example, a convolution layer, a pooling layer, and a fully connected layer. A plurality of convolution layers and pooling layers may be provided alternately. The convolution layer and the pooling layer extract features of an image input through the input layer by calculation using the nodes of each layer. The fully connected layer combines data from which features have been extracted by the convolution layer and the pooling layer into one node, and outputs feature variables transformed by an activation function. The feature variables are output to the output layer through the fully connected layer.

[0168] The output layer includes one or more nodes. Based on the feature variables input from the fully connected layer of the intermediate layer, the output layer converts the feature variables into probabilities using a softmax function, and outputs an estimation result regarding the height of the load. The output form of the estimation result by the output layer is arbitrary. For example, the output layer may be configured with n nodes from the first node to the nth node, and each node constituting the output layer may output a probability regarding the height of the load, such as the probability that the height of the load exceeds the upper limit from the first node, the probability that the height of the load reaches the upper limit from the second node, the probability that the height of the load is 90% of the upper limit from the third node, and the probability that the height of the load is 80% of the upper limit from the fourth node. The number of nodes constituting the output layer and the contents output from each node are not limited to the above, and can be designed as appropriate.

[0169] The estimation device 100 can generate a learning model LM2 as shown in Fig. 26 by collecting a large amount of image data captured by the imaging device 86 and data on the height of the loaded object when the image data was captured (e.g., actual measured value), and learning using the collected image data and height data as training data. Also, instead of a configuration in which the learning model LM2 is generated by the estimation device 100, a configuration in which the learning model LM2 is generated by an external server and the learned learning model LM2 is acquired from the external server may be adopted. The estimation device 100 stores the learning model LM2 generated by the device itself or the learning model LM2 acquired from the external server in the storage unit 102.

[0170] When the control unit 101 of the estimation device 100 acquires image data output from the imaging device 86 in step S704 of the flowchart shown in Fig. 26, the control unit 101 inputs the acquired image data to the learning model LM2 and executes a calculation using the learning model LM2. The control unit 101 judges the loading state (height of the loaded objects in this example) by referring to the calculation result by the learning model LM2. At this time, the control unit 101 can judge the loading state of the shipping box 4 by selecting the state with the highest probability among the probabilities output from each node of the output layer.

[0171] The control unit 101 notifies the loading state of the packing box 4 determined in step S705 (step S706). Fig. 27 is a schematic diagram showing an example of notifying the loading state of the packing box 4. The control unit 101 can notify the loading state by outputting text information indicating the height of the loaded items from the output unit 105 to the display device 120. The example of Fig. 27 shows a state in which text information indicating that the height of the loaded items exceeds an upper limit is displayed on the display device 120.

[0172] As described above, in this embodiment, the state of the specially equipped vehicle 1 can be detected, and the detected state of the specially equipped vehicle 1 can be notified to the occupant.

[0173] In the flowchart shown in Fig. 24, the loading state is determined and notified after the state of the packing box 4 is determined and notified, but the order of execution may be set arbitrarily. Also, only one of the determination and notification of the state of the packing box 4 and the determination and notification of the loading state may be executed.

[0174] In addition, in this embodiment, a configuration has been described in which the state of the specially equipped vehicle 1 is detected by detecting the state of the cargo box 4 and the state of the cargo in the cargo box 4, and the detection results are reported, but the inclination (roll and pitch) of the specially equipped vehicle 1 measured by the inclinometer 82 and an upper limit value for the inclination may also be reported.

[0175] Furthermore, in this embodiment, the status of the specially equipped vehicle 1 is displayed on the display device 120, but it may also be notified to an external management server. At this time, an identifier for identifying the specially equipped vehicle 1 and the position information of the specially equipped vehicle 1 measured by the GPS receiver 87 may be added, and the position information and the status of the specially equipped vehicle 1 may be managed by the management server for each specially equipped vehicle 1. Furthermore, the display device 120 is not limited to being provided in a visible portion of the cargo box 4, and may be a mobile terminal such as a smartphone carried by the worker.

[0176] In addition, in this embodiment, the height of the cargo is estimated using the learning model LM2, but the height of the cargo may also be estimated by analyzing the image obtained by the imaging device 86 and identifying the height position of the cargo within the image.

[0177] The embodiments disclosed herein are illustrative in all respects and should not be considered as limiting. The scope of the present invention is defined by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims.

[0178] For example, in the present embodiment, a dump truck equipped with an estimation device 100 for estimating a load weight, a display device 120 for notifying information related to the load weight estimated by the estimation device 100, and a dump device 3 has been described as an example of the specially equipped vehicle 1, but the present invention is not limited to dump trucks and can be applied to various specially equipped vehicles. For example, the present invention can be applied to specially equipped vehicles such as refuse trucks, mixer trucks, tank trucks, suction vehicles, and container mounting and dismounting vehicles.

[0179] In addition, in the first to seventh embodiments, the specially equipped vehicle 1 is configured to include the estimation device 100, but the estimation device 100 may be a computer provided outside the specially equipped vehicle 1. For example, the computer may be a server device communicably connected to a control device mounted on the specially equipped vehicle 1. In this case, the estimation device 100 may acquire measurement data such as the amount of distortion, hydraulic pressure, tilt angle, and temperature from the control device mounted on the specially equipped vehicle 1 via communication, and estimate the load weight of the specially equipped vehicle 1 by inputting the acquired measurement data into the learning model LM1, etc. [Explanation of symbols]

[0180] 1 Special vehicles 2 Truck chassis 3 Dump device 4 packing boxes 5 Hoist mechanism 20 Cab 21 Chassis frame 22F front wheel 22R rear wheel 23F,23R Axle 30 Subframe 81 Strain Sensor 82 Inclinometer 83 Thermometer 84 Pressure Gauge 85 Inclinometer 86 Imaging Device 87 GPS Receiver 100 Estimator 101 Control section 102 Storage section 103 Operation section 104 Input section 105 Output section 106 Communications Department PG1 Study Program PG2 Estimation Program TB1 Measurement table LM1,LM2 learning models

Claims

1. An acquisition unit that acquires measurement data relating to the magnitude of hydraulic pressure from a pressure gauge that measures the magnitude of hydraulic pressure acting on a hydraulic actuator for a specially equipped vehicle equipped with a hydraulic actuator for lifting and lowering a cargo box; A determination unit that distinguishes and determines a first stop state in which the container is stopped after being raised and a second stop state in which the container is stopped after being lowered; an estimation unit that estimates the load weight in the packing box by selectively using either a first learning model configured to output a calculation result for the load weight when measurement data including a magnitude of hydraulic pressure measured in the first stop state is input according to the stop state identified by the identification unit, or a second learning model configured to output a calculation result for the load weight when measurement data including a magnitude of hydraulic pressure measured in the second stop state is input; a notification unit that notifies information regarding the load weight estimated by the estimation unit; A load weight estimation system for special purpose vehicles.

2. a control unit that controls the operation of the hydraulic actuator so that the container is stopped after being raised when a load weight estimation instruction is given, When the container is stopped after being raised under the control of the control unit, the estimation unit estimates the load weight of the container by inputting the measurement data acquired by the acquisition unit into the first learning model. The system for estimating a load weight of a specially equipped vehicle according to claim 1.

3. The input to the first learning model and the second learning model further includes at least one of an inclination angle, a distortion, and a temperature of the specially equipped vehicle. The system for estimating a load weight of a specially equipped vehicle according to claim 1.

4. A calibration unit that calibrates the estimation result by the estimation unit according to each specially equipped vehicle. The specially equipped vehicle load weight estimation system according to claim 1, comprising:

5. a determining unit for determining a loading state of the load according to the load weight estimated by the estimating unit; Equipped with The notification unit notifies the loading state in a manner corresponding to a determination result of the determination unit. The system for estimating a load weight of a specially equipped vehicle according to claim 1.

6. The notification unit includes a display device provided at a visible portion of a packing box in which the load is loaded, and displays the loading state on the display device in a display mode corresponding to the determination result. The system for estimating the load weight of a specially equipped vehicle according to claim 5.

7. A state detection unit that detects the state of the vehicle Equipped with The notification unit notifies information regarding the state detected by the state detection unit. The system for estimating a load weight of a specially equipped vehicle according to claim 1.

8. The state detected by the state detection unit includes at least one of a state of a packing box in which the load is loaded and a loading state of the load. The system for estimating the load weight of a specially equipped vehicle according to claim 7.

Images