Load calculation device and construction machine including same

The load calculation device uses a posture detection device and machine learning model to predict excavated material weight during excavation, addressing the instability of load pressures and improving loading efficiency by optimizing load management.

WO2026070388A1PCT designated stage Publication Date: 2026-04-02HITACHI CONSTRUCTION MACHINERY CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Filing Date
2025-09-10
Publication Date
2026-04-02

AI Technical Summary

Technical Problem

Existing load calculation methods for construction machinery struggle to accurately determine the weight of excavated material in the bucket during excavation operations due to unstable load pressures caused by ground reaction forces, leading to inefficiencies in loading and production volume management.

Method used

A load calculation device equipped with a posture detection device, load detection device, and a machine learning model that predicts the weight of excavated material based on the detected posture and load values, using a machine learning model trained on time-series data from previous operations.

Benefits of technology

Enables accurate calculation of excavated material weight during both excavation and transport operations, optimizing loading efficiency by preventing overloading or underloading, thereby enhancing production efficiency and reducing the need for rework.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2025032014_02042026_PF_FP_ABST
    Figure JP2025032014_02042026_PF_FP_ABST
Patent Text Reader

Abstract

Provided is a load computation device that computes the weight of an object to be excavated in a bucket in excavation and loading work of a construction machine, said load computation device comprising: a load calculation unit that calculates the weight of the object to be excavated in the bucket on the basis of detection values of the posture and load of a work device, which are detected by a posture detection device and a load detection device; and a load prediction unit that computes a prediction value which the weight of the object to be excavated in the bucket can reach, on the basis of the detection values of the posture and load of the work device, which are detected by the posture detection device and the load detection device, and a machine learning model which has been constructed using, as learning data, prescribed data that indicates a time-series change in an operation state of the construction machine and that has been acquired during the excavation and loading work which has been executed in advance.
Need to check novelty before this filing date? Find Prior Art

Description

Load calculation device and construction machine equipped therewith

[0001] The present invention relates to a load calculation device for calculating the weight (load) of an excavated object in a bucket during excavation and loading operations of a construction machine equipped with a work device including a bucket, and to a construction machine equipped with the same.

[0002] Construction machinery equipped with articulated work devices, such as hydraulic excavators, performs excavation and loading operations in mines and other locations, where materials such as ore and soil are excavated and loaded onto transport machinery such as dump trucks. In this excavation and loading operation, a series of actions are generally performed: "excavation," which involves excavating the target material and scooping it up with the bucket; "transportation," which involves rotating the bucket after the excavation to move it onto the transport machine's bed; and "release," which involves releasing the excavated material held in the bucket onto the transport machine's bed after the transport operation. By repeating this series of actions, the excavated material is loaded onto the transport machine's bed. Here, since the transport machine has a maximum load capacity, it is necessary to avoid overloading the transport machine during excavation and loading operations using construction machinery. On the other hand, insufficient loading of excavated material will lead to a decrease in production volume. Thus, from the standpoint of legal compliance and production efficiency, it is desirable to load the excavated material without exceeding or underloading the transport machine's maximum load capacity.

[0003] In response to such demands, a technology is known for calculating the weight of soil (load) in a bucket during excavation and loading operations of construction machinery (see, for example, Patent Document 1). Specifically, the shovel described in Patent Document 1 calculates the weight of the load using different calculation methods depending on the mode, which is switched according to the operating state of the shovel (e.g., boom raising operation, slewing operation, or other) and the position of the work object, among several modes related to the timing of detecting the weight of the load loaded in the bucket of the work machine (working device). In this technology, one of the methods for calculating the weight of the load uses the balance condition (calculation formula) of the moment (torque) around the rotating part of the work machine, and the load pressure of the hydraulic actuator (boom cylinder or bucket cylinder) that drives the work machine and the detected value (angle) of a physical quantity related to the posture of the work machine are input.

[0004] Japanese Patent Publication No. 2021-31904

[0005] In the technology described in Patent Document 1, as mentioned above, a method is employed to calculate the weight of the load using the load pressure of a hydraulic actuator (boom cylinder or bucket cylinder) as an input value, based on the balance condition of the moment (torque) around the rotating part of the work machine. This calculation method can accurately calculate the weight of the load in the bucket during the transport operation of the excavation and loading work. This is because the load pressure of the hydraulic actuator generated to hold the load in the bucket is stable during the transport operation. On the other hand, during the excavation operation of the excavation and loading work, the load pressure of the hydraulic actuator fluctuates greatly and is unstable due to the influence of ground reaction forces, etc. Therefore, it is difficult to accurately calculate the weight of the load in the bucket during the excavation operation using this calculation method.

[0006] However, there is a demand to know the weight of the material being excavated in the bucket even during the excavation process. Knowing the weight of the material being excavated during the excavation process, that is, before the material is transported and loaded onto the transport machine, would allow for fine-tuning of the excavation work to optimize the amount loaded onto the transport machine. For example, it would help to reduce the number of excavation and loading operations due to insufficient excavation volume, or the need to reload the material due to excessive excavation volume, thereby reducing the efficiency of the excavation and loading operations.

[0007] The present invention is based on the above-mentioned matters and aims to provide a load calculation device capable of calculating the weight of the excavated object in the bucket during either the excavation operation or the transport operation in the excavation and loading work of construction machinery, and a construction machine equipped therewith.

[0008] The present invention includes multiple means for solving the above problems. One example of a solution is a load calculation device for calculating the weight of an excavated object in a bucket during excavation and loading work of a construction machine equipped with a work device including a bucket, a posture detection device for detecting the posture of the work device, and a load detection device for detecting the load acting on the work device, characterized in that the device comprises a load calculation unit that calculates the weight of an excavated object in the bucket based on the detected value of the posture of the work device detected by the posture detection device and the detected value of the load of the work device detected by the load detection device, and a load prediction unit that calculates a predictable value of the weight of an excavated object in the bucket based on the detected value of the posture of the work device detected by the posture detection device and the detected value of the load of the work device detected by the load detection device, and a machine learning model constructed using predetermined data showing the time-series change in the operating state of the construction machine acquired during previously performed excavation and loading work as learning data.

[0009] According to one example of the solution, during the excavation operation in the excavation and loading work, the load prediction unit can calculate a predicted value for the weight of the excavated object in the bucket by using a machine learning model, and during the transport operation, the load calculation unit can directly calculate the weight of the excavated object in the bucket from the detected values ​​of the attitude detection device and the load detection device. In other words, the weight of the excavated object in the bucket can be calculated at either the excavation operation or the transport operation in the excavation and loading work of the construction machine. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

[0010] Figure 1 is an external view showing a hydraulic excavator, which is an example of a construction machine equipped with a load calculation device according to an embodiment of the present invention. Figure 1 is a schematic perspective view showing the equipment arranged inside the cab of the construction machine shown in Figure 1. Figure 1 is a hydraulic circuit showing the schematic configuration of the hydraulic system equipped in the construction machine shown in Figure 1. Figure 4 is a block diagram showing the hardware and functional configuration of the load calculation device according to the first embodiment of the present invention. Figure 4 is a flowchart showing an example of the processing procedure of the display generation unit in the load calculation device according to the first embodiment shown in Figure 4. Figure 1 shows the state of the hydraulic excavator immediately before the start of excavation and loading work and the display screen of the display device according to the output result of the load calculation device according to the first embodiment for the working state of the hydraulic excavator. Figure 2 shows the state of the hydraulic excavator immediately after the start of excavation operation in excavation and loading work and the display screen of the display device according to the output result of the load calculation device according to the first embodiment for the working state of the hydraulic excavator. Figure 3 shows the state of the excavation operation in the middle of the excavation operation in excavation and loading work of the hydraulic excavator and the display screen of the display device according to the output result of the load calculation device according to the first embodiment for the working state of the hydraulic excavator. Figure 10 shows a diagram illustrating the state immediately before the completion of the excavation operation or the start of the transport operation in the excavation and loading work of a hydraulic excavator, and a diagram showing the display screen of a display device corresponding to the output result of the load calculation device according to the first embodiment for the working state of the hydraulic excavator. Figure 10 is a block diagram showing the hardware and functional configuration of the load calculation device according to the second embodiment of the present invention. Figure 10 is a flowchart showing an example of the processing procedure of the operation determination unit in the load calculation device according to the second embodiment shown. Figure 10 is a flowchart showing an example of the processing procedure of the total load calculation unit in the load calculation device according to the second embodiment shown. Figure 10 is a flowchart showing an example of the processing procedure of the display generation unit in the load calculation device according to the second embodiment shown. Figure 10 shows a diagram illustrating the state immediately before the completion of the transport operation in the excavation and loading work of a hydraulic excavator, and a diagram showing the display screen of a display device corresponding to the output result of the load calculation device according to the second embodiment for the working state of the hydraulic excavator. Figure 10 shows a diagram illustrating the state of the release operation in the excavation and loading work of a hydraulic excavator, and a diagram showing the display screen of a display device corresponding to the output result of the load calculation device according to the second embodiment for the working state of the hydraulic excavator.This is a block diagram showing the hardware and functional configuration of a load calculation device according to a third embodiment of the present invention. This is a flowchart showing an example of the processing procedure of the addition value selection unit in the load calculation device according to the third embodiment shown in Figure 16. This is a flowchart showing an example of the processing procedure of the total load calculation unit in the load calculation device according to the third embodiment shown in Figure 16. This is a diagram showing the state immediately before the completion of the transport operation in the excavation and loading work of a hydraulic excavator, and the display screen of the display device according to the output result of the load calculation device according to the third embodiment for the working state of the hydraulic excavator. This is a diagram showing the state of the release operation in the excavation and loading work of a hydraulic excavator, and the display screen of the display device according to the output result of the load calculation device according to the third embodiment for the working state of the hydraulic excavator. This is a block diagram showing the hardware and functional configuration of a load calculation device according to a fourth embodiment of the present invention. This is a flowchart showing an example of the processing procedure of the relearning execution unit in the load calculation device according to the fourth embodiment shown in Figure 21. This is a flowchart showing an example of the processing procedure of the display generation unit in the load calculation device according to the fourth embodiment shown in Figure 21. This is a diagram showing the state after the completion of the excavation operation in the excavation and loading work of a hydraulic excavator, and the display screen of the display device according to the output result of the load calculation device according to the fourth embodiment for the working state of the hydraulic excavator. This figure shows a non-operational state of a hydraulic excavator in standby mode, and a figure showing the display screen of a display device corresponding to the output result of the load calculation device according to the fourth embodiment for the working state of the hydraulic excavator.

[0011] Hereinafter, embodiments of the load calculation device of the present invention will be described with reference to the drawings. The load calculation device according to this embodiment calculates the weight of the excavated object to be excavated and loaded onto a transport machine during the excavation and loading work of construction machinery. In this embodiment, a hydraulic excavator will be used as an example of a construction machine that performs excavation and loading work.

[0012] First, the general configuration of a hydraulic excavator, as an example of construction machinery, will be explained using Figures 1 and 2. Figure 1 is an external view showing a hydraulic excavator, which is an example of construction machinery equipped with a load calculation device according to the first embodiment. Figure 2 is a schematic perspective view showing the equipment arranged inside the cab of the construction machinery shown in Figure 1. Here, the explanation will be based on the perspective of a worker seated in the driver's seat.

[0013] In Figure 1, the hydraulic excavator 30, as a construction machine, is equipped with a work device 31 for performing excavation and loading operations, and a machine body 32 on which the work device 31 is rotatably mounted. The machine body 32 consists of a self-propelled traveling body 33 and a rotating body 34 that is rotatably mounted on the traveling body 33.

[0014] The work device 31 is a multi-jointed work device composed of multiple link members connected so as to be rotatable perpendicular to the rotation surface of the slewing body 34 for performing excavation and loading operations. The multiple link members consist of, for example, a boom 41, an arm 42, and a bucket 43 as a work tool. The base end of the boom 41 is rotatably supported at the front of the slewing body 34 via a boom pin. The base end of the arm 42 is rotatably supported at the tip of the boom 41 via an arm pin. The bucket 43 is rotatably supported at the tip of the arm 42 via a bucket pin. The boom 41, arm 42, and bucket 43 are driven by hydraulic actuators, the boom cylinder 62, arm cylinder 63, and bucket cylinder 64, respectively.

[0015] The vehicle 33 is equipped with, for example, crawler-type travel devices 45 on both the left and right sides (only the left side is shown). The travel devices 45 are driven by a hydraulic travel hydraulic motor (not shown), which is a hydraulic actuator.

[0016] The slewing body 34 is configured to be driven to rotate relative to the traveling body 33 by, for example, a slewing hydraulic motor 65 (see Figure 3, described later), which is a hydraulic actuator. The slewing body 34 includes a driver's cab 47 in which an operator sits, a building 48 for housing various equipment, and a counterweight 49 attached to the rear end of the building 48.

[0017] As shown in Figure 2, the operator's cab 47 is equipped with an operator's seat 51 where the worker sits, an operating device 52 for operating the hydraulic excavator 30, and a display device 53 for displaying various information such as the operating status of the hydraulic excavator 30. The operating device 52 consists of, for example, a right operating lever 52a for operating the boom 41 and bucket 43 of the work device 31, a left operating lever 52b for operating the arm 42 and slewing body 34 of the work device 31, and a travel operating lever 52c for operating the travel device 45. The right operating lever 52a and the left operating lever 52b are, for example, electric levers, and output an electrical signal corresponding to the amount and direction of lever operation to the control device 80 (see Figure 3 below), which will be described later. The display device 53 is configured to display a screen according to the output (calculation result) of the load calculation device 1 (see Figure 4 below), which will be described later according to this embodiment.

[0018] The building 48 houses various hydraulic equipment for operating the work device 31 and the machine body 32 (traveling body 33 and rotating body 34). The counterweight 49 is used to balance the weight with the work device 31.

[0019] The working device 31 is equipped with a posture detection device 55 that detects the posture of the working device 31, or in this embodiment, physical quantities related to the posture of the working device 31. The posture detection device 55 is composed of, for example, a first posture sensor 55a, a second posture sensor 55b, and a third posture sensor 55c that detect physical quantities related to the posture of the boom 41, arm 42, and bucket 43 that constitute the working device 31, respectively. The first posture sensor 55a, the second posture sensor 55b, and the third posture sensor 55c are, for example, a boom angle sensor attached to the boom pin, an arm angle sensor attached to the arm pin, and a bucket angle sensor attached to the bucket pin, respectively. The posture detection device 55 (55a, 55b, 55c) outputs sensor signals corresponding to the detected values ​​to the load calculation device 1 according to the first embodiment described later (see Figure 4 described later). The attitude detection device can be replaced with, for example, an inertial measurement unit (IMU) capable of detecting the inclination (angle) of the boom 41, arm 42, and bucket 43 with respect to a reference plane (e.g., a horizontal plane).

[0020] The rotating body 34 (vehicle body) is equipped with an attitude detection device 56 that detects physical quantities related to the attitude of the rotating body 34 (vehicle body). The attitude detection device 56 is, for example, an inertial measurement unit (IMU) capable of detecting the tilt (angle) and angular velocity of the rotating body 34 with respect to a reference plane (e.g., a horizontal plane), or a rotation angle sensor capable of detecting the rotation angle.

[0021] Next, the configuration of the hydraulic system of a construction machine equipped with a load calculation device according to the first embodiment will be described with reference to Figure 3. Figure 3 is a hydraulic circuit showing the schematic configuration of the hydraulic system of the construction machine shown in Figure 1. In Figure 3, solid lines indicate the lines of the main hydraulic circuit, and dashed lines indicate the lines of the pilot hydraulic circuit.

[0022] In Figure 3, the hydraulic excavator 30 is equipped with a hydraulic system 60 that drives the working device 31, the traveling body 33, and the slewing body 34 (see Figure 1 for both) by hydraulic pressure. The hydraulic system 60 includes a hydraulic pump 61 that is driven by a prime mover (not shown) to discharge pressurized oil, and a plurality of hydraulic actuators 62, 63, 64, 65 that are driven by the pressurized oil discharged from the hydraulic pump 61. Figure 3 shows an example of a hydraulic circuit configuration with only one hydraulic pump 61. However, the following discussion also holds true for hydraulic circuits with two or more hydraulic pumps. Also, in Figure 3, for the sake of simplicity, only the hydraulic circuits related to the boom cylinder 62, arm cylinder 63, and bucket cylinder 64 that drive the boom 41, arm 42, and bucket 43, respectively, which are components of the working device 31, and the slewing hydraulic motor 65 that rotates the slewing body 34 are shown, and the hydraulic circuit related to the traveling hydraulic motor that moves the traveling body 33 is omitted.

[0023] The boom cylinder 62, arm cylinder 63, bucket cylinder 64, and slewing hydraulic motor 65 are connected to the hydraulic pump 61 via first control valve 66, second control valve 67, third control valve 68, and fourth control valve 69, respectively. Each of the control valves 66, 67, 68, and 69 is connected to the hydraulic pump 61 via a discharge line 71. The first control valve 66, second control valve 67, third control valve 68, and fourth control valve 69 are connected to the boom cylinder 62, arm cylinder 63, bucket cylinder 64, and slewing hydraulic motor 65 via a pair of first actuator lines 72a, 72b (72), a pair of second actuator lines 73a, 73b (73), a pair of third actuator lines 74a, 74b (74), and a pair of fourth actuator lines 75a, 75b (75), respectively.

[0024] Each control valve 66, 67, 68, and 69 is, for example, a hydraulic pilot-operated spool valve, and has pressure-receiving parts on both ends of the spool valve to which the operating pilot pressure acts. Each control valve 66, 67, 68, and 69 is configured to change position (stroke) according to the magnitude of the operating pilot pressure.

[0025] The hydraulic system 60 further includes a pilot pump 77 that supplies the source pressure for generating the operating pilot pressure, and a solenoid valve unit 78 that reduces the hydraulic pressure supplied from the pilot pump 77 to generate the operating pilot pressure for a plurality of control valves 66, 67, 68, and 69. The solenoid valve unit 78 generates the operating pilot pressure for each of the control valves 66, 67, 68, and 69 in response to a command from the control device 80.

[0026] The hydraulic system 60 is provided with a load detection device that detects the load acting on the work device 31, which in this embodiment is the load pressure of the hydraulic actuator that drives the work device 31. The load detection device includes, for example, a first load pressure sensor 57 for detecting the load pressure of the boom cylinder 62, a second load pressure sensor 58 for detecting the load pressure of the arm cylinder 63, and a third load pressure sensor 59 for detecting the load pressure of the bucket cylinder 64. Specifically, a first bottom pressure sensor 57a for detecting the pressure in the bottom oil chamber B (hereinafter referred to as the bottom chamber) of the boom cylinder 62 and a first rod pressure sensor 57b for detecting the pressure in the rod oil chamber R (hereinafter referred to as the rod chamber) of the boom cylinder 62 are installed on a pair of first actuator lines 72a and 72b, respectively. A second bottom pressure sensor 58a for detecting the pressure in the bottom chamber B of the arm cylinder 63 and a second rod pressure sensor 58b for detecting the pressure in the rod chamber R of the arm cylinder 63 are installed on a pair of second actuator lines 73a and 73b, respectively. A third bottom pressure sensor 59a for detecting the pressure in the bottom chamber B of the bucket cylinder 64 and a third rod pressure sensor 59b for detecting the pressure in the rod chamber R of the bucket cylinder 64 are installed on a pair of third actuator lines 74a and 74b, respectively. Each of the pressure sensors 57a, 57b, 58a, 58b, 59a, and 59b outputs a sensor signal corresponding to the detected pressure value to the load calculation device 1 according to the first embodiment described later (see Figure 4 described later).

[0027] The control device 80 controls the solenoid valve unit 78 in response to electrical signals (operation signals) from the right operation lever 52a and the left operation lever 52b of the operating device 52. The control device 80 can, for example, include the calculation function of the load calculation device 1 according to the first embodiment.

[0028] In the construction machine 30 configured in this way, when the operating device 52 is operated by the operator of the construction machine 30, the control device 80 controls the opening amount of the solenoid valve unit 78 according to the operating direction and amount of the operating device 52. As a result, the solenoid valve unit 78 reduces the discharge pressure of the pilot pump 77 and generates an operating pilot pressure. In response to the operating pilot pressure generated by the solenoid valve unit 78, the first control valve 66, second control valve 67, third control valve 68, and fourth control valve 69 are driven, and pressurized oil from the hydraulic pump 61 is supplied to the boom cylinder 62, arm cylinder 63, and bucket cylinder 64. As the boom cylinder 62, arm cylinder 63, and bucket cylinder 64 extend and retract due to the pressurized oil from the hydraulic pump 61, the boom 41, arm 42, and bucket 43 rotate, respectively, and the posture of the work device 31 changes. Excavation and loading operations are performed by changing the posture of the work device 31.

[0029] The excavation and loading operation of the construction machine 30 generally consists of four operations in one cycle: an "excavation operation" in which the bucket 43 of the work device 31 excavates the target to be excavated and holds the excavated material in the bucket 43; a "transportation operation" in which the bucket 43 is rotated after the excavation operation and moved onto the loading platform of the transport machine; and a "discharge operation" in which the excavated material in the bucket 43 is released onto the loading platform of the transport machine after the transport operation. By repeating this series of three operations, the loading platform of the transport machine is filled with excavated material. The transport machine has a maximum load capacity set. Therefore, in excavation and loading operations using construction machines, it is necessary to suppress the decrease in production volume due to insufficient or overloading of excavated material onto the transport machine and the need for reloading. In other words, from the viewpoint of production efficiency, it is desirable to load the excavated material without any excess or deficiency relative to the maximum load capacity of the transport machine.

[0030] Therefore, the construction machine 30 is equipped with, for example, a load calculation device 1 (see Figure 4) that calculates the weight of the object to be excavated in the bucket 43.

[0031] [First Embodiment] Next, the hardware and functional configuration of the load calculation device according to the first embodiment will be described with reference to Figure 4. Figure 4 is a block diagram showing the hardware and functional configuration of the load calculation device according to the first embodiment.

[0032] In Figure 4, the load calculation device 1 accurately calculates the weight of the excavated object in the bucket 43 during both the excavation and transporting phases of the excavation and loading operation of the construction machine 30. The load calculation device 1 is electrically or communicatively connected to at least the posture detection device 55 (boom angle sensor 55a, arm angle sensor 55b, bucket angle sensor 55c) and the first load pressure sensor 57 (first bottom pressure sensor 57a, first rod pressure sensor 57b), and is configured to receive detection signals from the posture detection device 55 (physical quantities related to the posture of the work device 31) and detection signals from the first load pressure sensor 57 (load pressure of the boom cylinder 62). The load calculation device 1 uses the detection signals (sensor signals) from the posture detection device 55 and the first load pressure sensor 57 to calculate the weight of the excavated object in the bucket 43. The load calculation device 1 is electrically or communicatively connected to the display device 53, and generates display data to display the calculated weight of the excavated object in the bucket 43 and outputs it to the display device 53.

[0033] The hardware configuration of the load calculation device 1 includes, for example, a sensor signal input device 2 that receives sensor signals from a posture detection device 55 and a first load pressure sensor 57, a storage device 3 consisting of RAM or ROM, and a microcomputer including a processing device 4 consisting of a CPU or MPU. The sensor signal input device 2 converts the received sensor signals into signals that the processing device 4 can process. The storage device 3 has programs and various information stored in advance that are necessary to calculate the weight of the excavated object in the bucket 43 during the process from the excavation operation to the release operation in the excavation and loading operation of the construction machine 30. The processing device 4 reads programs and various information from the storage device 3 as appropriate and executes processing according to the programs to realize the various functions described later. The functional configuration of the load calculation device 1 includes, for example, a load calculation unit 11, a load prediction unit 12, and a display generation unit 13.

[0034] The load calculation unit 11 calculates the weight of the excavation object held in the bucket 43 by using a calculation formula derived from a predetermined equation with the detection values (sensor signals) of the attitude detection device 55 received by the sensor signal input unit 2 and the detection values (sensor signals) of the first load pressure sensor 57 as input values. The calculation result of the load calculation unit 11 (hereinafter sometimes referred to as a calculated value) is output to the display generation unit 13.

[0035] The predetermined calculation formula used by the load calculation unit 11 is derived as follows, for example. The dynamics of the hydraulic excavator 30 in which the articulated working device 31 composed of three link members, the boom 41, the arm 42, and the bucket 43, is attached to the revolving body 34 is expressed by the following formula (1).

[0036]

[0037] Here, τ, M, h, and g are the torque, inertia term, centrifugal force and Coriolis force term, and gravity term, respectively. Also, q, q dot, and q two dots are the position vector, velocity vector, and acceleration vector, respectively.

[0038] Specifically, formula (1) is expressed as formula (2). Here, θ is an angle, and θ two dots is an angular acceleration. The subscripts 1, 2, 3, and 4 of M, θ two dots, h, and g correspond to the revolving body 34, the boom 41, the arm 42, and the bucket 43, respectively.

[0039]

[0040] Among the formulas of formula (2), the formula of τ in the second row boom shows the balance formula of the moment around the rotation axis (boom pin) of the boom 41. From the formula of τ in this second row boom of the formula, the total weight m bucket obtained by adding the weight m payload of the bucket itself and the weight m 4 of the excavation object held in the bucket 43 can be derived. Therefore, the calculation formula for the weight m payload of the excavation object held in the bucket 43 can be obtained from formula (2). τ boom(Torque of boom 41) can be calculated from the driving force (support force) of boom cylinder 62 using the detected values of the first bottom pressure sensor 57a and the first rod pressure sensor 57b (the first load pressure sensor 57). Since M, h, and g are functions of the angle (the posture of the working device 31 and the slewing body 34) and the angular velocity, they can be calculated from the detected values of the posture detection devices 55 and 56. The load calculation unit 11 uses such a predetermined calculation formula to calculate the weight m of the excavation object in the bucket 43 payload Thereby. For the calculation of the inertial term and the gravitational term, specifications such as the shapes of the boom 41, the arm 42, the bucket 43, and the slewing body 34 stored in the storage device 3 in advance are used.

[0041] However, during the excavation operation of the excavation / loading work, the load calculation unit 11 has difficulty calculating the weight m of the excavation object in the bucket 43 payload with high accuracy. This is because in the predetermined calculation formula used by the load calculation unit 11, the detected values of the first bottom pressure sensor 57a and the first rod pressure sensor 57b are used as input values. The load pressure of the boom cylinder 62, which is the detected value of the first bottom pressure sensor 7a and the first rod pressure sensor 7b, is stable during the transportation operation in the excavation / loading work without large fluctuations, but large fluctuations occur during the excavation operation due to the influence of the excavation reaction force and the like. For this reason, in the calculation formula of the load calculation unit 11 that uses the detected values of the first bottom pressure sensor 57a and the first rod pressure sensor 57b as input values, it is difficult to calculate the weight m of the excavation object in the bucket 43 payload with high accuracy during the excavation operation. Therefore, the load calculation unit 11 outputs the calculation result only when the load determination condition is satisfied, such as when the load pressure of the boom cylinder 62 is in a stable state (the fluctuation of the load pressure is small and within a predetermined range). In other words, the load calculation unit 11 does not output the calculation result when the load determination condition, such as the load pressure of the boom cylinder 62 being in a fluctuating state (unstable state), is not satisfied. The predetermined calculation formula used by the load calculation unit 11 is not limited to this calculation method, and a known method capable of calculating the weight of the excavation object in the bucket 43 during the transportation operation can be used.

[0042] The load prediction unit 12 calculates an achievable value (hereinafter sometimes referred to as the predicted value) for the weight of the excavated object in the bucket 43 during the actual excavation and loading operation, based on a machine learning model constructed using predetermined data showing the time-series changes in the operating state of the construction machine 30 acquired during previously performed excavation and loading operations as training data. The calculation result of the load prediction unit 12 (e.g., the predicted value) is output to the display generation unit 13. As training data for the machine learning model, for example, time-series data of physical quantities relating to the posture of the work device and time-series data of the load pressure (bottom pressure and rod pressure) of the boom cylinder are used as input data, and the weight of the excavated object held in the bucket during the excavation and loading operation is used as training data. Furthermore, the load prediction unit 12 performs its prediction calculations by inputting, for example, time-series data (sensor signals received by the sensor signal input unit 2) of the detected values ​​(physical quantities related to the posture of the work device 31) from the posture detection device 55 during actual excavation and loading work, and time-series data (sensor signals received by the sensor signal input unit 2) of the detected values ​​(bottom pressure and rod pressure of the boom cylinder 62) from the first load pressure sensor 57, to a machine learning model.

[0043] The machine learning model used by the load prediction unit 12 is constructed using deep learning with a neural network that mimics the structure of the human brain. The neural network consists of three layers: an input layer, a hidden layer, and an output layer. The input layer receives, for example, time-series data of physical quantities (e.g., angle, angular velocity, angular acceleration) related to the posture of the work equipment in previously performed excavation and loading operations, and time-series data of the load pressure (bottom pressure and rod pressure) of the hydraulic actuator that drives the work equipment. The hidden layer inherits data from the input layer and generates feature quantities as output values. Here, for example, a recurrent neural network (RNN), a deep learning algorithm suitable for handling time-series data, is used. The output layer adjusts the output values ​​from the hidden layer to produce a one-dimensional output value. In the machine learning model used by the load prediction unit 12, the gradient of the intermediate layer with respect to the time-series input data is tuned so that the one-dimensional output value of the output layer (i.e., the predicted value of the weight of the excavated object to be scooped into the bucket 43) matches the training data (the weight of the excavated object held in the bucket by previously performed excavation and loading operations).

[0044] The load prediction unit 12 uses the machine learning model constructed as described above to predict the weight m of the excavated object to be scooped into the bucket 43, based on the change in the posture of the work device 31 and the change in the load pressure of the boom cylinder 62 during the excavation operation (time-series data detected by the posture detection device 55 and the first load pressure sensor 57) during the excavation operation. payload This makes it possible to predict with high accuracy. However, since the calculations performed by the load prediction unit 12 using a machine learning model require time-series data from the attitude detection device 55 and the first load pressure sensor 57, the calculation results of the load prediction unit 12 are not output until the excavation operation has progressed to a certain extent.

[0045] The display generation unit 13 generates display data for display on the display device 53 based on the output (calculation result) from the load calculation unit 11 or the load prediction unit 12. In this embodiment, the display generation unit 13 generates a display screen (screen layout and display image including the calculation result) as the display data for displaying the calculation result on the display device 53. The display screen as the display data generated by the display generation unit 13 is output to the display device 53. The display device 53 displays the display screen (display data) generated by the display generation unit 13.

[0046] Next, the processing procedure of the display generation unit in the load calculation device according to the first embodiment will be explained with reference to Figure 5. Figure 5 is a flowchart showing an example of the processing procedure of the display generation unit in the load calculation device according to the first embodiment shown in Figure 4.

[0047] In Figure 5, the display generation unit 13 of the load calculation device 1 shown in Figure 4 detects a signal from the load calculation unit 11 and determines whether or not there is an output (calculation result or calculated value) from the load calculation unit 11 (step S10). If there is no output from the load calculation unit 11 (in the case of NO), the process proceeds to step S20, while if there is an output from the load calculation unit 11 (in the case of YES), the process proceeds to step S50. As described above, the load calculation unit 11 performs calculations and outputs the calculation results when the load pressure of the boom cylinder 62 is in a stable state, but does not perform calculations and does not output calculation results when the load pressure is in a fluctuating state.

[0048] If the answer in step S10 is NO, the display generation unit 13 detects the signal from the load prediction unit 12 and determines whether or not there is an output (prediction result or predicted value) from the load prediction unit 12 (step S20). If there is no output from the load prediction unit 12 (if NO), the process proceeds to step S30, while if there is an output from the load prediction unit 12 (if YES), the process proceeds to step S40. As described above, the load prediction unit 12 outputs the prediction result after the excavation operation has progressed to a certain extent (for example, a few seconds to more than ten seconds after the start of the excavation operation).

[0049] If the answer in step S20 is NO, the display generation unit 13 generates a default display screen (display data) and outputs it to the display device 53 (step S30). With this, the display generation unit 13 terminates the display screen generation process (end).

[0050] If the answer in step S20 is YES, the display generation unit 13 generates a display screen (display data) that adds the words "payload prediction value" to the "numerical value" of the calculation result of the load prediction unit 12 and outputs it to the display device 53 (step S40). With this, the display generation unit 13 finishes the process of generating the display screen (end).

[0051] Furthermore, if the answer in step S10 is YES, the display generation unit 13 generates a display screen (display data) with the words "payload value" added to the "numerical value" of the calculation result from the load calculation unit 11, and outputs it to the display device 53 (step S50). With this, the display generation unit 13 completes the display screen (display data) generation process (end).

[0052] Next, the display screen of the display device corresponding to the output result of the load calculation device according to the first embodiment for excavation and loading operations of construction machinery will be explained using Figures 6 to 9. The upper figures of Figures 6 to 9 show the flow of the excavation and loading operation from just before the start of the work to just after the start of the excavation operation, the intermediate state of the excavation operation, and the state just before the start of the transport operation after the completion of the excavation operation. The lower figures of Figures 6 to 9 show the display screen of the display device corresponding to the output result of the load calculation device according to the first embodiment for the working state of the construction machinery.

[0053] The upper part of Figure 6 shows the hydraulic excavator 30 just before the start of the excavation and loading operation. More specifically, it shows the state just before the start of the excavation operation, with the bucket 43 brought close to the excavation surface 101 in order to perform the bench cut method.

[0054] At this time (before the start of excavation and loading work), the load calculation unit 11 and load prediction unit 12 of the load calculation device 1 do not output calculation results, so the load calculation device 1 (display generation unit 13) generates a default display screen corresponding to step S30 of the flowchart shown in Figure 5 and outputs it to the display device 53. As a result, the display device 53 displays a display screen (default display screen) that includes the string "payload value" and the numerical value "0t", as shown in the lower part of Figure 6. Since it is before the start of excavation and loading work, the weight of the excavation target held in the bucket 43 is set to "0t" by default.

[0055] The upper diagram of Figure 7 shows the transition from the state immediately before the start of the excavation and loading work shown in the upper diagram of Figure 6 to the state immediately after the start of the excavation operation. At this time, since the load calculation unit 11 and the load prediction unit 12 do not output calculation results, the load calculation device 1 (display generation unit 13) generates a default display screen corresponding to step S30 of the flowchart shown in Figure 5 and outputs it to the display device 53. However, since the time-series data of the detected values ​​from the attitude detection device 55 and the first load pressure sensor 57 have been input to the load prediction unit 12 due to the start of the excavation operation, the calculation of the load prediction unit 12 using a machine learning model is in progress. Therefore, as the default display screen, a display screen is generated in which "0t" in the display screen of the lower diagram of Figure 6 is changed to the string "Calculating" indicating that the load prediction unit 12 is processing, and a "Graph" showing the calculation progress. As a result, the display device 53 displays a display screen that includes the strings "Payload value" and "Calculating" in addition to a "Graph" showing the calculation progress, as shown in the lower diagram of Figure 7.

[0056] The upper diagram in Figure 8 shows the transition from the state immediately after the start of the excavation operation, as shown in the upper diagram in Figure 7, to an intermediate state of the excavation operation. Specifically, the hydraulic excavator 30 is in an intermediate state of excavation operation, with the bucket 43 positioned about halfway up the excavation surface 101.

[0057] During this time (during excavation), the load pressure of the boom cylinder 62 is in a fluctuating (unstable) state, so the calculation result of the load calculation unit 11 is not output. On the other hand, time-series data of the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 during the excavation operation, from the upper diagram of Figure 7 to the upper diagram of Figure 8, are input to the load prediction unit 12, and the prediction result using the machine learning model of the load prediction unit 12 is output. For this reason, the load calculation device 1 (display generation unit 13) generates a display screen showing the predicted value of the payload, which is the prediction result of the load prediction unit 12, in accordance with step S40 of the flowchart shown in Figure 5, and outputs it to the display device 53. As a result, the display device 53 displays a display screen that includes the string "Predicted Payload Value" and the "Value (for example, 30t)" of the prediction result of the load prediction unit 12, as shown in the lower diagram of Figure 8. The prediction result from the load prediction unit 12 is the predicted value of the load that can be reached when scooping up the excavated object into the bucket 43, assuming that the series of excavation operations shown from the upper diagram of Figure 7 to the upper diagram of Figure 8 continue without change, and is displayed on the display device 53 as the "payload prediction value".

[0058] The upper diagram of Figure 9 shows the transition from the intermediate state of the excavation operation shown in the upper diagram of Figure 8 to the state immediately before the start of the transport operation after the excavation operation has been completed. Specifically, the hydraulic excavator 30 has finished the excavation operation using the bench cut method and has lifted the bucket 43 away from the excavation surface 101.

[0059] At this time (when the excavation operation is completed and the bucket 43 is in the air), the load pressure of the boom cylinder 62 stabilizes, and the calculation result of the load calculation unit 11 is output. For this reason, the load calculation device 1 (display generation unit 13) generates a display screen showing the payload calculation value, which is the calculation result of the load calculation unit 11, according to step S50 of the flowchart shown in Figure 5, and outputs it to the display device 53. As a result, the display device 53 displays a display screen that includes the string "payload value" and the "value (for example, 31t)" of the calculation result of the load calculation unit 11, as shown in the lower part of Figure 9. The calculation result of the load calculation unit 11 is calculated using a predetermined calculation formula that shows the balance of the moment of the rotating part of the boom 41, when the hydraulic excavator 30 is in the posture shown in the upper part of Figure 9, and is more accurate than the calculation result predicted by the load prediction unit 12 based on the time-series data during the excavation operation. For this reason, the calculated value of the load calculation unit 11 is displayed on the display device 53 as the "payload value" with priority over the predicted value of the load prediction unit 12.

[0060] As described above, the load calculation device 1 according to this embodiment calculates the weight of the excavated object in the bucket 43 during both the excavation and transport operations of the construction machine 30, and displays the calculation result on the display device 53. When the excavation operation is in progress before completion, the load calculation device 1 (load prediction unit 12) inputs time-series data detected during the excavation operation by the attitude detection device 55 and the first load pressure sensor 57 to the above-mentioned machine learning model, thereby calculating the weight of the excavated object in the bucket 43 m payload This predicts the load that can be reached in the bucket 43, assuming that the series of excavation operations continue as is. payload Highly accurate prediction values ​​can be obtained for this purpose. By displaying these prediction values ​​on the display device 53, the operator of the construction machine 30 can be provided with information to determine whether the amount of material to be excavated on the transport machine is excessive or insufficient during the excavation operation.

[0061] Furthermore, when the excavation operation is completed and the bucket 43 is separated from the excavation surface 101, the load calculation device 1 (load calculation unit 11) inputs the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 into the predetermined calculation formula described above, thereby calculating the weight m of the excavated object in the bucket 43. payload The following calculation is performed. When the excavation operation is completed, the load pressure (bottom pressure and rod pressure) of the boom cylinder 62 detected by the first load pressure sensor 57 is in a stable state, so the weight m of the excavated object in the bucket 43 is calculated using a predetermined calculation formula. payload This can be calculated with high accuracy. By displaying the calculated value on the display device 53, the operator of the construction machine 30 can be provided with information to determine whether the amount of excavated material loaded onto the transport machine is excessive or insufficient during transport operations.

[0062] As described above, the load calculation device 1 according to the first embodiment calculates the weight of the excavated object in the bucket 43 during excavation and loading operations of a construction machine 30 which is equipped with a work device 31 including a bucket 43, a posture detection device 55 for detecting the posture of the work device 31, and a first load pressure sensor 57 as a load detection device for detecting the load acting on the work device 31. The load calculation device 1 comprises a load calculation unit 11 that calculates the weight of the excavated object in the bucket 43 based on the detected posture value of the work device 31 detected by the posture detection device 55 and the detected load value of the work device 31 detected by the first load pressure sensor 57 (load detection device), and a load prediction unit 12 that calculates a predicted value of the weight of the excavated object in the bucket 43 that can be reached, based on the detected posture value of the work device 31 detected by the posture detection device 55 and the detected load value of the work device 31 detected by the first load pressure sensor 57 (load detection device), and a machine learning model constructed using predetermined data showing the time-series changes in the operating state of the construction machine 30 acquired during previously performed excavation and loading work as learning data.

[0063] With this configuration, during the excavation operation in the excavation and loading work, the load prediction unit 12 can use a machine learning model to calculate a predicted value for the weight of the excavated object in the bucket 43, and during the transport operation, the load calculation unit 11 can directly calculate the weight of the excavated object in the bucket 43 from the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 (load detection device). In other words, the weight of the excavated object in the bucket 43 can be calculated at any stage of the excavation and transport operation in the excavation and loading work of the construction machine 30.

[0064] Furthermore, the load calculation device 1 according to this embodiment further includes a display generation unit 13 that generates display data for displaying the calculated weight of the excavation target from the load calculation unit 11 or the predicted weight of the excavation target from the load prediction unit 12 on the display device 53.

[0065] With this configuration, the load calculation device 1 outputs display data of the calculated value from the load calculation unit 11 or the predicted value from the load prediction unit 12 to the display device 53, so that in either the excavation operation or the transport operation in the excavation and loading work of the construction machine 30, the weight m of the excavated object in the bucket 43 is determined. payload The calculation results of the load calculation device 1 can be displayed on the display device 53. This provides the operator of the construction machine 30 with information to determine whether the amount of excavated material loaded onto the transport machine is excessive or insufficient during the excavation and transport operations. As a result, the operator can adjust the operation by the end of the excavation based on the predicted weight of the excavated material in the bucket 43 provided during the excavation, thereby enabling them to excavate the target weight.

[0066] Furthermore, in the load calculation device 1 according to this embodiment, if both the calculated value from the load calculation unit 11 and the predicted value from the load prediction unit 12 are available, the display generation unit 13 generates display data for displaying the calculated value from the load calculation unit 11 and outputs it to the display device 53.

[0067] With this configuration, if both the calculated value from the load calculation unit 11 and the predicted value from the load prediction unit 12 are available, the load calculation device 1 prioritizes outputting the display data of the calculated value from the load calculation unit 11 to the display device 53, thus enabling relatively high-precision calculation of the weight m of the excavated object in the bucket 43. payload By displaying this information on the display device 53, more accurate information can be provided to the operator of the construction machine 30.

[0068] Furthermore, in this embodiment, the load detection device (first load pressure sensor 57) is configured to detect the load pressure of the boom cylinder 62, which is a hydraulic actuator that drives the work device 31, as a physical quantity of the load acting on the work device 31. The machine learning model of the load prediction unit 12 of the load calculation device 1 is constructed by using time-series data of physical quantities related to the posture of the work device and time-series data of the load pressure of the boom cylinder (hydraulic actuator) acquired during previously performed excavation and loading work as input data, and by using the weight of the excavated object in the bucket acquired during the said previously performed excavation and loading work as training data. The load calculation unit 11 calculates the above-mentioned weight by inputting the detected value of the posture detection device 55 and the detected value of the first load pressure sensor 57 (load detection device) into a predetermined calculation formula. The load prediction unit 12 calculates the above-mentioned weight prediction by inputting the time-series data of the detected value of the posture detection device 55 and the time-series data of the detected value of the first load pressure sensor 57 (load detection device) into the machine learning model.

[0069] In this configuration, time-series data of physical quantities related to the posture of the work device and the load pressure of the boom cylinder (hydraulic actuator) acquired during previously performed excavation and loading operations are used as input data, and the weight of the excavated object in the bucket at that time is used as training data to construct a machine learning model. By inputting the physical quantities related to the posture of the work device 31 and the load pressure of the boom cylinder 62 (hydraulic actuator) detected by the posture detection device 55 and the first load pressure sensor 57 (load detection device) during the actual excavation and loading operations, the weight of the excavated object in the bucket 43 is predicted. payload Highly accurate prediction values ​​can be obtained for this.

[0070] [Second Embodiment] Next, a load calculation device according to the second embodiment of the present invention will be described with reference to Figures 10 to 15. In Figures 10 to 15, parts with the same reference numerals as those shown in Figures 1 to 9 are similar parts, so their detailed explanation will be omitted. First, the functional configuration of the load calculation device according to the second embodiment will be described with reference to Figure 10. Figure 10 is a block diagram showing the hardware and functional configuration of the load calculation device according to the second embodiment.

[0071] The load calculation device 1A according to the second embodiment shown in Figure 10 differs from the first embodiment in that it further includes the functions of an operation determination unit 15A and a total load calculation unit 16A, and consequently, the function of the display generation unit 13A has been changed. The other functional configurations of the load calculation device 1A according to the second embodiment (load calculation unit 11 and load prediction unit 12) are the same as in the first embodiment, and their explanation will be omitted.

[0072] Specifically, the operation determination unit 15A of the load calculation device 1A determines which of the excavation and loading operations the current operation corresponds to, based on the detected values ​​of the posture detection device 55 and the first load pressure sensor 57 (sensor signals received by the sensor signal input device 2). The operation determination unit 15A determines, for example, whether the operation is an "excavation operation" in which the bucket 43 of the work device 31 excavates the target to be excavated and holds the excavated material in the bucket 43; a "transportation operation" in which the bucket 43 is rotated after the excavation operation and moved onto the loading platform of the transport machine; a "discharge operation" in which the excavated material in the bucket 43 is discharged onto the loading platform of the transport machine after the transport operation; or any other operation. The determination result of the operation determination unit 15A is output to the total load calculation unit 16A.

[0073] The total load calculation unit 16A calculates the cumulative weight of the excavation target loaded onto the transport machine by the construction machine 30 (hereinafter referred to as the cumulative load value). Specifically, when the judgment result of the operation determination unit 15A switches from transport operation to earth discharge operation, the total load calculation unit 16A updates the cumulative load value by adding the calculation result of the load calculation unit 11 to the current cumulative load value and outputs the updated cumulative load value to the display generation unit 13A. On the other hand, when the judgment result of the operation determination unit 15A is anything other than that, even if the calculation result of the load calculation unit 11 has been output, the calculation result is not added to the current cumulative load value, and the unupdated cumulative load value is output as is to the display generation unit 13A. In other words, when the excavated material is released from the bucket 43 to the transport machine, the total load calculation unit 16A adds the calculation result of the load calculation unit 11 to the cumulative load value from the previous cycle of the excavation and loading operation, thereby updating the cumulative load value by the weight of the excavated material loaded onto the transport machine during the current cycle of excavation and loading operation.

[0074] The display generation unit 13A receives the calculation results from the load calculation unit and the load prediction unit, as in the first embodiment. Furthermore, the calculation result from the total load calculation unit 16A is also received. The display generation unit 13A generates a display screen as display data that displays either the output (calculation result) of the load calculation unit 11 or the load prediction unit 12 on the display device 53, as in the first embodiment. In addition, it generates a display screen as display data that displays the output (calculation result) from the total load calculation unit 16A on the display device 53. That is, the display generation unit 13A generates a display screen as display data that shows the weight of the excavated object held in the bucket 43 during the current excavation and loading operation, and the cumulative load value of the excavated object loaded onto the transport machine by repeating the excavation and loading operation.

[0075] Next, the processing procedure of the operation determination unit in the load calculation device according to the second embodiment will be explained with reference to Figure 11. Figure 11 is a flowchart showing an example of the processing procedure of the operation determination unit in the load calculation device according to the second embodiment shown in Figure 10.

[0076] In Figure 11, the operation determination unit 15A of the load calculation device 1A shown in Figure 10 receives signals (detection data from the attitude detection device 55 and the first load pressure sensor 57) from the sensor signal input device 2 (step S110), and determines which of the excavation and loading operations corresponds to the received signals (sensor detected values) (steps S120, S140, S160).

[0077] The operation determination unit 15A first determines whether the received signal (sensor detected value) matches the conditions for excavation operation (step S120). If the conditions for excavation operation are met (YES), the operation determination unit 15A outputs the excavation operation determination result to the total load calculation unit 16A (step S130), and the processing of the flow ends.

[0078] On the other hand, if the conditions for excavation operation are not met in step S120 (in the case of NO), the operation determination unit 15A determines whether the received signal (sensor detected value) meets the conditions for transport operation (step S140). If the conditions for transport operation are met (in the case of YES), the operation determination unit 15A outputs the transport operation determination result to the total load calculation unit 16A (step S150), and the processing of the flow ends.

[0079] On the other hand, if the conditions for transport operation are not met in step S140 (in the case of NO), the operation determination unit 15A determines whether the acquired signal (sensor detected value) meets the conditions for release operation (step S160). If the conditions for release operation are met (in the case of YES), the operation determination unit 15A outputs the release operation determination result to the total load calculation unit 16A (step S170), and the processing of that flow ends. On the other hand, if the conditions for release operation are not met (in the case of NO), the operation determination unit 15A outputs the determination results for other operations to the total load calculation unit 16A (step S180), and the processing of that flow ends.

[0080] In this way, the operation determination unit 15A determines whether the current operation corresponds to excavation or loading operation based on the signal (sensor detected value) received from the sensor signal input device 2. Note that the order of condition determination for excavation operation, transport operation, and discharge operation (steps S120, S140, S160) is arbitrary.

[0081] Next, the processing procedure of the total load calculation unit in the load calculation device according to the second embodiment will be explained with reference to Figure 12. Figure 12 is a flowchart showing an example of the processing procedure of the total load calculation unit in the load calculation device according to the second embodiment shown in Figure 10.

[0082] In Figure 12, the total load calculation unit 16A of the load calculation device 1A shown in Figure 10 receives the calculation results from the load calculation unit 11 and the judgment results from the operation judgment unit 15A (step S210).

[0083] Next, the total load calculation unit 16A determines whether the determination result from the operation determination unit 15A is a transport operation (step S220). If the determination result is not a transport operation (NO), the total load calculation unit 16A outputs the current cumulative load value to the display generation unit 13A (step S230), and the processing of that flow ends.

[0084] On the other hand, if the determination result in step S220 is a transport operation (YES), the total load calculation unit 16A outputs the current cumulative load value to the display generation unit 13A (step S240) and takes the next determination result from the operation determination unit 15A (step S250). Furthermore, the total load calculation unit 16A determines whether the determination result from the operation determination unit 15A is a release operation (step S260). If the determination result is not a release operation (NO), the total load calculation unit 16A executes steps S240 to S260 again. The total load calculation unit 16A repeats steps S240 to S260 until the determination result in step S260 is a release operation (YES).

[0085] If the determination result in step S260 is release operation (YES), the total load calculation unit 16A updates the cumulative load value by adding the calculation result of the load calculation unit 11 to the current cumulative load value, outputs the updated cumulative load value after the addition to the display generation unit 13A (step S270), and ends the processing of the flow. In other words, when the operation transitions from transport operation to release operation in one cycle of excavation and loading work, the total load calculation unit 16A updates the cumulative weight of the excavated object loaded on the transport machine (cumulative load value) by adding the weight of the excavated object held in the bucket 43.

[0086] Next, the processing procedure of the display generation unit in the load calculation device according to the second embodiment will be explained using Figure 13. Figure 13 is a flowchart showing an example of the processing procedure of the display generation unit in the load calculation device according to the second embodiment shown in Figure 10.

[0087] In Figure 13, the display generation unit 13A of the load calculation device 1A shown in Figure 10 first takes in the calculation result (cumulative load value) output from the total load calculation unit 16A (step S2A), and generates a display screen (display data) with the words "cumulative payload value" added to the "numerical value" of the output value of the total load calculation unit 16A, and outputs it to the display device 53 (step S4A).

[0088] Next, the display generation unit 13A executes the processes of steps S10 to S50, similar to the first embodiment. That is, the display generation unit 13A determines whether or not there is an output (calculation result) from the load calculation unit 11 (step S10), and if there is an output from the load calculation unit 11 (YES), it generates a display screen (display data) with the words "payload value" added to the "numerical value" of the calculation result from the load calculation unit 11 and outputs it to the display device 53 (step S50). On the other hand, if there is no output from the load calculation unit 11 (NO), the display generation unit 13A determines whether or not there is an output (prediction result) from the load prediction unit 12 (step S20), and if there is an output from the load prediction unit 12 (YES), it generates a display screen (display data) with the words "payload prediction value" added to the "numerical value" of the prediction result from the load prediction unit 12 and outputs it to the display device 53 (step S40). On the other hand, if there is no output from the load prediction unit 12 (in the case of NO), the display generation unit 13A generates a default display screen (display data) and outputs it to the display device 53 (step S30).

[0089] In this way, the display generation unit 13A generates a display screen (display data) that displays the weight of the excavated object in the bucket during one cycle of the excavation and loading work, as well as a display screen (display data) that displays the cumulative load value of the excavated object loaded onto the transport machine, through the processing of steps S2A to S4A and steps S10 to S50.

[0090] Next, the display screen of the display device corresponding to the output result of the load calculation device according to the second embodiment for the excavation and loading work of construction machinery will be explained using Figures 14 and 15. The upper figures of Figures 14 and 15 show the flow of the excavation and loading work from the completion of the excavation operation to the middle of the transport operation and then to the release operation. The lower figures of Figures 14 and 15 show the display screen of the display device corresponding to the output result of the load calculation device according to the second embodiment for the working state of the construction machinery.

[0091] The upper part of Figure 14 shows the hydraulic excavator 30 in a state where the bucket 43 is being aligned with the loading platform 201 of the transport machine 200 in order to load the excavated material onto the transport machine 200 during the excavation and loading operation.

[0092] At this time (when the transport operation is nearing its end and the bucket 43 is in the air), the load pressure of the boom cylinder 62 stabilizes, and the calculation result of the load calculation unit 11 is output. For this reason, the load calculation device 1A (display generation unit 13A) generates a display screen showing the payload calculation value, which is the calculation result of the load calculation unit 11, according to step S50 of the flowchart shown in Figure 13, and outputs it to the display device 53.

[0093] Furthermore, at this time (just before the end of the transport operation and before the start of the release operation), the operation determination unit 15A of the load calculation device 1A determines that it is a transport operation. Therefore, the total load calculation unit 16A of the load calculation device 1A outputs the cumulative load value before the update, which does not include the calculation result of the load calculation unit 11, to the display generation unit 13A, in accordance with step S240 of the flowchart shown in Figure 12. For this reason, the load calculation device 1A (display generation unit 13A) generates a display screen that shows the cumulative load value, which is the calculation result of the total load calculation unit 16A and does not include the calculation result of the load calculation unit 11, as the cumulative payload value, in accordance with step S4A of the flowchart shown in Figure 13, and outputs it to the display device 53.

[0094] As a result, the display device 53 displays a screen that includes the string "Payload Value" and the calculated value (e.g., 31t) from the load calculation unit 11, as shown in the lower diagram of Figure 14, as well as the string "Cumulative Payload Value" and the output value (e.g., 0t) from the total load calculation unit 16A. For the first cycle of the excavation and loading operation, the weight of the excavated object (cumulative load value) loaded on the loading platform 201 of the transport machine 200 is set to "0t" by default.

[0095] In this manner, the load calculation device 1A displays the cumulative load value on the display device 53, which does not include the weight of the excavated object held in the bucket 43, before the start of the discharge operation of the excavation and loading work.

[0096] The upper diagram of Figure 15 shows the transition from the state near the end of the transport operation shown in the upper diagram of Figure 14 to the release operation. Specifically, the bucket 43 of the hydraulic excavator 30 is driven in the dumping direction.

[0097] At this time (when the state has transitioned to the release operation), the operation determination unit 15A determines that the operation is a release operation, so the total load calculation unit 16A outputs the updated cumulative load value, which is the result of adding the calculation result of the load calculation unit 11, to the display generation unit 13A in accordance with step S270 of the flowchart shown in Figure 12. For this reason, the load calculation device 1A (display generation unit 13A) generates a display screen that shows the updated cumulative load value, which is the calculation result of the total load calculation unit 16A, as the cumulative payload value in accordance with step S4A of the flowchart shown in Figure 13, and outputs it to the display device 53.

[0098] As a result, the display device 53 displays a screen that includes the string "Payload Value" and the "Value (e.g., 31t)" of the calculation result of the load calculation unit 11, as well as the string "Cumulative Payload Value" and the "Value (e.g., 31t)" of the output result of the total load calculation unit 16A, as shown in the lower diagram of Figure 15. The "Cumulative Payload Value" is the value obtained by adding the "Calculated Payload Value," which is the calculation result of the load calculation unit 11, to the original value, i.e., 31t.

[0099] As described above, the load calculation device 1A according to this embodiment, similar to the first embodiment, calculates or predicts the weight of the excavated material in the bucket 43 during one cycle of the excavation and loading operation of the construction machine 30, and displays the calculation result on the display device 53. In addition, it calculates the cumulative load value of the excavated material loaded onto the transport machine 200 by the multiple cycles of the excavation and loading operation that have been performed so far, and displays the cumulative load value on the display device 53. This allows the operator of the construction machine 30 to check the cumulative load value of the transport machine 200 at the time of loading the excavated material onto the transport machine 200, and to determine the target value of the excavation weight for the next cycle of the excavation and loading operation before the excavation operation.

[0100] According to the load calculation device 1A of the second embodiment described above, similar to the first embodiment described above, during excavation, the load prediction unit 12 can use a machine learning model to calculate a predicted value of the weight of the excavated object in the bucket 43, and during transport, the load calculation unit 11 can directly calculate the weight of the excavated object in the bucket 43 from the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 (load detection device). In other words, the weight of the excavated object in the bucket 43 can be calculated in either the excavation operation or the transport operation during the excavation and loading work of the construction machine 30.

[0101] Furthermore, the load calculation device 1A according to this embodiment further comprises an operation determination unit 15A that determines the operation being performed by the hydraulic excavator 30 (construction machine), and a total load calculation unit 16A that calculates the cumulative weight of the excavated object loaded onto the transport machine 200 by the hydraulic excavator 30 (construction machine). The total load calculation unit 16A updates the cumulative value by adding the calculated value from the load calculation unit 11 when the determination result of the operation determination unit 15A switches from the transport operation to the release operation in a series of operations during the excavation and loading work.

[0102] With this configuration, the updated cumulative load value of the excavation target loaded onto the transport machine 200 can be acquired at the time of loading onto the transport machine 200, making it possible to determine the target value of the excavation weight for the next cycle of excavation and loading work before the excavation operation begins.

[0103] Furthermore, the load calculation device 1A according to this embodiment further includes a display generation unit 13A that generates display data for displaying on the display device 53 either the calculated value from the load calculation unit 11 or the predicted value from the load prediction unit 12, along with the cumulative value from the total load calculation unit 16A, and outputs this data to the display device 53.

[0104] With this configuration, the cumulative load value of the excavation target loaded onto the transport machine 200 is updated and displayed on the display device 53 at the time of loading onto the transport machine 200 in the current cycle of the excavation and loading work. As a result, the operator of the construction machine 30 can determine the target value of the excavation weight for the next cycle of the excavation and loading work before the excavation operation, based on the cumulative load value displayed on the display device 53.

[0105] [Third Embodiment] Next, a load calculation device according to the third embodiment of the present invention will be described with reference to Figures 16 to 20. In Figures 16 to 20, parts with the same reference numerals as those shown in Figures 1 to 15 are similar parts, so their detailed explanation will be omitted. First, the functional configuration of the load calculation device according to the third embodiment will be described with reference to Figure 16. Figure 16 is a block diagram showing the hardware and functional configuration of the load calculation device according to the third embodiment.

[0106] The load calculation device 1B according to the third embodiment shown in Figure 16 differs from the second embodiment in that it further includes the function of the addition value selection unit 14B, and consequently, the function of the total load calculation unit 16B has been changed. The other functional configurations of the load calculation device 1B according to the third embodiment (load calculation unit 11, load prediction unit 12, display generation unit 13A, operation determination unit 15A) are the same as in the second embodiment, and their description is omitted.

[0107] Specifically, the addition value selection unit 14B of the load calculation device 1B outputs either the calculation result of the load calculation unit 11 or the prediction result of the load prediction unit 12 to the total load calculation unit 16B. More specifically, when the calculation result of the load calculation unit 11 is output, the addition value selection unit 14B outputs the calculation result of the load calculation unit 11 to the total load calculation unit 16B. When the calculation result of the load calculation unit 11 is not output, but the prediction result of the load prediction unit 12 is output, the prediction result of the load prediction unit 12 is output to the total load calculation unit 16B. In other words, in situations where the calculation result of the load calculation unit 11 cannot be obtained, for example during excavation, the addition value selection unit 14B outputs the prediction value of the load prediction unit 12 as a substitute value for the addition calculation of the total load calculation unit 16B.

[0108] The total load calculation unit 16B calculates the cumulative weight of the excavated material loaded onto the transport machine according to the selection result of the addition value selection unit 14B (the selected value of either the calculated value of the load calculation unit 11 or the predicted value of the load prediction unit 12). Specifically, when the judgment result of the operation determination unit 15A switches from transport operation to soil discharge operation, the total load calculation unit 16B adds either the calculated value of the load calculation unit 11 or the predicted value of the load prediction unit 12, which is the selection result of the addition value selection unit 14B, to the current cumulative load value to update the cumulative load value and outputs the updated cumulative load value to the display generation unit 13A. On the other hand, if the judgment result of the operation determination unit 15A is anything other than this, even if the output value (selection result) of the addition value selection unit 14B is input, the current cumulative load value is not added to the current cumulative load value, and the current cumulative load value is output to the display generation unit 13A.

[0109] Next, the processing procedure of the addition value selection unit in the load calculation device according to the third embodiment will be explained using Figure 17. Figure 17 is a flowchart showing an example of the processing procedure of the addition value selection unit in the load calculation device according to the third embodiment shown in Figure 16.

[0110] In Figure 17, the addition value selection unit 14B of the load calculation device 1B shown in Figure 16 first determines whether or not there is an output (calculation result) from the load calculation unit 11 (step S310). If there is an output from the load calculation unit 11 (YES), the addition value selection unit 14B outputs the output value (calculated value) from the load calculation unit 11 to the total load calculation unit 16B (step S320), and the processing of that flow ends. On the other hand, if there is no output from the load calculation unit 11 (NO), the addition value selection unit 14B determines whether or not there is an output (prediction result) from the load prediction unit 12 (step S330).

[0111] If the load prediction unit 12 outputs an output in step S330 (YES), the addition value selection unit 14B outputs the output value (predicted value) from the load prediction unit 12 to the total load calculation unit 16B (step S340), and terminates the processing of that flow. On the other hand, if the load prediction unit 12 does not output an output (NO), the process returns to step S310. The processing in steps S310 and S330 is repeated until there is an output (calculation result) from the load calculation unit 11 or an output (prediction result) from the load prediction unit 12.

[0112] Next, the processing procedure of the total load calculation unit in the load calculation device according to the third embodiment will be explained with reference to Figure 18. Figure 18 is a flowchart showing an example of the processing procedure of the total load calculation unit in the load calculation device according to the third embodiment shown in Figure 16.

[0113] In Figure 18, the total load calculation unit 16B of the load calculation device 1B shown in Figure 15, unlike in the second embodiment, takes in the selection result from the addition value selection unit 14B (either the calculation result from the load calculation unit 11 or the prediction result from the load prediction unit 12) and also takes in the determination result from the operation determination unit 15A (step S210B).

[0114] Next, the total load calculation unit 16B, similar to the second embodiment, determines whether the determination result of the operation determination unit 15A is a transport operation (step S220). If the determination result is not a transport operation (NO), it outputs the current cumulative load value to the display generation unit 13A (step S230).

[0115] On the other hand, if the determination result in step S220 is a transport operation (YES), the total load calculation unit 16B outputs the current cumulative load value to the display generation unit 13A (step S240). Furthermore, it takes the next determination result from the operation determination unit 15A (step S250) and determines whether the determination result is a release operation or not (step S260). If the determination result is not a release operation (NO), the total load calculation unit 16B executes steps S240 to S260 again. The total load calculation unit 16B repeats steps S240 to S260 until the determination result in step S260 is a release operation (YES).

[0116] In step S260, if the determination result is release operation (YES), the total load calculation unit 16B updates the cumulative load value by adding the calculated value of the load calculation unit 11 or the predicted value of the load prediction unit 12, which is the selected value of the addition value selection unit 14B, and outputs the updated cumulative load value after the addition to the display generation unit 13A (step S270B). That is, when the excavation and loading operation transitions from transport operation to release operation, the total load calculation unit 16B adds either the calculated value of the load calculation unit 11 or the predicted value of the load prediction unit 12 to the cumulative load value, which is the weight of the excavated object loaded onto the transport machine up to one cycle before the excavation and loading operation. Unlike in the second embodiment, the total load calculation unit 16B can update the cumulative load value using the predicted value of the load prediction unit 12 even if the calculation result of the load calculation unit 11 is not output for some reason.

[0117] Next, the display screen of the display device corresponding to the output result of the load calculation device according to the third embodiment for the excavation and loading work of construction machinery will be explained using Figures 19 and 20. The upper figures of Figures 19 and 20 show the flow of the excavation and loading work from the completion of the excavation operation to the middle of the transport operation and to the release operation. The lower figures of Figures 19 and 20 show the display screen of the display device corresponding to the output result of the load calculation device according to the third embodiment for the working state of the construction machinery.

[0118] The upper part of Figure 19 shows the hydraulic excavator 30 in a state where the bucket 43 is being aligned with the loading platform 201 of the transport machine 200 in order to load the excavated material onto the transport machine 200 during the excavation and loading operation.

[0119] At this time (near the end of the transport operation), the calculation result of the load calculation unit 11 may not be available for some reason. On the other hand, the prediction result of the load prediction unit 12 is usually output based on the time-series data of the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 detected during the excavation operation. Therefore, the load calculation device 1B (display generation unit 13A) generates a display screen showing the payload prediction value, which is the prediction result of the load prediction unit 12, in accordance with step S40 of the flowchart shown in Figure 13, and outputs it to the display device 53.

[0120] Furthermore, at this time (just before the end of the transport operation and before the start of the release operation), the addition value selection unit 14B of the load calculation device 1B cannot obtain the calculation result from the load calculation unit 11, so it outputs the prediction result from the load prediction unit 12 to the total load calculation unit 16B in accordance with step S340 of the flowchart shown in Figure 17. The total load calculation unit 16B determines from the operation determination unit 15A of the load calculation device 1B that it is a transport operation, so it outputs the current cumulative load value to the display generation unit 13A in accordance with step S240 of the flowchart shown in Figure 18. For this reason, the load calculation device 1B (display generation unit 13A) generates a display screen showing the current cumulative load value, which is the calculation result of the total load calculation unit 16B, as the cumulative payload value in accordance with step S4A of the flowchart shown in Figure 13, and outputs it to the display device 53.

[0121] As a result, the display device 53 displays a screen that includes the string "Payload Prediction Value" and the "Value (e.g., 31t)" of the prediction result from the load prediction unit 12, as well as the string "Total Payload Value" and the "Value (e.g., 0t)" of the output result from the total load calculation unit 16B, as shown in the lower diagram of Figure 19. When the excavation and loading operation is in its first cycle, the total load value of the excavated object loaded onto the transport machine is set to "0t" by default.

[0122] The upper diagram in Figure 20 shows the transition from the state near the end of the transport operation shown in the upper diagram in Figure 19 to the release operation. Specifically, the bucket 43 of the hydraulic excavator 30 is driven in the dumping direction.

[0123] At this time (when the system has transitioned to the release operation), the total load calculation unit 16B, because the operation determination unit 15A's determination result is a release operation, outputs the updated cumulative load value obtained by adding the output value of the addition value selection unit 14B according to step S270B of the flowchart shown in Figure 18, to the display generation unit 13A. In this case, since the calculation result of the load calculation unit 11 has not yet been obtained, the addition value selection unit 14B outputs the prediction result of the load prediction unit 12 to the total load calculation unit 16B. Therefore, the load calculation device 1B (display generation unit 13A), according to step S4A of the flowchart shown in Figure 13, generates a display screen showing the updated cumulative load value obtained by adding the prediction value of the load prediction unit 12 (output value of the addition value selection unit 14B), which is the calculation result of the total load calculation unit 16B, as the cumulative payload value, and outputs it to the display device 53.

[0124] As a result, the display device 53 displays a screen that includes the string "Payload Prediction Value" and the "Value (e.g., 31t)" of the prediction result from the load prediction unit 12, as well as the string "Total Payload Value" and the "Value (e.g., 31t)" of the output result from the total load calculation unit 16B, as shown in the lower diagram of Figure 20. The "Total Payload Value" is the value obtained by adding the "Payload Prediction Value," which is the prediction result from the load prediction unit 12 (output value from the addition value selection unit 14B), i.e., 31t.

[0125] As described above, in the case where the calculation result of the load calculation unit 11 cannot be obtained for any reason, the load calculation device 1B according to this embodiment displays a payload prediction value, which is the predicted result of the excavation weight during one cycle of the excavation and loading operation of the construction machine 30, on the display device. In addition, when the operation transitions to release, the display device 53 displays an updated value on the transport machine 200 as the cumulative load value, by adding the predicted weight of the excavated object that may be excavated during that cycle of the excavation and loading operation. As a result, even if the calculation result of the load calculation unit 11 cannot be obtained, the predicted value of the load prediction unit 12 can be used as a substitute value to provide the operator of the construction machine 30 with a relatively high accuracy cumulative load value.

[0126] According to the load calculation device 1B of the third embodiment described above, similar to the second embodiment described above, during excavation, the load prediction unit 12 can use a machine learning model to calculate a predicted value of the weight of the excavated object in the bucket 43, and during transport, the load calculation unit 11 can directly calculate the weight of the excavated object in the bucket 43 from the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 (load detection device). In other words, the weight of the excavated object in the bucket 43 can be calculated in either the excavation operation or the transport operation during the excavation and loading work of the construction machine 30.

[0127] Furthermore, the load calculation device 1B according to this embodiment further includes an operation determination unit 15A that determines the operation being performed by the construction machine 30, a total load calculation unit 16B that calculates the cumulative weight of the excavation target loaded onto the transport machine by the construction machine 30, and an addition value selection unit 14B that outputs either the calculated value from the load calculation unit 11 or the predicted value from the load prediction unit 12 to the total load calculation unit 16B. The addition value selection unit 14B outputs the calculated value from the load calculation unit 11 when both the calculated value from the load calculation unit 11 and the predicted value from the load prediction unit 12 are available. The total load calculation unit 16B updates the cumulative value by adding the output value from the addition value selection unit 14B when the determination result of the operation determination unit 15A switches from the transport operation to the release operation in a series of operations in the excavation and loading work.

[0128] With this configuration, even if a calculated value from the load calculation unit 11 cannot be obtained for any reason during a certain cycle of excavation and loading work, the cumulative value of the total load calculation unit 16B is updated by using the predicted value from the load prediction unit 12 as a substitute value for the calculated value from the load calculation unit 11. This ensures that the cumulative weight of the excavated material loaded onto the transport machine by the excavation and loading work performed so far is reliably provided.

[0129] Furthermore, if a calculated value is obtained from the load calculation unit 11, the calculated value from the load calculation unit 11 is preferentially added in the calculation of the total load calculation unit 16B to update the cumulative value, thereby maintaining a high level of accuracy in the cumulative weight value which is the calculation result of the total load calculation unit 16B.

[0130] Furthermore, the load calculation device 1B according to this embodiment further includes a display generation unit 13A that generates display data for displaying on the display device 53 either the calculated value from the load calculation unit 11 or the predicted value from the load prediction unit 12, and the cumulative value from the total load calculation unit 16B, and outputs this data to the display device 53. If the calculated value from the load calculation unit 11 has been obtained, the display generation unit 13A generates display data for displaying the calculated value from the load calculation unit 11 and the cumulative value from the total load calculation unit 16B.

[0131] With this configuration, even if the calculated value from the load calculation unit 11 cannot be obtained for any reason, the predicted weight of the excavated material held in the bucket 43 during one cycle of excavation and loading work, and the cumulative weight of the excavated material that would have been loaded onto the transport machine by the excavation and loading work so far, are displayed on the display device 53. Therefore, the operator of the construction machine 30 can constantly check the weight information of the excavated material during the excavation and loading work of the construction machine 30 on the display device 53.

[0132] Furthermore, if a calculated value is obtained from the load calculation unit 11, the display data of the calculated value from the load calculation unit 11 is preferentially displayed on the display device 53, and the display data of the cumulative value (calculation result of the total load calculation unit 16B), which has been updated by preferentially adding the calculated value from the load calculation unit 11, is also displayed on the display device 53. As a result, highly accurate weight information can be provided to the operator of the construction machine 30.

[0133] [Fourth Embodiment] Next, a load calculation device according to the fourth embodiment of the present invention will be described with reference to Figures 21 to 25. In Figures 21 to 25, parts with the same reference numerals as those shown in Figures 1 to 20 are similar parts, so their detailed explanation will be omitted. First, the functional configuration of the load calculation device according to the fourth embodiment will be described with reference to Figure 21. Figure 21 is a block diagram showing the hardware and functional configuration of the load calculation device according to the fourth embodiment.

[0134] The load calculation device 1C according to the fourth embodiment shown in Figure 21 differs from the second embodiment in that it further includes the function of a relearning execution unit 18C, and consequently, the functions of the load prediction unit 12C and the display generation unit 13C have been changed. The other functional configurations of the load calculation device 1C according to the fourth embodiment (load calculation unit 11, operation determination unit 15A, total load calculation unit 16A) are the same as in the second embodiment, and their description is omitted.

[0135] Specifically, the retraining execution unit 18C of the load calculation device 1C retrains the machine learning model used in the load prediction unit 12C based on the time-series data of the detected values ​​from the posture detection device 55 and the first load pressure sensor 57, as well as the calculation results from the load calculation unit 11. The machine learning model, which has been corrected and updated through retraining, is output to the load prediction unit 12C. In addition, when the retraining execution unit 18C is performing retraining of the machine learning model, it outputs information indicating that retraining is in progress to the display generation unit 13C. On the other hand, it also outputs information indicating that retraining of the machine learning model is not in progress.

[0136] In detail, the retraining execution unit 18C retrains the machine learning model used in the load prediction unit 12C by using time-series data of physical quantities related to the posture of the work equipment detected by the posture detection device 55 during the actual excavation and loading work of the hydraulic excavator 30, and time-series data of the bottom pressure and rod pressure of the boom cylinder 62 detected by the first load pressure sensor 57, i.e., the sensor signals received by the sensor signal input device 2, as input data, and by using the calculation result of the load calculation unit 11 during the said excavation and loading work (weight of the excavated object held in the bucket 43) as training data. As mentioned above, the machine learning model used by the load prediction unit 12C is constructed using deep learning with a neural network. Through retraining of the machine learning model, the gradient of the intermediate layer with respect to the time-series data of the detected values ​​of the posture detection device 55 and the first load pressure sensor 57 (input data) is tuned so that the one-dimensional output value of the output layer (i.e., the predicted weight of the excavated object that can be held in the bucket 43) matches the training data (calculation result of the load calculation unit 11).

[0137] The retraining execution unit 18C collects time-series data from the attitude detection device 55 and the first load pressure sensor 57, as well as the calculation results from the load calculation unit 11, during the "excavation operation" and "transportation operation" of the series of operations in the excavation and loading work. On the other hand, during operations other than the "excavation operation," "transportation operation," and "release operation" of the series of operations in the excavation and loading work, or during "other operations" which are operations other than excavation and loading work, the retraining of the machine learning model is performed using the collected time-series data and the calculation results from the load calculation unit 11. In detail, if the judgment result of the operation determination unit 15A is one of "excavation operation," "transportation operation," or "release operation," the retraining execution unit 18C collects time-series data from the attitude detection device 55 and the first load pressure sensor 57, as well as the calculation results from the load calculation unit 11. On the other hand, if the judgment result of the operation determination unit 15A is "any other operation," the retraining of the machine learning model is performed.

[0138] The load prediction unit 12C uses a machine learning model that has been retrained and modified by the retraining execution unit 18C to predict the maximum weight that the excavation target held in the bucket 43 can reach, based on the time-series data detected by the attitude detection device 55 and the first load pressure sensor 57 (i.e., the change in the attitude of the work device 31 and the change in the load pressure of the boom cylinder 62). payload The load prediction unit 12C uses a machine learning model that has been improved according to the actual excavation and loading work of the hydraulic excavator 30, so it can calculate the maximum weight m of the excavated material that can be scooped up into the bucket 43 during the excavation operation. payload It is now possible to predict this with high accuracy.

[0139] The display generation unit 13C generates a display screen (similar to the screen in the second embodiment) as display data for displaying the calculation result of either the load calculation unit 11 or the load prediction unit 12 and the calculation result of the total load calculation unit 16A on the display device 53. In addition, it generates a display screen as display data for displaying "Learning" on the display device 53 according to the relearning flag from the relearning execution unit 18C. The generated display screen (display data) is output to the display device 53.

[0140] Next, the processing procedure of the relearning execution unit in the load calculation device according to the fourth embodiment will be explained with reference to Figure 22. Figure 22 is a flowchart showing an example of the processing procedure of the relearning execution unit in the load calculation device according to the fourth embodiment shown in Figure 21.

[0141] In Figure 22, the relearning execution unit 18C of the load calculation device 1C shown in Figure 21 starts the flow shown in Figure 22 when it detects the output from the sensor signal input device 2 and the output (calculation result) from the load calculation unit 11. The relearning execution unit 18C takes in the judgment result from the operation judgment unit 15A (step S410) and determines whether the judgment result is "other operation" or not (step S420).

[0142] If the result in step S420 is NO (the result of the operation determination unit 15A is not "other operation"), the retraining execution unit 18C collects sensor data from the sensor signal input unit 2 and the calculation results from the load calculation unit 11 as data for retraining (step S430). Furthermore, the retraining execution unit 18C outputs information to the display generation unit 13C indicating that the machine learning model is not being retrained (step S440), and terminates the processing of the flow.

[0143] On the other hand, if the answer in step S420 is YES (the result of the operation determination unit 15A is "other operation"), the retraining execution unit 18C uses the sensor data collected from the sensor signal input device 2 and the calculation results of the load calculation unit 11 to retrain the machine learning model used in the load prediction unit 12C (step S450). Furthermore, the retraining execution unit 18C outputs information to the display generation unit 13C indicating that the machine learning model is being retrained (step S460), and the processing of this flow is terminated.

[0144] Next, the processing procedure of the display generation unit in the load calculation device according to the fourth embodiment will be explained using Figure 23. Figure 23 is a flowchart showing an example of the processing procedure of the display generation unit in the load calculation device according to the fourth embodiment shown in Figure 21.

[0145] In Figure 23, the display generation unit 13C of the load calculation device 1C shown in Figure 21 performs the processing of steps S2A and S4A, similar to the case of the second embodiment. That is, the display generation unit 13C generates a display screen (display data) in which the "numerical value" of the calculation result of the total load calculation unit 16A is accompanied by the words "cumulative payload value", and outputs it to the display device 53.

[0146] Furthermore, the display generation unit 13C performs the processing of steps S10 to S50 in the same manner as in the second embodiment. That is, if there is an output from the load calculation unit 11 (if the answer is YES in step S10), the display generation unit 13C generates a display screen (display data) in which the "numerical value" of the calculation result of the load calculation unit 11 is accompanied by the words "payload value" (step S50). On the other hand, if there is no output from the load calculation unit 11 (if the answer is NO in step S10) and there is an output from the load prediction unit 12 (if the answer is YES in step S20), the display generation unit 13C generates a display screen (display data) in which the "numerical value" of the prediction result of the load prediction unit 12 is accompanied by the words "payload prediction value" (step S40). Also, if there is no output from the load calculation unit 11 (if the answer is NO in step S10) and there is no output from the load prediction unit 12 (if the answer is NO in step S20), the display generation unit 13C generates a default display screen (display data) (step S30).

[0147] In addition, the display generation unit 13C receives a signal from the retraining execution unit 18C (information indicating whether or not training is in progress) (step S60), synthesizes information corresponding to the signal (training in progress or not training in progress) with the display screen generated in any of steps S30 to S50, and outputs the synthesized display screen (display data) to the display device 53 (step S70). In this way, the display generation unit 13C generates a display screen (display data) indicating that the machine learning model is being retrained through the processing in steps S60 to S70.

[0148] Next, the display screen of the display device corresponding to the output result of the load calculation device according to the fourth embodiment for excavation and loading work of construction machinery will be explained using Figures 24 and 25. The upper part of Figure 24 shows the operating state of the construction machinery, while the upper part of Figure 25 shows the standby state of the construction machinery. The lower parts of Figures 24 and 25 show the display screen of the display device corresponding to the output result of the load calculation device according to the fourth embodiment for the working state of the construction machinery.

[0149] The hydraulic excavator 30 shown in Figure 24 has finished excavation using the bench cut method and has lifted the bucket 43 away from the excavation surface 101.

[0150] At this time (when the excavation operation is completed and the bucket 43 is in the air), the calculation result of the load calculation unit 11 is output, and the load calculation device 1C (display generation unit 13C) generates a display screen showing the payload calculation value of the calculation result of the load calculation unit 11 according to step S50 of the flowchart shown in Figure 23 and outputs it to the display device 53. As a result, the display device 53 displays a display screen that includes the string "payload value" and the "value (for example, 31t)" of the calculation result of the load calculation unit 11, as shown in the lower part of Figure 24.

[0151] Furthermore, at this time, since the determination result of the operation determination unit 15A is a transport operation, the total load calculation unit 16A outputs the cumulative load value, which has not been added to the calculation result of the load calculation unit 11, to the display generation unit 13C in accordance with step S240 of the flowchart shown in Figure 12. For this reason, the load calculation device 1C (display generation unit 13C) generates a display screen that shows the cumulative load value, which is the calculation result of the total load calculation unit 16A and does not include the calculation result of the load calculation unit 11, as the cumulative payload value, in accordance with step S4A of the flowchart shown in Figure 23, and outputs it to the display device 53. Note that the load calculation device 1C (display generation unit 13C) outputs a display screen in which information indicating the "learning in progress" state has not been added in accordance with step S70 of the flowchart shown in Figure 23, in accordance with the processing of step S440 of the flowchart shown in Figure 22 of the relearning execution unit 18C.

[0152] As a result, the display device 53 displays a screen that includes, in addition to the string "Payload Value" and the calculated value (e.g., 31t) from the load calculation unit 11, the string "Cumulative Payload Value" and the output value (e.g., 250t) from the total load calculation unit 16A, as shown in the lower diagram of Figure 24.

[0153] The hydraulic excavator 30 shown in Figure 25 is in a standby state with a parked position. At this time, the load calculation device 1C (display generation unit 13C) generates a display screen showing the cumulative load value, which is the calculation result of the total load calculation unit 16A, as the cumulative payload value, and a display screen showing the payload calculation value, which is the calculation result of the load calculation unit 11, in accordance with steps S4A and S50 of the flowchart shown in Figure 23, and outputs them to the display device 53. As a result, the display device 53 displays a display screen that includes the string "Payload Value" and the "Value (for example, 31t)" of the calculation result of the load calculation unit 11, as well as the string "Cumulative Payload Value" and the "Value (for example, 250t)" of the output result of the total load calculation unit 16A, as shown in the lower part of Figure 25.

[0154] Furthermore, at this time (when the hydraulic excavator is in standby mode), the determination result of the operation determination unit 15A is "other operation," so the retraining execution unit 18C performs retraining of the machine learning model according to steps S450 and S460 of the flowchart shown in Figure 22, and outputs information indicating that retraining is in progress to the display generation unit 13C. As a result, the load calculation device 1C (display generation unit 13C) synthesizes a screen containing the string "Learning" with the aforementioned display screen according to step S70 of the flowchart shown in Figure 23 and outputs it to the display device 53. As a result, the display device 53 displays a display screen containing the string "Learning" drawn within a frame, as shown in the lower part of Figure 25.

[0155] As described above, the load calculation device 1C (retraining execution unit 18C) according to this embodiment collects time-series data of the detected values ​​from the attitude detection device 55 and the first load pressure sensor 57 and the calculation results from the load calculation unit 11 as retraining data during predetermined operations of the excavation and loading work of the hydraulic excavator 30, and retrains the machine learning model of the load prediction unit 12C using the retraining data during operations other than predetermined operations. As a result, the machine learning model used in the load prediction unit 12C is modified and updated, thereby improving the accuracy of the predicted values ​​of the load prediction unit 12C.

[0156] According to the load calculation device 1C of the fourth embodiment described above, similar to the second embodiment described above, during excavation, the load prediction unit 12C can use a machine learning model to calculate a predicted value of the weight of the excavated object in the bucket 43, and during transport, the load calculation unit 11 can directly calculate the weight of the excavated object in the bucket 43 from the detected values ​​of the attitude detection device 55 and the first load pressure sensor 57 (load detection device). In other words, the weight of the excavated object in the bucket 43 can be calculated in either the excavation operation or the transport operation during the excavation and loading work of the construction machine 30.

[0157] Furthermore, the load calculation device 1C according to this embodiment further includes a retraining execution unit 18C that retrains the machine learning model of the load prediction unit 12C by using the time-series data of the detected values ​​from the posture detection device 55 and the time-series data of the detected values ​​from the first load pressure sensor 57 (load detection device) as input data, and using the calculated values ​​from the load calculation unit 11 as training data. The load prediction unit 12C updates the machine learning model by the retraining performed by the retraining execution unit 18C.

[0158] With this configuration, the time-series data of the values ​​detected by the attitude detection device 55 and the first load pressure sensor 57 (load detection device) collected during the actual excavation and loading work of the construction machine 30, along with the corresponding calculated values ​​from the load calculation unit 11, are used as relearning data to retrain and update the machine learning model of the load prediction unit 12C. As a result, as the operating time of the construction machine 30 increases, the deviation of the predicted values ​​of the load prediction unit 12C from the calculated values ​​of the load calculation unit 11 can be suppressed. In other words, the accuracy of the predicted values ​​of the load prediction unit 12C can be improved by relearning the machine learning model of the load prediction unit 12C.

[0159] Furthermore, the display generation unit 13C in the load calculation device 1C according to this embodiment is configured to generate display data to indicate that the machine learning model is being retrained when the retraining execution unit 18C is performing retraining of the machine learning model, and output this data to the display device 53.

[0160] With this configuration, the display device 53 shows that the machine learning model is being retrained, so that information about the status of the machine learning model can be provided to the operator of the construction machine 30.

[0161] [Other Embodiments] In the embodiments described above, examples were shown in which the load calculation devices 1, 1A, 1B, and 1C output calculation results during the excavation and loading work of the hydraulic excavator 30. However, these embodiments can also be applied to various construction machines other than hydraulic excavators that are capable of excavation and loading work.

[0162] Furthermore, the present invention is not limited to this embodiment and includes various modifications. The embodiments described above are explained in detail for the purpose of clearly illustrating the present invention and are not necessarily limited to those having all the configurations described. It is possible to replace parts of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add configurations from other embodiments to the configuration of one embodiment. In addition, it is possible to add, delete, or replace parts of the configuration of each embodiment with other configurations.

[0163] For example, in this embodiment, an example of a configuration in which load calculation devices 1, 1A, 1B, and 1C are provided on the construction machine 30 is shown. However, it is also possible to configure the load calculation device to be provided on a remote control device that remotely controls the construction machine 30.

[0164] Furthermore, in this embodiment, the load calculation devices 1, 1A, 1B, and 1C are shown as having a display generation unit 13, 13A, and 13C that generates a display screen showing the calculation results of the load calculation unit 11 or the load prediction units 12 and 12C as display data and outputs it to the display device 53 of the construction machine 30. However, the load calculation devices 1, 1A, 1B, and 1C can be configured without a display generation unit. For example, when a construction machine performs excavation and loading work by automatic control, it is not necessary to output a display screen (display data) to the display device 53 that provides information to the operator. In this case, the load calculation device can output the calculation results of the load calculation unit 11 or the load prediction units 12 and 12C to the control device that performs automatic control of the construction machine. As a result, the control device can determine control command values ​​such as the target trajectory of the excavation operation in the excavation and loading work based on the calculation results of the load calculation unit 11 or the load prediction units 12 and 12C.

[0165] Furthermore, in this embodiment, when constructing the machine learning models for the load prediction units 12 and 12C of the load calculation devices 1, 1A, 1B, and 1C, an example was shown in which time-series data of physical quantities related to the posture of the work device and time-series data of the load pressure (bottom pressure and rod pressure) of the boom cylinder were used as training data. However, the machine learning models of the load prediction units 12 and 12C can also use time-series data of the load pressure (bottom pressure and rod pressure) of the arm cylinder and bucket cylinder that drive the arm and bucket constituting the work device, in addition to time-series data of physical quantities related to the posture of the work device and time-series data of the load pressure (bottom pressure and rod pressure) of the boom cylinder as training data. In this case, the load prediction units 12 and 12C input time-series data of the detected values ​​from the first load pressure sensor 57, as well as time-series data of the detected values ​​from the second load pressure sensor 58 (bottom pressure and rod pressure of the arm cylinder 63) and the detected values ​​from the third load pressure sensor 59 (bottom pressure and rod pressure of the bucket cylinder 64) to the machine learning model to calculate an achievable predicted weight of the excavated object to be scooped up by the bucket 43.

[0166] Similarly, in the fourth embodiment, when the retraining execution unit 18C of the load calculation device 1C retrains the machine learning model, an example was shown in which time-series data of the detected values ​​of the posture detection device 55 (physical quantities related to the posture of the work device 31) and time-series data of the detected values ​​of the first load pressure sensor 57 (bottom pressure and rod pressure of the boom cylinder 62) were used as retraining data. However, if the machine learning model of the load prediction unit is constructed using time-series data of the load pressure of the arm cylinder and bucket cylinder that drive the arm and bucket, in addition to time-series data of the physical quantities related to the posture of the work device and the load pressure of the boom cylinder, the retraining execution unit can be configured to execute using time-series data of the detected values ​​of the second load pressure sensor 58 (bottom pressure and rod pressure of the arm cylinder 63) and time-series data of the detected values ​​of the third load pressure sensor 59 (bottom pressure and rod pressure of the bucket cylinder 64), in addition to the time-series data of the detected values ​​of the posture detection device 55 and the first load pressure sensor 57, as retraining data.

[0167] 1, 1A, 1B, 1C...Load calculation device, 11...Load calculation unit, 12, 12C...Load prediction unit, 13, 13A, 13C...Display generation unit, 14B...Addition value selection unit, 15A...Operation determination unit, 16A, 16B...Total load calculation unit, 18C...Relearning execution unit, 30...Hydraulic excavator (construction machine), 31...Working device, 43...Bucket, 53...Display device, 55...Attitude detection device, 57...First load pressure sensor (load detection device)

Claims

1. A load calculation device for calculating the weight of an excavated object in a bucket during excavation and loading work of a construction machine equipped with a working device including a bucket, a posture detection device for detecting the posture of the working device, and a load detection device for detecting the load acting on the working device, comprising: a load calculation unit that calculates the weight of an excavated object in the bucket based on a detected value of the posture of the working device detected by the posture detection device and a detected value of the load of the working device detected by the load detection device; and a load prediction unit that calculates a predicted value of the weight of an excavated object in the bucket based on a machine learning model constructed using the detected value of the posture of the working device detected by the posture detection device and the detected value of the load of the working device detected by the load detection device, and predetermined data showing the time-series changes in the operating state of the construction machine acquired during previously performed excavation and loading work as learning data.

2. A load calculation device according to claim 1, further comprising a display generation unit that generates display data for displaying on a display device the calculated value of the weight of the excavation target from the load calculation unit, or the predicted value of the weight of the excavation target from the load prediction unit.

3. A load calculation device according to claim 2, wherein the display generation unit generates display data for displaying the calculated value from the load calculation unit and outputs it to the display device when both the calculated value from the load calculation unit and the predicted value from the load prediction unit have been obtained.

4. A load calculation device according to claim 1, wherein the load detection device is configured to detect the load pressure of a hydraulic actuator that drives the work device as a physical quantity of load acting on the work device; the machine learning model of the load prediction unit is constructed by using time-series data of physical quantities relating to the posture of the work device and time-series data of the load pressure of the hydraulic actuator acquired during the previously performed excavation and loading work as input data, and by using the weight of the object to be excavated in the bucket acquired during the previously performed excavation and loading work as training data; the load calculation unit calculates the calculated weight by inputting the detected value of the posture detection device and the detected value of the load detection device to a predetermined calculation formula; and the load prediction unit calculates the predicted weight by inputting the time-series data of the detected value of the posture detection device and the time-series data of the detected value of the load detection device to the machine learning model.

5. A load calculation device according to claim 4, further comprising a retraining execution unit that retrains the machine learning model by using time-series data of the detected values ​​from the posture detection device and the time-series data of the detected values ​​from the load detection device as input data and the calculated values ​​from the load calculation unit as training data, wherein the load prediction unit updates the machine learning model by the retraining performed by the retraining execution unit.

6. A load calculation device according to claim 5, further comprising a display generation unit that generates display data for displaying the calculated value from the load calculation unit or the predicted value from the load prediction unit on a display device, wherein the display generation unit further generates display data for indicating that the machine learning model is being retrained when the retraining execution unit is performing retraining of the machine learning model, and outputs this to the display device.

7. A load calculation device according to claim 1, further comprising: an operation determination unit that determines the operation being performed by the construction machine; and a total load calculation unit that calculates the cumulative weight of the excavated object loaded onto the transport machine by the construction machine, wherein the total load calculation unit updates the cumulative value by adding the calculated weight of the excavated object from the load calculation unit when the determination result of the operation determination unit switches from a transport operation to a release operation in a series of operations in the excavation and loading work.

8. A load calculation device according to claim 7, further comprising a display generation unit that generates display data for displaying on a display device either the calculated value from the load calculation unit or the predicted value from the load prediction unit and the cumulative value from the total load calculation unit, and outputs this data to the display device.

9. A load calculation device according to claim 1, further comprising: an operation determination unit that determines the operation being performed by the construction machine; a total load calculation unit that calculates the cumulative weight of the excavation target loaded onto the transport machine by the construction machine; and an addition value selection unit that outputs either the calculated weight of the excavation target from the load calculation unit or the predicted weight of the excavation target from the load prediction unit to the total load calculation unit, wherein the addition value selection unit outputs the calculated value from the load calculation unit when both the calculated value from the load calculation unit and the predicted value from the load prediction unit are available, and the total load calculation unit updates the cumulative value by adding the output value from the addition value selection unit when the determination result of the operation determination unit switches from a transport operation to a release operation in a series of operations in the excavation and loading work.

10. A load calculation device according to claim 9, further comprising a display generation unit that generates display data for displaying on a display device either the calculated value from the load calculation unit or the predicted value from the load prediction unit and the cumulative value from the total load calculation unit, and outputs this data to the display device, wherein the display generation unit generates display data for displaying the calculated value from the load calculation unit and the cumulative value from the total load calculation unit when the calculated value from the load calculation unit is available.

11. A construction machine comprising: a work device including a bucket; a posture detection device for detecting the posture of the work device; a load detection device for detecting the load acting on the work device; a load calculation device for calculating the weight of the excavated object in the bucket during excavation and loading work; and a display device for displaying the calculation results of the load calculation device, wherein the load calculation device comprises: a load calculation unit for calculating the weight of the excavated object in the bucket based on the detected posture value of the work device detected by the posture detection device and the detected load value of the work device detected by the load detection device; a load prediction unit for calculating a predictable value of the weight of the excavated object in the bucket based on the detected posture value of the work device detected by the posture detection device and the detected load value of the work device detected by the load detection device, and a machine learning model constructed using predetermined data showing the time-series changes in the operating state of the construction machine acquired during previously performed excavation and loading work as learning data; A construction machine characterized by having a display generation unit that generates display data for displaying the calculated weight of the excavation target from the load calculation unit, or the predicted weight of the excavation target from the load prediction unit, and outputs it to the display device.

Citation Information

Patent Citations

  • Multi-source information fusion-based fullness rate identification method and system

    CN117492023A

  • Intelligent online prediction device and method for effective load of engineering machinery

    CN118643637A

  • Work machine

    JP2018154976A

  • Control device of construction machine

    JP2021055289A

  • Work machine

    JP2024127611A