Construction machine operation prediction method, system, and program

Abstract

The present invention provides a construction machine operation prediction method, system, and program that can predict the operation of a construction machine. [Solution] The construction machinery operation prediction method of the present invention comprises a model generation step S2 in which a prediction model that predicts the operation of the construction machinery based on environmental information that represents the surrounding environment of the construction machinery, operation information that represents the operating state of the construction machinery, and operation information that represents input operations of an input unit that inputs instructions to operate the construction machinery is generated by using corresponding past environmental information, operation information, operation information, and prediction target information that represents the operation of the construction machinery to be predicted; and a prediction step S4 in which the prediction model generated in the model generation step is used to predict the operation of the construction machinery by using the environmental information, operation information, and operation information at the prediction time that is to be predicted.

Description

[Technical Field]

[0001] The present invention relates to a construction machine operation prediction method, a construction machine operation prediction system, and a construction machine operation prediction program for predicting the operation of a construction machine. [Background technology]

[0002] There are various types of construction machines depending on the type of work they are used for, and construction machines are used at various work sites (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2021 / 241716 Summary of the Invention [Problem to be solved by the invention]

[0004] When a construction machine is operated by an operator (user), there is a desire to ensure safety around the construction machine. Furthermore, while efficient cooperation is desired, the construction machine is operated by an operator (user), and there is uncertainty in the operation, making it difficult to predict the operation of the construction machine. Therefore, if the operation of the construction machine can be predicted, the safety (degree of safety) of the predicted operation of the construction machine can be evaluated, and safety can be improved by taking measures according to the evaluation results.

[0005] The present invention has been made in consideration of the above-mentioned circumstances, and its object is to provide a construction machinery operation prediction method, a construction machinery operation prediction system, and a construction machinery operation prediction program that can predict the operation of a construction machine. [Means for solving the problem]

[0006] After extensive investigation, the inventors have found that the above object can be achieved by the present invention described below. That is, a construction machine operation prediction method according to one aspect of the present invention comprises: a model generation step of generating a prediction model that predicts the operation of a construction machine based on environmental information that represents the surrounding environment of the construction machine, operation information that represents the operating state of the construction machine, and operation information that represents input operations of an input unit that inputs instructions to operate the construction machine, by using corresponding past environmental information, operation information, and operation information, and prediction target information that represents the operation of the construction machine that is to be predicted; and a prediction step of predicting the operation of the construction machine by using the environmental information, operation information, and operation information at the time of prediction that is to be predicted to the prediction model generated in the model generation step.

[0007] Such a construction machine operation prediction method generates a prediction model, making it possible to predict the operation of the construction machine.

[0008] In another aspect, in the above-mentioned construction machine operation prediction method, the construction machine is equipped with a work attachment which is a mechanism for performing predetermined work according to the type of construction machine, and the work attachment is equipped with a boom, an arm connected to the tip of the boom, and a bucket swingably attached to the tip of the arm, and the operation information includes the swing speed and swing acceleration of the work attachment, the boom angle of the boom, the arm angle of the arm, and the bucket angle of the bucket.

[0009] Such a construction machine operation prediction method can generate a prediction model using operation information including the swing speed and swing acceleration of the work attachment, the boom angle of the boom, the arm angle of the arm, and the bucket angle of the bucket.

[0010] In another aspect, in the above-mentioned construction machine operation prediction method, the environmental information includes the work position of the construction machine, and the positions of at least one of construction machines and obstacles in the vicinity of the construction machine.

[0011] Such a construction machine operation prediction method can generate a prediction model using environmental information including the work position of the construction machine and the positions of at least one of construction machines and obstacles in the vicinity of the construction machine.

[0012] In another aspect, in the above-mentioned construction machine operation prediction methods, the construction machine is equipped with a work attachment which is a mechanism for performing a predetermined task according to the type of the construction machine, the operation of the construction machine is the operation of the work attachment, and the operation information includes the amount of operation of the work attachment.

[0013] Such a construction machine operation prediction method can generate a prediction model using operation information including the amount of movement of the work attachment.

[0014] In another aspect, in the above-mentioned construction machine operation prediction method, the model generation step generates the prediction model using a Gaussian process regression method, and the prediction step represents the predicted operation of the construction machine, including the uncertainty of the predicted result.

[0015] This construction machinery operation prediction method expresses the predicted construction machinery operation including the uncertainty of the prediction result, so that the accuracy of the prediction result can be recognized and the uncertainty of the construction machinery operation can be recognized by the range of variation.

[0016] In another aspect, in the above-mentioned construction machine operation prediction method, the prediction step represents the predicted operation of the construction machine as time-series data from the time of the prediction.

[0017] Such a construction machine operation prediction method expresses the predicted operation of the construction machine as time-series data from the time of prediction, so that the predicted operation of the construction machine can be recognized over time.

[0018] In another aspect, the above-mentioned construction machine operation prediction method further comprises an interference determination step of determining whether or not the operation of the construction machine will interfere with the operation of another construction machine different from the construction machine that operates automatically according to a preset operation plan, based on the operation of the construction machine predicted in the prediction step and the operation of the operation plan.

[0019] Such a construction machine operation prediction method can predict interference between the operation of a construction machine and the operation of another construction machine.

[0020] In another aspect, the above-mentioned construction machine operation prediction method further includes an operation plan re-creation step of creating an operation plan for the other construction machine so as to avoid interference if it is determined in the interference determination step that interference will occur.

[0021] Such a construction machine operation prediction method creates an operation plan for the other construction machine so as to avoid interference, thereby making it possible to avoid interference.

[0022] A construction machinery operation prediction system according to another aspect of the present invention comprises a model generation unit that generates a prediction model that predicts the operation of a construction machinery based on environmental information that represents the surrounding environment of the construction machinery, operation information that represents the operating state of the construction machinery, and operation information that represents the input operation of an operating lever that inputs instructions to operate the construction machinery, by using corresponding past actual environmental information, operation information, and operation information, and prediction target information that represents the operation of the construction machinery to be predicted; and a prediction unit that predicts the operation of the construction machinery by using the prediction model generated by the model generation unit and the environmental information, operation information, and operation information at the prediction time to be predicted.

[0023] Such a construction machine operation prediction system generates a prediction model, thereby making it possible to predict the operation of a construction machine.

[0024] A construction machine operation prediction program according to another aspect of the present invention is a program for causing a computer to function as the above-described construction machine operation prediction system.

[0025] Such a construction machine operation prediction program generates a prediction model, making it possible to predict the operation of the construction machine. [Effects of the Invention]

[0026] The construction machine operation prediction method, construction machine operation prediction system, and construction machine operation prediction program according to the present invention can predict the operation of a construction machine. [Brief explanation of the drawings]

[0027] [Figure 1] FIG. 1 is a side view showing a hydraulic excavator according to an embodiment. [Figure 2] 1 is a block diagram mainly showing the configuration of a construction machine operation prediction device according to an embodiment. [Figure 3] FIG. 10 is a diagram for explaining prediction target information indicating the operation of a construction machine to be predicted. [Figure 4] 4 is a flowchart showing the operation of the construction machine operation prediction device. [Figure 5] FIG. 10 is a diagram showing an example of a predicted turning speed. [Figure 6] FIG. 10 is a diagram illustrating a prediction result of a turning speed as another example. [Figure 7] 10A and 10B are diagrams for explaining the operation of two first and second construction machines in a modified embodiment. [Figure 8] FIG. 4 is a diagram for explaining the operation plan of the second construction machine. [Figure 9] 1 is a flowchart of the entire process executed by the control system 100 according to the first embodiment. [Figure 10] FIG. 10 is a diagram illustrating a model for calculating the distance between two buckets. [Figure 11] FIG. 10 is a diagram showing a model in which the distance is the distance between the tip coordinate of the bucket of the β machine 1B and the straight line formed by the work attachment of the α machine 1A. [Figure 12] 4 is a flowchart of a calculation process according to a first embodiment. [Figure 13] 4 is a graph showing the transition of the swing speed of the work machine according to the first embodiment. [Figure 14] FIG. 10 is a diagram showing a state in which the work plan of the α machine 1A according to the first embodiment has been changed. [Figure 15] FIG. 1 is a schematic diagram showing the working state of the α machine 1A and the β machine 1B. [Figure 16] 10 is a schematic diagram showing the transition of the rotation speed of the α machine 1A according to the first embodiment. FIG. [Figure 17] FIG. 10 is a diagram showing a state in which the work plan of the α machine 1A according to the second embodiment has been changed. [Figure 18] 10 is a schematic diagram showing the state of operation of the α machine 1A and the β machine 1B according to the second embodiment. FIG. [Figure 19] FIG. 10 is a diagram showing a state in which the work plan of the α machine 1A according to the third embodiment has been changed. [Figure 20] 10 is a schematic diagram showing the operation of the α machine 1A and the β machine 1B according to the third embodiment. FIG. DETAILED DESCRIPTION OF THE INVENTION

[0028] Hereinafter, one or more embodiments of the present invention will be described with reference to the drawings. However, the scope of the invention is not limited to the disclosed embodiments. In addition, components with the same reference numerals in each drawing indicate the same components, and their description will be omitted as appropriate. In this specification, when referring to a general term, a reference numeral without a subscript is used, and when referring to an individual component, a reference numeral with a subscript is used.

[0029] The construction machinery operation prediction system in one embodiment includes a model generation unit that generates a prediction model that predicts the operation of the construction machinery based on environmental information that represents the surrounding environment of the construction machinery, operation information that represents the operating state of the construction machinery, and operation information that represents input operations of an input unit that inputs instructions to operate the construction machinery, by using corresponding past environmental information, operation information, and operation information, and prediction target information that represents the operation of the construction machinery to be predicted; and a prediction unit that predicts the operation of the construction machinery by using the environmental information, operation information, and operation information at the prediction time to be predicted, in the prediction model generated by the model generation unit.

[0030] Below, such a construction machinery operation prediction system, as well as the construction machinery operation prediction method and construction machinery operation prediction program implemented therein, will be described in more detail.The construction machinery operation prediction system may be configured by interconnecting an input / output terminal device that inputs and outputs data, one or more arithmetic processing devices (e.g., server devices) that execute various arithmetic processing, and one or more database devices that store (manage) various data so that they can communicate with each other, or at least some of these input / output terminal devices, one or more arithmetic processing devices, and one or more database devices may be configured as an integrated unit, and interconnected so that they can communicate with the remainder.Here, the construction machinery operation prediction system will be described using an example of a construction machinery operation prediction device in which all of these are integrated.

[0031] The construction machine (construction machine) HS whose operation is predicted by the construction machine operation prediction device PD in the embodiment is a machine for performing a predetermined task, such as a hydraulic excavator, a crane, a wheel roller, etc. Here, as an example, a case will be described in which the construction machine HS is a hydraulic excavator HS, but of course, this construction machine HS may be another type of construction machine.

[0032] FIG. 1 is a side view showing a hydraulic excavator in an embodiment. FIG. 2 is a block diagram mainly showing the configuration of a construction machine operation prediction device in an embodiment. FIG. 3 is a diagram for explaining prediction target information that indicates the operation of a construction machine to be predicted. The horizontal axis of FIG. 3 represents elapsed time, and the vertical axis represents turning speed. FIG. 3 shows a left turn when a right turn is considered positive.

[0033] 1 and 2, the hydraulic excavator HS in this embodiment is a device for performing work that can be performed by an excavator, and includes a lower traveling body 101 having a left traveling crawler 101L and a right traveling crawler (not shown), an upper frame 102 that is rotatably mounted on the lower traveling body 101 and serves as a base, a machinery room 103 that is mounted on the upper frame 102 and that houses an engine such as an internal combustion engine and a hydraulic pump that generates power to drive a work attachment 104 (described below), the work attachment 104 that is mounted on the upper frame 102 so that it can be raised and lowered, and a cab room (operation room) 105 that is equipped with an operating lever CL for inputting commands to operate the hydraulic excavator HS. The upper frame 102, the machinery room 103, the work attachment 104, and the cab room (operation room) 105 together form a so-called upper rotating body.

[0034] The work attachment 104 is a mechanism for performing a predetermined task according to the type of the construction machine HS. In this embodiment, since the work attachment 104 is a hydraulic excavator HS, the work attachment 104 has, for example, a boom 141, an arm 142 connected to the tip of the boom 141, and a bucket 143 swingably attached to the tip of the arm 142. The boom 141 is raised and lowered relative to the upper frame 102 by the extension and retraction of a boom cylinder 144a. The arm 142 swings relative to the boom 141 by the extension and retraction of an arm cylinder 144b. The bucket 143 swings relative to the arm 142 by the extension and retraction of a bucket cylinder 144c. Note that in this embodiment, since the work attachment 104 is a hydraulic excavator HS, the bucket 143 is attached to the tip of the arm 142, but an appropriate type of tip attachment may be attached according to the type of the construction machine HS.

[0035] The control levers CL (CL-1 to CL-4) include a swing control lever CL-1 for inputting an instruction to swing the upper frame 102, a boom control lever CL-2 for inputting an instruction to operate the boom 141, an arm control lever CL-3 for inputting an instruction to operate the arm 142, and a bucket control lever CL-4 for inputting an instruction to operate the bucket 143. The input operation of the swing control lever CL-1 inputs the start and end of a right (counterclockwise) swing and the start and end of a left (clockwise) swing. For example, the start of a swing occurs when the swing control lever CL-1 is tilted from its neutral position, and the end of a swing occurs when the tilted position is returned to the neutral position. The swing angle is measured by a swing angle sensor SN-1 that measures angles. For example, the swing angle is measured with the front of the construction machine HS as the reference 0 [degrees]. A change in the boom angle is input by the boom control lever CL-2. The boom angle is the angle formed between the upper frame 102 and the boom 141 (the angle of the boom 141 from the horizontal). The boom cylinder 144a is provided with a boom cylinder stroke sensor SN-2 that measures its stroke length, and the boom angle is measured based on the detection result of the boom cylinder stroke sensor SN-2. For example, a correspondence relationship between the stroke length of the boom cylinder 144a and the boom angle is stored in advance in the memory unit 5, and the stroke length of the boom cylinder 144a is converted to the boom angle using this correspondence. Alternatively, an angle sensor (boom angle sensor) that measures the boom angle may be provided instead of the boom cylinder stroke sensor SN-2. A change in the arm angle is input by the arm operation lever CL-3. The arm angle is the angle formed between the boom 141 and the arm 142. The arm cylinder 144b is provided with an arm cylinder stroke sensor SN-3 that measures its stroke length, and the arm angle is measured based on the detection result of the arm cylinder stroke sensor SN-3. For example, the correspondence relationship between the stroke length of the arm cylinder 144b and the arm angle is stored in advance in the storage unit 5, and the stroke length of the arm cylinder 144b is converted into the arm angle using this correspondence relationship.Alternatively, an angle sensor (arm angle sensor) that measures the arm angle may be provided instead of the arm cylinder stroke sensor SN-3. A change in bucket angle is input by the bucket operation lever CL-4. The bucket angle is the angle formed between the arm 142 and the bucket 143 (the swing angle of the bucket 143). The bucket cylinder 144c is provided with a bucket cylinder stroke sensor SN-4 that measures its stroke length, and the bucket angle is measured based on the detection result of the bucket cylinder stroke sensor SN-4. For example, a correspondence relationship between the stroke length and bucket angle of the bucket cylinder 144c is stored in advance in the memory unit 5, and the stroke length of the bucket cylinder 144c is converted to a bucket angle using this correspondence relationship. Alternatively, an angle sensor (bucket angle sensor) that measures the bucket angle may be provided instead of the bucket cylinder stroke sensor SN-4. The swing operation lever CL-1, boom operation lever CL-2, arm operation lever CL-3, and bucket operation lever CL-4 are configured with, for example, multi-axis lever switches, multi-axis joysticks, or the like.

[0036] The operating lever CL corresponds to an example of an input section (first input section, construction machine input section) for inputting instructions to operate the construction machine.

[0037] In this embodiment, the construction machine HS is equipped with an environment measurement unit RS and a communication interface CI.

[0038] The environment measurement unit RS is a device that measures the surrounding environment of the construction machine HS. The environment measurement unit RS is attached to the construction machine HS (for example, above the cab room 105) and is a device that measures the external shape of objects that exist within the working area of ​​the work attachment 104, and is configured with, for example, a LiDAR. More specifically, the environment measurement unit RS measures three-dimensional point cloud data that represents the shape of the terrain within the working area of ​​the work attachment 104.

[0039] The communication interface CI is a device for transmitting and receiving communication signals to and from external devices, and is configured to include, for example, a data communication card, a communication interface circuit conforming to the IEEE802.11 standard, and the like.

[0040] The construction machine HS is communicably connected to the construction machine operation prediction device PD via the communication interface CI and the communication network NW. The construction machine HS repeatedly measures the surrounding environment, swing angle, boom cylinder stroke length (or boom angle), arm cylinder stroke length (or arm angle), and bucket cylinder stroke length (or bucket angle) of the construction machine HS using the environment measurement unit RS, swing angle sensor SN-1, boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), and bucket cylinder stroke sensor SN-4 (or bucket angle sensor) in synchronization at predetermined sampling intervals, detects the operation amount of the swing operation lever CL-1 in synchronization with this, and sends these to the construction machine operation prediction device PD.

[0041] The construction machine operation prediction device PD includes, for example, a control processing unit 1, a second input unit 2, an output unit 3, an interface unit (IF unit) 4, and a storage unit 5, as shown in FIG.

[0042] The second input unit (prediction device input unit) 2 is connected to the control processing unit 1 and is a device that inputs various commands, such as a command to start generating a prediction model or a command to start prediction, and various data necessary for operating the construction machinery operation prediction device PD, such as past performance information, to the construction machinery operation prediction device PD, and is, for example, a keyboard, a mouse, or multiple input switches to which predetermined functions are assigned. The output unit 3 is connected to the control processing unit 1 and is a device that outputs the commands, data, prediction results, etc. input from the second input unit 2 under the control of the control processing unit 1, and is, for example, a display device such as a CRT display, LCD (liquid crystal display) or organic EL display, or a printing device such as a printer.

[0043] The second input unit 2 and the output unit 3 may be configured as a touch panel. In the case of configuring this touch panel, the second input unit 2 is a position input device that detects and inputs an operation position, such as a resistive film type or a capacitive type, and the output unit 3 is a display device. In this touch panel, a position input device is provided on the display surface of the display device, and one or more input content candidates that can be input are displayed on the display device. When a user touches the display position showing the input content they want to input, the position is detected by the position input device, and the display content displayed at the detected position is input to the construction machine operation prediction device PD as the user's operation input content. With such a touch panel, the user can easily intuitively understand the input operation, and therefore a construction machine operation prediction device PD that is easy for the user to use is provided.

[0044] The IF unit 4 is connected to the control processing unit 1 and is a circuit that inputs and outputs data to and from, for example, an external device under the control of the control processing unit 1, and is, for example, an interface circuit for RS-232C, which is a serial communication method, an interface circuit using the Bluetooth (registered trademark) standard, an interface circuit using the USB standard, etc. The IF unit 4 is also a communication interface circuit that transmits and receives communication signals to and from external devices, such as a data communication card or a communication interface circuit conforming to the IEEE802.11 standard, etc.

[0045] The storage unit 5 is connected to the control processing unit 1 and is a circuit that stores various predetermined programs and various predetermined data under the control of the control processing unit 1.

[0046] The various predetermined programs include, for example, a control processing program, and the control processing program includes, for example, a control program, a model generation program, and a prediction program. The control program controls each of the sections 2 to 5 of the construction machine operation prediction device PD according to the function of each section. The model generation program is a program that generates a prediction model that predicts the operation of the construction machine HS based on environmental information that represents the surrounding environment of the construction machine HS, operation information that represents the operating state of the construction machine HS, and operation information that represents the input operation of the control lever CL that inputs instructions to operate the construction machine HS, by using corresponding past environmental information, operation information, and operation information, and operation information that represents prediction target information that represents the operation of the construction machine HS that is to be predicted. The prediction program is a program that predicts the operation of the construction machine HS by using the environmental information, operation information, and operation information at the time of prediction that is to be predicted in the prediction model generated by the model generation program.

[0047] The various types of predetermined data include, for example, past performance information including associated past performance environment information, operation information, manipulation information, and prediction target information, various intermediate calculation results, and final calculation results, and other data necessary for executing each of these programs. The storage unit 5 includes, for example, a nonvolatile storage element such as a read-only memory (ROM) or a rewritable nonvolatile storage element such as an electrically erasable programmable read-only memory (EEPROM). The storage unit 5 also includes a random access memory (RAM) that serves as the working memory of the control processing unit 1 and stores data generated during execution of the predetermined programs. The storage unit 5 may also be configured with a hard disk drive or solid-state drive (SSD) with a relatively large storage capacity.

[0048] The storage unit 5 functionally includes a past performance information storage unit 51. The past performance information storage unit 51 stores environmental information, operation information, manipulation information, and prediction target information of past performances that are associated with each other.

[0049] The environmental information is information that represents the surrounding environment of the construction machine HS. More specifically, the environmental information includes the work position of the construction machine, and the positions of at least one of other construction machines and obstacles around the construction machine. For example, the environmental information may include the positions of other construction machines within a predetermined work area when the construction machine HS is stopped. Alternatively, the environmental information may include predetermined obstacles within the predetermined work area when the construction machine HS is stopped. Obstacles include, for example, fences, prefabricated buildings, concrete, wood, soil and asphalt used in work by other construction machines, and soil moldings (e.g., slopes) placed at a construction site. The other construction machines and the predetermined obstacles are each appropriately set in advance. The other construction machines are detected from the measurement results within the work area obtained by the environmental measurement unit RS, for example, by pattern matching, and their coordinates are calculated as environmental information. Similarly, the predetermined obstacles are detected from the measurement results within the work area obtained by the environmental measurement unit RS, for example, by pattern matching, and their coordinates are calculated as environmental information.

[0050] Alternatively, for example, the environmental information includes the vertex positions of a predetermined number of objects in a predetermined working area in descending order of their height when the construction machine HS is stopped. The predetermined number is set in advance as appropriate. For example, when a hydraulic excavator HS is used to move piled earth, the environmental information includes the vertex positions of a predetermined number of piles in descending order of their height. In this embodiment, the environmental information is obtained based on the measurement results of the environmental measurement unit RS. For example, the measurement results of the environmental measurement unit RS are received from the construction machine HS via the communication interface CI and the communication network NW by the construction machine operation prediction device PD, and the construction machine operation prediction device PD uses the control processing unit 1 to determine, as environmental information, the coordinates of a predetermined number of objects in descending order of their height from the measurement results within the working area in the measurement results of the environmental measurement unit RS.

[0051] The operation information is information that represents the operation state of the construction machine HS. In this embodiment, the operation state of the construction machine HS is the operation state of the work attachment 104. More specifically, the operation information includes the swing speed of the upper rotating body (swing speed [rpm, deg / s] around the rotation axis of the upper rotating body), swing acceleration, the boom angle of the boom 141, the arm angle of the arm 142, and the bucket angle of the bucket 143. For example, the measurement results of the swing angle sensor SN-1, boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), and bucket cylinder stroke sensor SN-4 (or bucket angle sensor) are received by the construction machine operation prediction device PD from the construction machine HS via the communication interface CI and the communication network NW, as described above. The construction machinery operation prediction device PD determines the boom angle, arm angle, and bucket angle based on the measurement results of the boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), and bucket cylinder stroke sensor SN-4 (or bucket angle sensor) using the control processing unit 1. The swing speed is determined by measuring the swing angle at a predetermined sampling interval using the swing angle sensor SN-1 and dividing the change in swing angle between samples by the sampling interval, and the swing acceleration is determined by differentiating the swing speed with respect to time. Alternatively, the swing speed may be measured using, for example, an inertial measurement unit (IMU) (not shown). The operating state of the work attachment 104 can be determined by determining the swing speed, swing acceleration, boom angle, arm angle, and bucket angle.

[0052] The operation information is information representing the input operation of the control lever CL, which inputs instructions to operate the construction machine HS. More specifically, the operation information includes the movement amount of the work attachment 104. In this embodiment, the prediction model is a model that predicts the swing operation of the work attachment 104, and therefore the operation information is the movement amount of the swing operation lever CL-1, in other words, the swing pilot pressure corresponding to the movement amount of the work attachment 104. Alternatively, for example, the operation information may be an electrical signal representing the swing pilot pressure. As described above, the movement amount of the swing operation lever CL-1 is received by the construction machine operation prediction device PD from the construction machine HS via the communication interface CI and the communication network NW.

[0053] In this embodiment, as described above, the prediction model is a model that predicts the swing operation of the work attachment 104, and so the operation information is the amount of operation of the swing operation lever CL-1, but is not limited to this and may be the amount of operation of the boom operation lever CL-2, the amount of operation of the arm operation lever CL-3, the amount of operation of the bucket operation lever CL-4, etc., depending on the model.

[0054] The measurement results of the environment measurement unit RS, swing angle sensor SN-1, boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), bucket cylinder stroke sensor SN-4 (or bucket angle sensor), and the operation amount of the swing operation lever CL-1 were obtained at the same timing by operating the construction machine HS in the past, and as described above, environmental information and operation information were obtained, and the environmental information, operation information, and operation information from the same timing are associated with each other and stored in the memory unit 5. The past performance information memory unit 51 stores multiple pieces of such associated environmental information, operation information, and operation information from past performances.

[0055] The prediction target information is information that represents the operation of the construction machine HS that is to be predicted. In this embodiment, the operation of the construction machine HS that is to be predicted is time-series data from the time of prediction, for example, the time transition of the swing speed. This time transition of the swing speed is represented by, for example, a rise delay time L1, a rise time σ1, a steady value of the swing speed (target swing speed) r1, a fall delay time L2, and a fall time σ2, as shown in FIG. 3 . The rise delay time L1 is the delay time from the timing when the swing operation lever CL-1 is input to start the swing (start operation timing) to the actual start of the swing. The rise time σ1 is the time from the timing when the swing actually starts (swing start timing) to the time when the swing speed reaches 63.2% of the steady value r1 of the swing speed (after the rise delay time L1). The fall delay time L2 is the delay time from the timing when the swing operation lever CL-1 is input to end the swing (end operation timing) to the time when the swing speed actually starts to decelerate. The fall time σ2 is the time from the timing when the deceleration of the turning speed actually starts (deceleration start timing) (after the fall delay time L2) until the turning speed reaches 36.8[%] of the steady value r1 of the turning speed (the time until the turning speed decelerates from the steady value r1 of the turning speed to 63.2[%] of that value. These rise delay time L1, rise time σ1, steady value of the turning speed (target turning speed) r1, fall delay time L2 and fall time σ2 are determined by manual operation by the user or by automatic calculation by the control processing unit 1, and are used to determine the turning speed. The sampling timing when the operation amount of the turning lever CL-1 exceeds a predetermined threshold is defined as the start operation timing, and the turning speed for a right turn is defined as positive and the turning speed for a left turn is defined as negative. The environmental information, motion information, and operation information sampled at each sampling timing from when the absolute value of the turning speed exceeds the threshold until when it falls below the threshold again are associated with a rise delay time L1, a rise time σ1, a steady-state value of the turning speed (target turning speed) r1, a fall delay time L2, and a fall time σ2, and are stored in the past performance information storage unit 51.If the timing when the absolute value of the turning speed exceeds the threshold and then falls below the threshold again is defined as the turning end timing, the environmental information, motion information, and operation information sampled at each sampling timing from the start operation timing to the turning end timing are each associated with the same values ​​of rise delay time L1, rise time σ1, steady-state value of turning speed (target turning speed) r1, fall delay time L2, and fall time σ2.

[0056] In this embodiment, the environmental information, operation information and operation information of the past performance are received by the construction machinery operation prediction device PD from the construction machinery HS via the communication interface CI and the communication network NW, but the management server device that manages the environmental information, operation information and operation information of the past performance may receive, store and accumulate the environmental information, operation information and operation information from the construction machinery HS, for example, via a communication network, and the construction machinery operation prediction device PD may receive the environmental information, operation information and operation information of the past performance from the management server device, for example, via a communication network, and store it in the past performance information storage unit 51. Alternatively, the construction machine operation prediction device PD may be mounted on the construction machine HS, and the control processing unit 1 may be connected to the environment measurement unit RS, swing angle sensor SN-1, boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), bucket cylinder stroke sensor SN-4 (or bucket angle sensor), and swing operation lever CL-1, and may obtain the measurement results of the environment measurement unit RS, swing angle sensor SN-1, boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), and bucket cylinder stroke sensor SN-4 (or bucket angle sensor), as well as the operation amount of the swing operation lever CL-1, and store and accumulate these in the past performance information memory unit 51.

[0057] Alternatively, in the above description, the environmental information, swing speed, swing acceleration, boom angle, arm angle and bucket angle are determined by the construction machinery operation prediction device PD, but they may also be determined by the construction machinery HS and received by the construction machinery operation prediction device PD from the construction machinery HS via the communication interface CI and the communication network NW.

[0058] The control processing unit 1 is a circuit that controls each of the units 2 to 5 of the construction machine operation prediction device PD according to the function of each unit, generates a prediction model based on past performance data, and predicts the operation of the construction machine HS. The control processing unit 1 is configured, for example, with a CPU (Central Processing Unit) and its peripheral circuits. When the control processing program is executed, the control processing unit 1 is functionally configured to include a control unit 11, a model generation unit 12, a prediction unit 13, and an interference determination unit 14.

[0059] The control unit 11 controls each of the units 2 to 5 of the construction machine operation prediction device PD in accordance with the function of each unit, and is in charge of overall control of the construction machine operation prediction device PD.

[0060] The model generation unit 12 generates a prediction model by using the environmental information, operation information, manipulation information, and prediction target information of past actual results that are associated with each other and stored in the past actual result information storage unit 51. The prediction model is a model that predicts the operation of the construction machine HS (predicted value of the prediction target information) based on the environmental information, operation information, and manipulation information. In this embodiment, the model generation unit 12 generates the prediction model using a Gaussian process regression method.

[0061] The prediction unit 13 predicts the behavior of the construction machine HS by using the environmental information, operation information, and manipulation information at the time of prediction to be made in the prediction model generated by the model generation unit 12. The prediction unit 13 expresses the predicted behavior of the construction machine HS, including the uncertainty of the prediction result. More specifically, in this embodiment, the prediction unit 13 expresses the predicted behavior of the construction machine HS using the degree of dispersion (for example, variance).

[0062] More specifically, the number of items of environmental information, behavioral information, operation information, and prediction target information in the past performance for generating a prediction model is k, and a vector having elements of the environmental information, behavioral information, and operation information for the kth item is denoted by x (k) Let y be the vector whose element is the k-th prediction target information. (k) Then (k=1, 2, . . . , m), the following equation 1 is formed, and the matrix element of n rows and n' columns is expressed by the following equation 2. For example, if there are seven pieces of past performance information, equation 1 becomes the following equation 3 (m=7).

[0063]

number

[0064]

number

[0065]

number

[0066] Here, N is a symbol that indicates that a Gaussian distribution (normal distribution) is used as the kernel function. * is a vector whose elements are the environment information, the motion information, and the operation information at the time of prediction. * is a vector whose elements are the top position of the pile of soil, the swing speed, the swing acceleration, the boom angle of the boom 141, the arm angle of the arm 142, the bucket angle of the bucket 143, and the operation amount of the swing operation lever CL-1 at the time of prediction. * In this embodiment, the elements of the vector y are five elements: the rise delay time L1, the rise time σ1, the steady value r1 of the turning speed, the fall delay time L2, and the fall time σ2. *is any one of the rise delay time L1, rise time σ1, steady value r1 of the turning speed, fall delay time L2, and fall time σ2. Therefore, for the rise delay time L1, Equation 1 and Equation 2 are generated, and a prediction model for predicting the rise delay time L1 is generated; for the rise time σ1, Equation 1 and Equation 2 are generated, and a prediction model for predicting the rise time σ1 is generated; for the steady value r1 of the turning speed, Equation 1 and Equation 2 are generated, and a prediction model for predicting the steady value r1 of the turning speed is generated; for the fall delay time L2, Equation 1 and Equation 2 are generated, and a prediction model for predicting the fall delay time L2 is generated; and for the fall time σ2, Equation 1 and Equation 2 are generated, and a prediction model for predicting the fall time σ2 is generated.

[0067] The element y surrounded by the dashed line of the vector (vertical vector) y' in Equation 1 (1) ···y (m) Let y be the vector (vertical vector) (note that y (k) (The matrix in Equation 1 is also a vector as mentioned above.) The part of the matrix in Equation 1 from row 1, column 1 to row m, column m is the matrix K. The part of the matrix in Equation 1 from row m+1, column 1 to row m+1, column m+1 is the transposed vector k * T (superscript T is vector (vertical vector) k * (The symbol θ represents the transpose of θ), vector y, and matrix K are respectively substituted with past performance information on the prediction target, environmental information, motion information, and operation information, and the hyperparameters θ1, θ2, and θ3 in Equation 2 are found using, for example, well-known solution methods such as the gradient method or the Markov chain Monte Carlo method, to generate each prediction model.

[0068] Vector x * , the probability p(y * |x * , x, y) is given by Equation 4, and k in Equation 4 * T K -1 y is the formula for calculating the predicted value, and k(x * , x *)-k * T K -1 k * is the formula for calculating the variance of this predicted value. This gives the predicted value and variance of the operation of the construction machine HS.

[0069]

number

[0070] The environmental information can provide the rotation start position (rotation start angle) and rotation end position (rotation end angle) of the upper rotating body, and the rotation end position (rotation end angle) may be used for prediction of the prediction model. For example, in the case of the above-mentioned work of excavating soil from a pile of earth piled up to the fourth position PA', rotating from the fourth position PA' to the third position PB', discharging the soil to the third position PB', rotating from the third position PB' to the fourth position PA', and returning to the fourth position PA', the environmental information will include the peak position of the pile of earth piled up to the fourth position PA' and the peak position of the pile of earth piled up to the third position PB' (see FIG. 6 described later). The rotation start position (rotation start angle) and rotation end position (rotation end angle) of the upper rotating body may be input from the second input unit 2.

[0071] The control unit 11 outputs the operation of the construction machine HS predicted by the prediction unit 13 to the output unit 3.

[0072] The control processing unit 1, second input unit 2, output unit 3, IF unit 4 and storage unit 5 can be configured by, for example, a desktop or notebook computer.

[0073] Next, the operation of this embodiment will be described. Fig. 4 is a flowchart showing the operation of the construction machine operation prediction device. Fig. 5 is a diagram showing the predicted results of the swing speed, as an example. The horizontal axis of Fig. 5 is elapsed time, and the vertical axis is swing speed. Fig. 6 is a diagram for explaining the predicted results of the swing speed, as another example. Fig. 6A shows the positional relationship between the construction machine and an obstacle, and Fig. 6B shows the predicted results of the swing speed, as another example. The horizontal axis of Fig. 6B is elapsed time, and the vertical axis is swing speed.

[0074] When the construction machinery operation prediction device PD configured as above is powered on, it initializes the necessary parts and starts operation. The control processing unit 1 is functionally configured with a control unit 11, a model generation unit 12, and a prediction unit 13 by executing the control processing program. It is assumed that the past performance information storage unit 51 stores environment information, operation information, manipulation information, and prediction target information of past performances that are associated with each other.

[0075] 4, when an instruction to generate a prediction model is received from the user at the second input unit 2 (S1), the construction machinery operation prediction device PD causes the model generation unit 12 of the control processing unit 1 to generate a prediction model by using the environmental information, operation information, manipulation information, and prediction target information from past performance that are stored in correspondence with each other in the past performance information memory unit 51, and stores the generated prediction model in the memory unit 5 (S2). Note that in process S1, instead of the user inputting an instruction to generate a prediction model, the environmental information, operation information, and manipulation information from past performance may be accumulated, and when a predetermined amount of each information has been accumulated, the generation of the prediction model may be started automatically.

[0076] Next, the construction machine operation prediction device PD determines whether or not prediction has started using the prediction unit 13 of the control processing unit 1 (S3). For example, if the prediction unit 13 receives an instruction to start prediction from the user via the second input unit 2, it determines that prediction has started, and if the instruction to start prediction has not been received, it determines that prediction has not started. If the result of the determination is that prediction has started (Yes), the prediction unit 13 then executes process S4. On the other hand, if the result of the determination is that prediction has not started (No), the prediction unit 13 returns the process to process S3. Therefore, the prediction unit 13 repeatedly executes process S3 until it determines that prediction has started.

[0077] In process S4, the construction machine operation prediction device PD predicts the operation of the construction machine HS at the time of prediction. More specifically, the timing when execution of process S4 starts is set as the prediction time, and the construction machine operation prediction device PD transmits a signal requesting each piece of data (data request signal) to the construction machine HS via the control unit 11 of the control processing unit 1. The construction machine HS repeatedly measures the surrounding environment, swing angle, boom cylinder stroke length (or boom angle), arm cylinder stroke length (or arm angle), and bucket cylinder stroke length (or bucket angle) of the construction machine HS in synchronization at predetermined sampling intervals using the environment measurement unit RS, swing angle sensor SN-1, boom cylinder stroke sensor SN-2 (or boom angle sensor), arm cylinder stroke sensor SN-3 (or arm angle sensor), and bucket cylinder stroke sensor SN-4 (or bucket angle sensor), and in synchronization with this, measures the operation amount of the swing operation lever CL-1, the operation amount of the boom operation lever CL-2, the arm operation Assuming that the operation amount of the lever CL-3 and the operation amount of the bucket operation lever CL-4 are detected, the construction machine HS transmits (replies) to the construction machine operation prediction device PD a signal (data notification signal) containing the surrounding environment, swing angle, boom cylinder stroke length (or boom angle), arm cylinder stroke length (or arm angle), bucket cylinder stroke length (or bucket angle), operation amount of the swing operation lever CL-1 obtained at the sampling timing (most recent sampling timing) immediately prior to the timing of receiving this data request signal, and the swing angle obtained at the sampling timing immediately prior to this most recent sampling timing. Upon receiving the data notification signal, the construction machine operation prediction device PD determines the environmental information, operation information, and operation information at the time of prediction based on the data contained in this data notification signal, and predicts the operation of the construction machine HS by using this obtained environmental information, operation information, and operation information at the time of prediction in a prediction model. For example, predicted values ​​of the rise delay time L1, rise time σ1, steady-state value r1 of the turning speed, fall delay time L2, and fall time σ2, as well as the variances of each of these predicted values, are calculated, and the time transition of the turning speed shown in Figure 5 is predicted.In FIG. 5, the solid line is a graph of true values, the dashed line is a graph of predicted values, the one-dot-dash line is a graph of values ​​obtained by adding variance to the predicted values, and the two-dot-dash line is a graph of values ​​obtained by subtracting variance from the predicted values. Alternatively, for example, as shown in FIG. 6A, if the work attachment 104 is rotated while gripping soil from a pile of earth at position PA', the work attachment 104 will interfere with (collide with) an obstacle BA, and the work attachment 104 will rotate along the trajectory TR shown by the dashed line in FIG. 6A. In this case, prediction models corresponding to such trajectory TR are created based on past performance data, and the time transition of the rotation speed shown in FIG. 6B is predicted. In FIG. 6B, the solid line is a graph of predicted values, the one-dot-dash line is a graph of values ​​obtained by adding variance to the predicted values, and the two-dot-dash line is a graph of values ​​obtained by subtracting variance from the predicted values.

[0078] Returning to Fig. 4, the construction machine operation prediction device PD then causes the control unit 11 of the control processing unit 1 to output the predicted operation of the construction machine HS to the output unit 3 (S5), and the predicted operation of the construction machine HS is output to the outside by the output unit 3. For example, a graph of the time progression of the swing speed shown in Fig. 5 is displayed on the output unit 3. Note that the control unit 11 may output the predicted operation of the construction machine HS to an external device via the IF unit 4 as necessary.

[0079] Next, the construction machine operation prediction device PD determines whether or not this process has ended using the control unit 11 of the control processing unit 1 (S6). For example, if the control unit 11 receives an instruction to end prediction from the user via the second input unit 2, it determines that prediction has ended, and if the instruction to end prediction has not been received, it determines that prediction has not ended. If the result of the determination is that prediction has ended (Yes), the control unit 11 ends this process. On the other hand, if the result of the determination is that prediction has not ended (No), the control unit 11 returns the process to step S4.

[0080] As described above, the construction machine operation prediction system in the embodiment (one example of which is the construction machine operation prediction device PD) and the construction machine operation prediction method and construction machine operation prediction program implemented therein generate a prediction model, making it possible to predict the operation of the construction machine HS. The construction machine operation prediction system (one example of which is the construction machine operation prediction device PD), construction machine operation prediction method, and construction machine operation prediction program described above can create a prediction model that includes uncertainties that suit the worker, and can calculate a work plan for a cooperative machine that is suitable for the worker.

[0081] The above-mentioned construction machinery operation prediction system (one example of which is the construction machinery operation prediction device PD), construction machinery operation prediction method, and construction machinery operation prediction program can generate a prediction model using operation information including the rotation speed and rotation acceleration of the work attachment 104, the boom angle of the boom 141, the arm angle of the arm 142, and the bucket angle of the bucket 143.

[0082] The above-mentioned construction machinery operation prediction system (one example of which is the construction machinery operation prediction device PD), construction machinery operation prediction method, and construction machinery operation prediction program can generate a prediction model using environmental information including the working position of the construction machinery and the positions of at least one of construction machinery and obstacles around the construction machinery.

[0083] The above-mentioned construction machine operation prediction system (one example of which is the construction machine operation prediction device PD), construction machine operation prediction method, and construction machine operation prediction program can generate a prediction model using operation information including the operation amount of the work attachment 104.

[0084] The above-mentioned construction machinery operation prediction system (one example of which is the construction machinery operation prediction device PD), construction machinery operation prediction method, and construction machinery operation prediction program are expressed as time series data from the time of prediction, so that the predicted operation of the construction machinery HS can be recognized over time.

[0085] The above-mentioned construction machinery operation prediction system (one example of which is the construction machinery operation prediction device PD), construction machinery operation prediction method, and construction machinery operation prediction program represent the predicted operation of the construction machinery HS, including the uncertainty of the prediction result, so that the certainty of the prediction result can be recognized, and the uncertainty of the construction machinery operation can be recognized by the range of variation.

[0086] In the above-described embodiment, the construction machine operation prediction device PD may further include an interference determination unit 14 in the control processing unit 1, as shown by the dashed line in FIG. 1 , that determines whether or not the operation of a construction machine (first construction machine) HS1 will interfere with the operation of a different construction machine (second construction machine) HS2 that is different from the first construction machine HS1 and operates automatically according to a preset operation plan, based on the operation of the first construction machine HS1 predicted by the prediction unit 13 and the operation of the operation plan (in this case, it can also be considered an interference prediction method). The interference refers to the operation of one construction machine HS interfering with the operation of the other construction machine HS, i.e., at least one of the construction machines HS cannot continue its operation as is, for example, when they approach each other to within a predetermined distance. In this modified embodiment, the construction machine operation prediction device PD can predict interference between the operation of the first construction machine HS1 and the operation of the second construction machine HS2.

[0087] Fig. 7 is a diagram for explaining the state of operation of two construction machines, a first and a second, in a modified embodiment, and Fig. 8 is a diagram for explaining the operation plan of the second construction machine.

[0088] For example, as shown in FIG. 7, first and second construction machines HS1 and HS2 cooperate to transport soil to the bed of truck DC. Truck DC is parked with its bed located at first position PA. The first construction machine HS1 is located at the center between fourth position PA' and third position PB', and is manually operated by an operator (first operator (e.g., a worker)) to excavate soil from the pile of soil piled up at fourth position PA', swing from fourth position PA' to third position PB', discharge the soil at third position PB', swing from third position PB' to fourth position PA', and return to fourth position PA'. This process is then repeated. The second construction machine HS2 is located at the center between second position PB and first position PA, and is automatically operated according to the motion plan. As shown in FIG. 8, the motion plan involves sequentially swinging from a first position PA to a second position PB, excavating a pile of earth at the second position PB, swinging from the second position PB to the first position PA, and discharging earth at the first position PA, and these steps are then repeatedly performed. The third position PB' and the second position PB are close to each other, and a pile of earth has been formed between the third position PB' and the second position PB. The time progression of the swing speed is set appropriately in advance, and the swing operation, excavation operation, and earth discharge operation are each appropriately programmed in advance. Therefore, the movement trajectory of the work attachment of the second construction machine HS2 over time can be determined in advance, and the operation times required for swing operation, excavation operation, and earth discharge operation can also be determined in advance. The second construction machine HS2 receives an instruction to start executing the motion plan from an operator (a second operator (e.g., a site supervisor, etc.)), and executes the motion plan at that timing. Therefore, the position of the bucket of the work attachment of the second construction machine HS2 at a certain timing (for example, the central position of the bucket) can be obtained.

[0089] Here, the operation of the first construction machine HS1 is predicted at the prediction time point, and the presence or absence of interference from the prediction time point onwards is determined. As explained using Figure 4, the time transition of the swing speed of the first construction machine HS1 is predicted as shown in Figure 5.

[0090] When the change in the swing speed of the first construction machine HS1 over time is predicted, the interference determination unit 14 determines the movement trajectory of the work attachment 104 of the first construction machine HS1 over time based on this predicted change in the swing speed over time, and determines the positions of the bucket 143 of the first construction machine HS1 at each timing at preset time intervals. Next, the interference determination unit 14 determines the positions of the bucket of the second construction machine HS2 at each of these timings. Based on the positions of the bucket 143 of the first construction machine HS1 at each of these timings and the positions of the bucket of the second construction machine HS2 at each of the aforementioned timings, the interference determination unit 14 determines the Euclidean distances between the bucket 143 of the first construction machine HS1 and the bucket of the second construction machine HS2 at each of the aforementioned timings. The interference determination unit 14 determines whether any of the Euclidean distances at each of the aforementioned timings is equal to or less than a threshold value that is appropriately set in advance. The interference determination unit 14 determines that there is interference if there is a Euclidean distance equal to or less than the threshold, and determines that there is no interference if there is no Euclidean distance equal to or less than the threshold. When it determines that there is interference, the interference determination unit 14 outputs that there is interference to the output unit 3. The output unit 3 displays, for example, a message indicating that there is interference, and issues a warning that there is interference.

[0091] Furthermore, in the above-described modified embodiment, the interference determination unit 14 may further create an operation plan for the second construction machine HS2 so as to prevent interference when it is determined that interference will occur (in this case, it can also be considered as a method for regenerating an operation plan). According to this, an operation plan for the other construction machine is created so as to prevent interference, so that interference can be avoided.

[0092] In the following description, the α machine 1A corresponds to the second construction machine HS2 described above, and the β machine 1B corresponds to the first construction machine HS1 described above. FIG. 9 is a flowchart of the entire process executed by the control system 100 according to the first embodiment. For example, when a predetermined start button located in the cab 105 is pressed, the cooperative operation process of the α machine 1A and the β machine 1B is started. The start button is not limited to being located in the cab 105, but may be provided on a terminal (PC, tablet, smartphone) or remote control device located outside the cab 105. As an example, the bucket 143 of the α machine 1A is located at A1, and the bucket 143 of the β machine 1B is located at C1 by an operator in advance. Note that the initial positions of the buckets 143 are not limited to those located by the operator.

[0093] When the process starts, the operation plan of the α machine 1A is acquired (step S1A). In addition, the operation of the β machine 1B (the coordinated machine) is predicted using the most recent data of the β machine 1B (step S1B). Next, a decision on whether to re-plan is made (step S2A). For example, using the acquired time-series data and model, the operation in the near future is predicted from the most recent operation results, and it is determined whether re-planning is necessary.

[0094] If it is determined that re-planning is not necessary (NO in step S2A), the process proceeds to step S3A, where the current parameters such as σ, L, and r1 are maintained and the process ends. On the other hand, if it is determined that re-planning is necessary (YES in step S2A), the process proceeds to step S4A.

[0095] In step S4A, a parameter search is performed. Here, multiple sets of parameters (σ1, r1) are calculated. Here, as an example, it is assumed that 100 sets of parameters are set. Then, using Equation 1A described below, the possibility of interference is evaluated for each parameter set (step S5A).

[0096] The evaluation of step S5A will be described in detail below. Based on each parameter set, the distance D(t) between the buckets 143 of the α machine 1A and the β machine 1B in the simulation under replanning is calculated at time t t is used as a variable. Furthermore, a magnitude relationship between a distance threshold D_lim stored in advance in the storage unit 5 and the minimum value of the distance D(t) is compared. Note that D_lim is a threshold set to avoid contact and interference between the buckets 143. This value may be set by the operator.

[0097] Fig. 10 is a diagram showing a model for calculating the distance between two buckets 143. In Fig. 10, two black circles represent the centers of rotation of the upper rotating bodies of the hydraulic excavators, and two white circles represent the tips of the buckets 143. In this embodiment, the distance D(t) between the two buckets can be calculated based on the known Euclidean norm.

[0098] On the other hand, Fig. 11 is a diagram showing a model in which the distance D(t) is the distance between the tip coordinate of the bucket 143 of the β machine 1B and the straight line formed by the work attachment 104 of the α machine 1A. In this case, the distance D(t) can be expressed by the following equation 1A. Note that in equation 1A, the symbols α and β of the hydraulic excavators (α machine 1A and β machine 1B) are added to the coefficients of the X coordinate, Z coordinate, and straight line equations in the figure, respectively.

[0099]

number

[0100] Of the D(t) calculated by each of the above two models, the smaller one is defined as Min(t). When Min(D(t)) < D_lim, if the parameter to be evaluated is selected, the minimum value of the distance between the buckets 143 will be smaller than the threshold value, and eventually, there is a possibility that the two working machines will interfere. Therefore, the Flag for control switching is set to 1, and this parameter is excluded from the candidates. On the other hand, when D_lim ≤ Min(D(t)), if the parameter to be evaluated is selected, the minimum value of the distance between the buckets 143 will be greater than or equal to the threshold value, and the two working machines can perform their operations without interference. Therefore, the Flag is set to 0, and this parameter is retained as a candidate parameter.

[0101] The above evaluation is performed for each parameter set. When the evaluation for the set number of times is completed (YES in step S6A), among them, for example, considering other conditions such as the end time of a specific operation that the α machine 1A should satisfy, the optimal parameter is selected, and σ, L, and r1 are updated, thereby changing the operation plan (parameter change) of the α machine 1A (step S7A). Specifically, several parameters that can complete the operation for a predetermined time while keeping Flag = 0 are calculated, and each parameter is evaluated by Equation 2A, which will be described in detail later. Then, the one with the best evaluation value is selected. If the evaluation for the set number of times has not been completed in step S6A (NO in step S6A), the process returns to step S4A and is repeated.

[0102] Next, in the operation plan of the latest α machine 1A, it is determined whether the magnitude of the delay time L0 is greater than a preset time threshold L W (step S8A). The delay time L0 corresponds to the time (waiting time) when the start time of the operation of the α machine 1A is delayed in order to avoid interference between the α machine 1A and the β machine 1B. Also, the time threshold L W is a threshold set to determine whether the α machine 1A can additionally perform another operation that was not included in the original operation plan. In the present embodiment, as this other operation, the operation of the α machine 1A to loosen (stir) the earth and sand at the point A1, as described later, is exemplified.

[0103] In step S8A, L W <When L < L0 (YES in step S8A), a work plan regarding the above-mentioned other work is additionally obtained from the storage unit 5, and by combining it with the locally modified (delayed) work plan, the work plan of the final alpha machine 1A is changed (step S9A). Also, in step S8A, when L0 ≦ L W (NO in step S8A), the process ends without adding the above-mentioned other work.

[0104] When the above process is completed, the latest work plan is input to the alpha machine 1A, and the collaborative work is started. At this time, since the work plan of the alpha machine 1A is corrected as necessary so as to prevent interference between the buckets 143 of the alpha machine 1A and the beta machine 1B, it becomes possible to perform work efficiently and safely with a plurality of hydraulic excavators at the work site.

[0105] FIG. 12 is a flowchart of the arithmetic processing according to the first aspect. In step 7A of FIG. 9, the case where the work plan of the alpha machine 1A is changed, that is, the values of the acceleration / deceleration speed σ, the acceleration / deceleration time L, and the maximum speed r1 are updated will be described in detail.

[0106] As shown in FIG. 12, each parameter included in the work plan of the alpha machine 1A is obtained. At this time, the work plan of the alpha machine 1A includes information regarding the target angle r θ (t) of the bucket 143 accompanying the turning operation of the upper swing body. As described above, this information may be stored as teaching information by the operator operating the alpha machine 1A in advance.

[0107] Furthermore, as described above, the threshold value D_lim for preventing interference between the buckets 143 of the α machine 1A and the β machine 1B is input from the storage unit 5. Furthermore, an operation plan request signal for starting the processing of FIG. 9 is input. As described above, this signal is transmitted, for example, by pressing a switch button, and is also transmitted every time the bucket 143 of the α machine 1A reaches A1, B1, and the processing of FIG. 9 is repeated. Furthermore, information regarding the actual swing angle of the α machine 1A that is obtained is also input.

[0108] After predetermined calculations, the values ​​of acceleration σ1, deceleration σ2, maximum speed r1, acceleration start time L1, and deceleration start time L2 are updated and input to the controllers of the α units A1 and B1, and the upper rotating body is controlled at the target swing speed r. Fig. 13 is a graph showing the transition of the swing speed r of the α unit 1A (hydraulic excavator) determined by these parameters.

[0109] In response to the above parameter settings, a command signal for controlling the movement of the upper rotating body and the work attachment 104 is input to the controller of the α unit 1A based on known PI (Proportional Integral) control. At this time, the controller functions as a responsive controller using the acceleration / deceleration rates σ1 and σ2 as inputs. The above command signal corresponds to the operation lever signal (pilot pressure: Pi pressure) input from the operation unit when the α unit 1A is operated manually. In response to this signal, a valve command value is input to the drive unit to change the flow rate and destination of hydraulic oil supplied to the hydraulic actuator from a control valve included in the drive unit (not shown) of the upper rotating body and the work attachment 104. As a result, the α unit 1A can operate in accordance with the latest work plan.

[0110] In this aspect, as an example, a known real-valued GA (RGA: Real-coded Genetic Algorithms) is used as a method for optimization calculation. Equation 2A shows the evaluation function in this real-valued GA. By using this evaluation function, the work plan of the α machine 1A can be completed as soon as possible without the α machine 1A and the β machine 1B interfering with each other during operation. Note that this evaluation function will be described in detail later.

[0111] [Number]

[0112] The smaller this evaluation function is, under the condition of D_lim≦Min(D(t)), parameters are set such that the working efficiency of the α machine 1A is improved. Note that this evaluation function becomes J = 1 under the condition of Min(D(t)) < D_lim. In Equation 2A, Min(D(t)) is represented by D K (k).

[0113] In Equation 2A, r θ (t) means the arrival (target) turning angle of the α machine 1A. For example, in A1, r θ (t) is 0 degrees, and in B1, r θ (t) is 180 degrees. Also, θ α (t) means the time-series turning angle of the α machine 1A. This value corresponds to the result reflecting the search parameters L1, L2, r, σ1, and σ2. Furthermore, D(t) means the distance between the buckets 143 of the α machine 1A and the β machine 1B. Also, as described above, D_lim is a threshold value set to avoid contact and interference between the buckets 143. This value may be set by an operator.

[0114] In the above calculation, the information of the prediction of the β machine 1B is included in the condition of "Min(D(t)) < D_lim", and a new (revised) work plan of the α machine 1A is created under the consideration of this prediction. Depending on the parameters obtained by this optimization calculation, the movements of the α machine 1A and the β machine 1B are roughly classified into a plurality of patterns to be described later.

[0115] FIG. 14 is a diagram showing a state where the work plan of the α machine 1A according to the first aspect is changed. FIG. 15 is a schematic diagram showing the working states of the α machine 1A and the β machine 1B. FIG. 16 is a schematic diagram showing the transition of the turning speed of the α machine 1A according to the first aspect.

[0116] Here, when the optimized L1 is set to L0, in this embodiment, as shown in FIG. 14, the operation of the α machine 1A for excavating the soil at point A1 (holding it in the bucket 143) is delayed by the delay time L0 seconds, thereby preventing interference between the α machine 1A and the β machine 1B. In particular, in the time zone indicated by the oblique lines in FIG. 14, as shown in FIG. 15, after the β machine 1B leaves from point B2 to the C1 side, the α machine 1A approaches point B1. As a result, interference between the two machines can be stably prevented.

[0117] Furthermore, in this aspect, regarding the magnitude of the above delay time L0, as a result of the evaluation in step S8A of FIG. 9, L W < L0, so that the stirring operation in the A1 excavation area is added as a separate operation in the same step S9A (see FIG. 14).

[0118] Regarding the above, when the work plan of the α machine 1A is updated to a plan coordinated with the β machine 1B, waiting time occurs in the operation of the α machine 1A in the work procedure as per the prior plan. Therefore, by performing in advance an operation that can be executed from the subsequent work processes during this waiting time, a decrease in productivity can be prevented.

[0119] Here, the transition of the turning speed of the α machine 1A after the work plan is changed as described above will be described with reference to FIG. 16. In FIG. 16, the delay time L0 in FIG. 14 is shown as L 11 That is, L 11 is the waiting (separate operation possible) time before excavation. Also, L 12 is the turning deceleration timing when moving from A1 to B1. r1 is the turning speed when moving from A1 to B1. σ 11 represents the rise time to the target turning speed when moving from A1 to B1, and σ11 The smaller the value, the faster the target value is reached. 12 is the fall time from the target turning speed when going from A1 to B1. Similarly, L 21 is the waiting time (other work possible) when returning from B1 to A1, and L 22 is the timing of deceleration when returning from B1 to A1. Also, r2 is the turning speed when returning from B1 to A1. Furthermore, σ 21 is the rise time to the target turning speed when returning from B1 to A1, and σ 22 is the time it takes to fall from the target turning speed when returning from B1 to A1. In addition, since the turning direction is different when going from A1 to B1 and when returning from B1 to A1, the sign of the speed on the vertical axis is distinguished as positive or negative, as shown in Figure 16.

[0120] In the optimization using the evaluation function mentioned above, this delay time L 11 In addition, the transition of the turning speed when turning from point A1 to B1 (each parameter r1, σ 11 , σ 12 , L 12 ), the transition of the turning speed when turning from point B1 to A1 (each parameter r2, σ 21 , σ 22 , L 22 ) are optimized. In this case, the times required for the excavation and earth removal operations at points A1 and B1 are set as fixed values ​​based on a prior operation plan, etc.

[0121] FIG. 17 is a diagram showing a state in which the work plan of the α machine 1A according to the second embodiment has been changed. FIG. 18 is a schematic diagram showing the state of work of the α machine 1A and the β machine 1B according to the second embodiment. In this embodiment, as in the first embodiment, the excavation work of the α machine 1A is delayed by L0 seconds. In this example, as shown in FIG. 18, the α machine 1A and the β machine 1B move back and forth left and right so as not to interfere with each other's work. In other words, this is an example in which it is possible to perform work efficiently while preventing interference simply by delaying the start time of the work of the α machine 1A. Also, in this example, the magnitude of the delay time L0 was evaluated in step S8A of FIG. 9, and it was found that the delay time L0 was L0. <L WTherefore, no additional work was added (see Figure 17).

[0122] FIG. 19 is a diagram showing a state in which the work plan of the α unit 1A according to the third embodiment has been changed. FIG. 20 is a schematic diagram showing the state of work of the α unit 1A and the β unit 1B according to the third embodiment. In this embodiment, the work plan of the α unit 1A is modified, and the time required for turning is shortened. In other words, this is an example in which the acceleration σ1, deceleration σ2, and maximum speed r1 of the α unit 1A during turning operation are modified to be larger. In this case, as shown in FIG. 20, after the α unit 1A moves away from point B1 toward the A1 side, the β unit 1B approaches point B2. As a result, interference between the two units can be stably prevented.

[0123] In addition to the above, data sets for other construction site environmental conditions may be acquired in addition to the environmental information, operation information, and manipulation information, and a prediction model may be generated using the group of data sets. In the prediction process, by inputting the environmental information, operation information, and manipulation information of the machine to be predicted, it is possible to predict the turning operation up to the turning end angle that can be specified by the environmental information.

[0124] In order to express the present invention, the present invention has been properly and sufficiently described above through the embodiments with reference to the drawings, but it should be recognized that those skilled in the art can easily change and / or improve the above-mentioned embodiments. Therefore, unless the changes or improvements made by those skilled in the art are at a level that causes departure from the scope of the claims described in the claims, such changes or improvements are interpreted as being included in the scope of the claims. [Explanation of symbols]

[0125] HS Construction machinery (e.g. hydraulic excavators) HS1 First Construction Machine HS2 No. 2 Construction Machinery NW communication network PD Construction machinery operation prediction system (an example of a construction machinery operation prediction device) 1 Control processing section 2 Second input section 3 Output section 4 Interface section (IF section) 5 Storage section 11 Control section 12 Model Generation Unit 13 Prediction Department 14 Interference detection section 51 Past performance information storage unit

Claims

1. a model generation process for generating a prediction model that predicts the behavior of the construction machine based on environmental information that represents the surrounding environment of the construction machine, operation information that represents the operating state of the construction machine, and operation information that represents input operations of an input unit that inputs instructions to operate the construction machine, by using environmental information, operation information, and operation information of past results that are associated with each other, and prediction target information that represents the behavior of the construction machine to be predicted; a prediction step of predicting the operation of the construction machine by using environmental information, operation information, and manipulation information at the time of prediction to be predicted in the prediction model generated in the model generation step. Method for predicting construction machinery operation.

2. The construction machine is equipped with a work attachment, which is a mechanism for performing a predetermined task according to the type of the construction machine, The work attachment includes a boom, an arm connected to a tip of the boom, and a bucket swingably attached to the tip of the arm, The operation information includes a swing speed and a swing acceleration of the work attachment, a boom angle of the boom, an arm angle of the arm, and a bucket angle of the bucket. The construction machine operation prediction method according to claim 1 .

3. The environmental information includes a work position of the construction machine, and the positions of at least one of construction machines and obstacles in the vicinity of the construction machine. The construction machine operation prediction method according to claim 1 .

4. The construction machine is equipped with a work attachment, which is a mechanism for performing a predetermined task according to the type of the construction machine, the operation of the construction machine is an operation of the work attachment, The operation information includes an amount of movement of the work attachment. The construction machine operation prediction method according to claim 1 .

5. the model generation step generates the prediction model using a Gaussian process regression method; The prediction step represents the operation of the construction machine as a result of the prediction, including uncertainty of the result of the prediction. The construction machine operation prediction method according to claim 1 .

6. The prediction step represents the predicted operation of the construction machine as time-series data from the time of the prediction. The construction machine operation prediction method according to claim 1 .

7. The method further comprises an interference determination step of determining whether or not an operation of the construction machine will interfere with an operation of another construction machine different from the construction machine that automatically operates according to a preset operation plan, based on the operation of the construction machine predicted in the prediction step and the operation of the operation plan. The construction machine operation prediction method according to claim 1 .

8. The method further includes a motion plan re-creation step of creating a motion plan for the other construction machine so as to prevent interference when it is determined that interference will occur in the interference determination step. The construction machine operation prediction method according to claim 7.

9. a model generation unit that generates a prediction model that predicts the behavior of the construction machine based on environmental information that represents the surrounding environment of the construction machine, operation information that represents the operating state of the construction machine, and operation information that represents the input operation of an operation lever that inputs instructions to operate the construction machine, by using environmental information, operation information, and operation information of past results that are associated with each other, and prediction target information that represents the behavior of the construction machine that is to be predicted; a prediction unit that predicts the operation of the construction machine by using environmental information, operation information, and operation information at the time of prediction to be predicted in the prediction model generated by the model generation unit, Construction machinery operation prediction system.

10. A construction machine operation prediction program for causing a computer to function as the construction machine operation prediction system according to claim 9.

Images