Weight change system

Abstract

To enable an operator to perform an operation of a forklift properly even if the operator loses its posture.SOLUTION: A weight change system S is a system used for a forklift and includes a weight change part 21, a posture specification part 303, and a change command part 310. The weight change part 21 changes weights of a pedal 14 and a lever 15 of the forklift or a weight of one of these components. The posture specification part 303 specifies whether an operator loses its posture. The change command part 310 controls the weight change part 21 to change the weights of the pedal 14 and the lever 15 or the weight of one of these components when the operator loses its posture.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a weight changing system for changing the weight of a pedal and / or lever of a forklift. [Background technology]

[0002] There are various types of forklifts, such as reach forklifts and counter-load forklifts. Forklifts are equipped with pedals and levers, and the number and functions of these pedals vary depending on the type of forklift.

[0003] As disclosed in Patent Document 1, a reach forklift is equipped with a brake pedal and a presence pedal. The presence pedal is configured to be able to stop travel and loading, and the driver cannot travel or load unless he or she presses the presence pedal. As disclosed in Patent Document 2, a counter-type forklift is equipped with an accelerator pedal and a brake pedal, and an engine-powered counter-type forklift also has a clutch pedal. Among battery-powered forklifts, automatic forklifts also have an inching pedal instead of a clutch pedal. By pressing the inching pedal, the driver can put the forklift into a half-clutch state and cut off the traveling power.

[0004] As disclosed in Patent Document 3, a reach forklift includes a lift lever for raising and lowering the forks, a tilt lever for changing the vertical angle of the forks, a reach lever for moving the lifting device forward and backward, and an accelerator lever for driving the forklift. As disclosed in Patent Document 4, a counter-load forklift includes a tilt lever and an accelerator lever. As disclosed in Patent Document 5, levers are generally configured to swing forward or backward from a neutral position and automatically return to the neutral position by a spring when not in operation. The invention disclosed in Patent Document 5 uses an actuator to assist lever operation when the amount of operation of the operating lever exceeds a predetermined value to reduce operator fatigue. Other documents that disclose technologies for assisting vehicle lever operation include Patent Document 6, for example.

[0005] For example, drivers operate reach forklifts while standing upright, but they become tired after long periods of driving. This can lead to drivers losing proper posture, resulting in slow pedal and lever operation and sudden acceleration and braking that would not be performed during normal driving. Furthermore, the driver's field of vision may be narrowed by the load on the forks, causing the driver to lean left and right or forward and backward to maintain visibility. This can lead to unexpected accidents if the driver continues to operate the forklift in the normal manner while in this state. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-047759 [Patent Document 2] Japanese Patent Application Publication No. 2017-166663 [Patent Document 3] JP 2016-47745 A [Patent Document 4] Japanese Patent Application Laid-Open No. 2017-88390 [Patent Document 5] Japanese Patent Application Laid-Open No. 2002-370898 [Patent Document 6] Japanese Patent Application Laid-Open No. 2003-301942 Summary of the Invention [Problem to be solved by the invention]

[0007] Therefore, an object of the present invention is to provide a system that can operate a forklift appropriately even when the forklift's posture is disturbed. [Means for solving the problem]

[0008] In order to solve the above problem, the weight change system according to the present invention comprises: 1. A system for use in a forklift, comprising: a weight changing unit that changes the weight of a pedal and / or lever of the forklift; a posture identification unit that identifies whether the driver's posture is disturbed; The vehicle is provided with a change command unit that controls the weight change unit to change the weight of the pedal and / or the lever when the driver's posture is disturbed.

[0009] The weight change system preferably comprises: A foot camera is further provided to capture an image of the driver's right foot, left foot, or both feet and generate a foot image. The posture specifying unit specifies whether both feet, the right foot, the left foot, or both feet of the driver are in predetermined positions based on the foot image, and determines whether the driver's posture is disturbed.

[0010] The weight change system preferably comprises: Further provided is a head camera that photographs the driver's head and generates a head image, The posture specifying unit specifies whether the driver's head is in a predetermined position based on the head image, and specifies whether the driver's posture is disturbed.

[0011] The weight change system preferably comprises: A hand camera is further provided to capture an image of the driver's right hand, left hand, or both hands to generate a hand image; The posture specifying unit specifies whether the driver's hands are in a predetermined position based on the hand image, and specifies whether the driver's posture is irregular.

[0012] The weight change system preferably comprises: a body camera that captures at least a portion of the driver's body to generate a body image; a skeletal information acquisition unit that acquires skeletal information of the driver from the body image, The posture identifying unit identifies whether the driver's posture is disturbed based on the skeletal information acquired by the skeletal information acquiring unit.

[0013] The weight change system preferably comprises: The system further includes a trained model that pre-learns a correlation between an image of a driver and a likelihood score of the driver performing an undesirable operation, and outputs a likelihood score when an image of the driver is input; The posture identification unit identifies whether the driver's posture is irregular based on the output possibility score.

[0014] The weight change system preferably comprises: The posture specifying unit further specifies whether the posture of the driver is disturbed or not based on whether the forklift is moving forward, backward or handling cargo when specifying whether the posture of the driver is disturbed or not.

[0015] The weight change system preferably comprises: The posture identifying unit further identifies that the driver's posture is unstable only when the driver's posture is substantially unstable and continues for a predetermined period of time.

[0016] The weight change system preferably comprises: The vehicle further includes a change notice unit that notifies the driver that the weight of the pedal and / or lever will be changed.

[0017] In order to solve the above problem, a forklift according to the present invention includes any one of the weight changing systems described above.

[0018] In order to solve the above problem, a weight change program according to the present invention includes: A program used in a weight change system including a weight change unit that changes the weight of a pedal and / or lever of a forklift, and a computer, The computer determines whether the driver's posture is disturbed; When the driver's posture is disturbed, the weight changing unit is controlled to change the weight of the pedal and / or the lever. [Effects of the Invention]

[0019] The weight change system according to the present invention allows the forklift to be operated appropriately even when the posture is disturbed. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a side view of a reach forklift equipped with a weight changing system according to an embodiment of the present invention. [Figure 2] FIG. 1 is a top view of a reach forklift, showing the driver's seat and proper driver position. [Figure 3] FIG. 1 is a side view of a counter-load forklift equipped with a weight changing system. [Figure 4] FIG. 1 is a block diagram of a weight change system. [Figure 5] FIG. 1 is a side view of a reach forklift driver leaning forward improperly. [Figure 6] FIG. 1 is a side view showing a counter-load forklift driver leaning inappropriately to the right. [Figure 7]This is a perspective view from the rear of a counter-load forklift, where A shows the operator properly gripping the handle, and B shows the operator improperly gripping the handle. [Figure 8] FIG. 1 is a flow diagram illustrating the operation of the weight modification system. DETAILED DESCRIPTION OF THE INVENTION

[0021] Hereinafter, an embodiment of the weight change system according to the present invention will be described with reference to the accompanying drawings. The weight change system S according to this embodiment is configured to change the weight of both the pedal and the lever, but this is merely an example, and the weight change system according to the present invention may be configured to change the weight of either the pedal or the lever.

[0022] <Forklift configuration> FIG. 1 is a side view of a reach forklift 1a equipped with a weight change system S, and FIG. 2 is a plan view of the forklift 1a. FIG. 3 is a side view of a countertop forklift 1b equipped with a weight change system S. FIG. 4 is a block diagram of the weight change system S. First, the configuration of each of the forklifts 1a and 1b will be described. As shown in FIGS. 1 to 4, the forklifts 1a and 1b each include a plurality of wheels 10, a vehicle body 11, a driver's seat 12, a head guard 13, pedals 14, a lever 15, left and right masts 16, left and right forks 17, a camera 18, a speaker 19, a weight change unit 21, and a control unit 30. The system including the camera 18, the speaker 19, the weight change unit 21, and the control unit 30 constitutes the weight change system S of this embodiment.

[0023] A plurality of wheels 10 are mounted on a vehicle body 11. A driver's seat 12 is mounted on the vehicle body 11, and a head guard 13 has a support pillar 13a and is supported by the support pillar 13a and is mounted above the driver's seat 12. A grip 13b is provided on the support pillar 13a on the rear right side of the forklift 1b, which the driver H holds with his right hand when reversing.

[0024] The pedals 14 of the forklift 1a include a brake pedal and a presence pedal, while the pedals 14 of the forklift 1b include an accelerator pedal, a brake pedal, and an inching pedal. The brake pedal of the forklift 1a is a so-called deadman's brake that is released when depressed. Each pedal is provided below the driver's seat 12. The "pedal" in the present invention may include at least one of the above pedals.

[0025] The lever 15 of the forklift 1b includes a lift lever and a tilt lever, and the forklift 1a also includes an accelerator lever. Each lever is provided in front of the driver's seat 12. The lever 15 of the forklifts 1a and 1b may include a slide lever for sliding the fork 17 left and right. The "lever" in the present invention may include at least one of the above levers.

[0026] The left and right forks 17 are configured to be able to be raised and lowered via the left and right masts 16, and the operator H operates a lift lever to raise and lower the forks 17 to perform cargo handling work.

[0027] Camera 18 is fixed to head guard 13 and captures the head, hands, feet, and entire body of driver H to generate head images, hand images, foot images, and body images. Camera 18 corresponds to the "foot camera," "head camera," "hand camera," and "body camera" of the present invention. The positions of camera 18 are merely examples and are not limiting.

[0028] Camera 18 may be configured with multiple cameras. When camera 18 is configured with multiple cameras, it may be configured with a head camera that takes an image of the head of driver H to generate a head image, a hand camera that takes an image of the right hand, left hand, or both hands of driver H to generate a hand image, a foot camera that takes an image of the right foot, left foot, or both feet of driver H to generate a foot image, and a body camera that takes an image of at least a part of the body of driver H to generate a body image. In this case, too, the position of each camera is not limited to head guard 13, and is provided in a position suitable for generating each image. Furthermore, each camera may also have the function of another camera.

[0029] The speaker 19 emits a warning voice to notify the driver H of a weight change, which will be described later. For example, if the weight of the pedal 14 and the lever 15 is to be increased, the warning voice may be something like "The weight of the pedal and the lever will be increased," or if the weight of the pedal 14 and the lever 15 is to be decreased, the warning voice may be something like "The weight of the pedal and the lever will be decreased."

[0030] The weight change unit 21 changes the weight of the lever 15 and the pedal 14 when operated based on commands from a change command unit 310 (see FIG. 4), which will be described later. The weight change unit 21 has a pedal weight change unit that changes the weight of the pedal 14, and a lever weight change unit that changes the weight of the lever 15. Note that if the weight change unit 21 is configured to change the weight of either the pedal 14 or the lever 15, it will have either a pedal weight change unit or a lever weight change unit.

[0031] The pedal weight change unit changes the weight of the pedal 14 when the driver H presses it down, based on a command from the change command unit 310. The pedal weight change unit is configured by, for example, a device that can adjust the weight by hydraulic pressure or a device that can adjust the weight by a spring.

[0032] Furthermore, for example, the pedal weight change unit may be an assist device that assists the depression force of the pedal 14. In this case, the pedal weight change unit may adjust the assist force based on a command from the change command unit 310, and may substantially change the weight of the pedal 14 when the driver H depresses it.

[0033] Alternatively, the pedal weight change unit may be configured to change the magnitude of the power or braking force of each device responsive to the amount of depression force on the pedal 14, for example, based on a command from the change command unit 310. In this way, the pedal weight change unit effectively changes the weight of the pedal 14.

[0034] The lever weight changing section is configured by, for example, a device that can adjust the weight of the lever 15 by hydraulic pressure, or a device that can adjust the weight of the lever 15 by a spring.

[0035] Furthermore, for example, the lever weight change unit may be an assist device that assists in the operation according to the weight of the lever 15. In this case, the lever weight change unit may adjust the assist force based on a command from the change command unit 310, and may substantially change the weight of the lever 15 when the driver H operates it.

[0036] Alternatively, the lever weight change unit may be configured to change the magnitude of the power of each device responsive to the amount of operation of the lever 15, for example, based on a command from the change command unit 310. In this way, the lever weight change unit effectively changes the weight of the lever 15.

[0037] The control unit 30 is configured by a computer arranged inside the vehicle body 11, and has an arithmetic unit, a storage device, and a memory. The storage device stores a weight change program that causes the computer to function as a skeleton information acquisition unit 302, a posture identification unit 303, a change notification unit 308, and a change command unit 310, which will be described later.

[0038] <Functional configuration> Next, a description will be given of the functional configuration of the weight change system S. As shown in Fig. 4, the control unit 30 has a skeleton information acquisition unit 302, a posture identification unit 303, a change notice unit 308, and a change command unit 310.

[0039] The skeletal information acquisition unit 302 has a first trained model. The first trained model has undergone pre-supervised learning using a deep neural network with a large amount of body images as training data. As a result, when a body image is input, the first trained model outputs skeletal information of the driver H.

[0040] The skeletal information acquisition unit 302 inputs the body image into the first learned model and outputs skeletal information of the driver H. In this way, the skeletal information acquisition unit 302 acquires skeletal information of the driver H.

[0041] The posture identification unit 303 identifies the posture of the driver H based on at least one of the acquired skeletal information, body image, head image, foot image, and hand image. Then, the posture identification unit 303 identifies the driving situation of the forklift 1a (1b), and determines whether the posture of the driver H is disturbed based on the driving situation and the posture of the driver H. The operation of the posture identification unit 303 will be described in detail below.

[0042] The posture identification unit 303 identifies whether the forklift 1a (1b) is stopped or is moving forward, backward, or handling cargo.

[0043] The posture identification unit 303 identifies the position of the head of the driver H based on the head image, and determines whether the head of the driver H is in a predetermined position according to the forward movement, reverse movement or loading / unloading movement of the forklift 1a (1b).

[0044] Alternatively, the posture identification unit 303 has a second trained model. This second trained model is trained in advance using training data in which a head image and the movement of the forklift 1a (1b) are input data, and a score indicating whether the head of the driver H is in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) (hereinafter referred to as the "head position score") is output data. When the head image and the movement of the forklift 1a (1b) are input, the second trained model outputs the head position score. The training method for the second trained model is not particularly limited and may be a deep neural network. The posture identification unit 303 inputs a head image to the second trained model, causes it to output a head position score, and determines whether the head of the driver H is in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) based on the head position score.

[0045] As a result, the posture identifying unit 303 identifies that the head position is not in a predetermined position when the driver H is leaning his / her body inappropriately as shown in, for example, FIGS.

[0046] Furthermore, the posture identification unit 303 identifies the position of the driver H's right hand, left hand, or both hands based on the hand image, and determines whether the driver H's right hand, left hand, or both hands are in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b).

[0047] Alternatively, the posture identification unit 303 has a third trained model. This third trained model is trained in advance using training data in which a hand image and the movement of the forklift 1a (1b) are input data and a score indicating whether the right hand, left hand, or both hands of the driver H are in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) (hereinafter referred to as a "hand position score") is output data. When the hand image and the movement of the forklift 1a (1b) are input, the third trained model outputs the hand position score. The training method for the third trained model is not particularly limited and may be a deep neural network. The posture identification unit 303 inputs a hand image to the third trained model, causes it to output a hand position score, and determines whether the right hand, left hand, or both hands of the driver H are in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) based on the hand position score.

[0048] 7A and 7B are perspective views of the countertop forklift 1b from the rear, with Fig. 7A showing the driver H properly gripping the grip 13b and Fig. 7B showing the driver H improperly gripping the grip 13b. The posture identification unit 303 determines, through the above operation, that the driver H's hands are not in the proper position when the driver H is not gripping the grip 13b while reversing, for example, as shown in Fig. 7B. Additionally, the posture identification unit 303 determines that the driver H's right hand should be above the knee when the lever 15 is not being operated, as shown in Fig. 3, and that the driver H's hands are not in the proper position when the right hand is resting on the lever 15 when the lever 15 is not being operated.

[0049] Furthermore, the posture identification unit 303 identifies the position of the right foot, left foot, or both feet of the driver H based on the foot image, and determines whether the right foot, left foot, or both feet of the driver H are in a predetermined position according to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b).

[0050] Alternatively, the posture identification unit 303 has a fourth trained model. This fourth trained model is trained in advance using training data in which a foot image and the movement of the forklift 1a (1b) are input data and a score indicating whether the right foot, left foot, or both feet of the driver H are in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) (hereinafter referred to as a "foot position score") is output data. When the foot image and the movement of the forklift 1a (1b) are input, the fourth trained model outputs the foot position score. The training method for the fourth trained model is not particularly limited and may be a deep neural network. The posture identification unit 303 inputs a foot image to the fourth trained model, causes it to output a foot position score, and determines whether the right foot, left foot, or both feet of the driver H are in a predetermined position corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) based on the foot position score.

[0051] As a result, the posture identification unit 303 identifies that the feet of the driver H are not in the correct position, for example, when the feet of the driver H are sticking out from the driver's seat 12, the angle of the feet is inappropriate in the reach forklift 1a, or the feet are placed inappropriately on the pedals 14 in the counter-mounted forklift 1b.

[0052] Furthermore, the posture identification unit 303 identifies the posture of the driver H based on the acquired skeletal information, and determines whether this posture is a predetermined posture corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b).

[0053] Alternatively, the posture identification unit 303 has a fifth trained model. This fifth trained model is trained in advance using training data in which skeletal information and the movement of the forklift 1a (1b) are input data, and a score indicating that the skeletal information of the driver H is predetermined skeletal information corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) (hereinafter referred to as the "skeletal score") is output data. When the skeletal information and the movement of the forklift 1a (1b) are input, the fifth trained model outputs the skeletal score. The training method for the fifth trained model may be a deep neural network, and is not particularly limited. The skeletal information used in the fifth trained model includes one or more of the entire body, upper body, entire arms, forearms and hands, lower body, and entire lower legs. The posture identification unit 303 inputs skeletal information to the fifth trained model, causes it to output a skeletal score, and determines whether the skeletal information of the driver H is predetermined skeletal information corresponding to the forward movement, reverse movement, or loading / unloading movement of the forklift 1a (1b) based on the skeletal score.

[0054] As a result, in the reach forklift 1a, the posture identification unit 303 identifies that the posture of the driver H is not the predetermined posture if the driver H's body is facing left, when it is appropriate for the driver H's body to face diagonally left as shown in Figures 1 and 2. The posture identification unit 303 also identifies that the posture of the driver H is not the predetermined posture if the driver H's posture is tilted as shown in Figures 5 and 6, or if the driver H is reversing without gripping the grip 13b as shown in Figure 7B.

[0055] As explained above, the posture identification unit 303 identifies that the driver H's posture is disturbed if the driver H's head position, hand position, or foot position is not in the specified position, or if the driver H's posture is not the specified posture.

[0056] When the posture specification unit 303 specifies that the posture of the driver H is disturbed, the change notice unit 308 notifies the driver H by voice from the speaker 19 that the weight of the pedal 14 and the lever 15 will be changed.

[0057] When the posture specification unit 303 identifies that the posture of the driver H is disturbed and the change notice unit 308 gives notice of a weight change, the change command unit 310 controls the weight change unit 21 to increase the weight of the pedal 14 and the lever 15.

[0058] As a result, when the driver H's posture is disturbed, the change command unit 310 increases the weight of the pedal 14 and the lever 15, thereby suppressing operations such as sudden starts and sudden braking. In particular, when an obstacle in front of the driver H obstructs the driver H's field of vision, causing the driver H's posture to become disturbed as shown in Figures 5 and 6, the weight of the lever 15 and the pedal 14 can be increased to cause the driver H to carefully perform loading and accelerating operations. Furthermore, the weight change system S can give the driver H an opportunity to correct his or her posture by making the change notification unit 308 aware of the disturbance in his or her posture.

[0059] <Weight change system flow> Next, the operation of the weight change system S will be explained again with reference to the flow chart of FIG.

[0060] (1) First, the weight change system S takes an image of the driver H using the camera 18 and generates a head image, a hand image, a foot image, and a body image (see S (step) 1 in FIG. 8).

[0061] (2) Next, the weight change system S acquires skeletal information based on the generated body image by the skeletal information acquisition unit 302 (see S2 in FIG. 8).

[0062] (3) Next, the weight change system S identifies the driving situation, and the posture identification unit 303 identifies the head position, hand position, foot position, and posture of the driver H based on each image, and determines whether these positions and postures are the specified positions and postures according to the driving situation (see S3 in Figure 8).

[0063] (4) Next, when the weight change system S determines that the driver H's posture is disturbed (Yes in S4 of Figure 8), the change notification unit 308 notifies the driver of a change in the weight of the pedal 14 and the lever 15 (see S5 of Figure 8).

[0064] (5) Next, the weight change system S changes the weight of the pedal 14 and the lever 15 to be heavier (see S6 in FIG. 8). As a result, when the driver H's posture is disturbed, the weight change system S can suppress sudden starts and sudden braking and can encourage careful operation. The timing for returning the weight of the pedal 14 and the lever 15 to the original weight may be, for example, when the posture identification unit 303 determines that the driver H's posture is not disturbed, or when a predetermined time has elapsed since the weight of the pedal 14 and the lever 15 was changed. Furthermore, when returning the weight of the pedal 14 and the lever 15 to the original weight, the change notification unit 308 may provide a notification that the weight of the pedal 14 and the lever 15 will be changed to the original weight.

[0065] Although one embodiment of the weight change system according to the present invention has been described above, the present invention is not limited to the above embodiment. The weight change system according to the present invention may be implemented, for example, by combining the above embodiments, or by implementing each of the following modifications individually or in combination.

[0066] <Modification>

[0067] Any of the components of the control unit 30 may be configured, for example, by a server computer provided on the cloud. In this case, the weight change system S changes the weight of the pedal 14 and / or the lever 15 by communicating with this server computer.

[0068] The forklift 1a (1b) may be provided with a monitor instead of the speaker 19, and the change notice unit 308 may use the monitor to give notice that the weight of the lever 15 and / or the pedal 14 will be changed.

[0069] The posture identification unit 303 may identify that the driver H's posture is unstable only when the driver H's posture is substantially continued for a predetermined period of time. This allows the weight change system S to prevent frequent changes to the weight of the pedal 14 and the lever 15 when the driver H looks around or checks the load status for a short period of time, such as three seconds. The phrase "substantially continued" means that, for example, if the weight of the pedal 14 and the lever 15 is changed when the driver H's posture is unstable for three consecutive minutes, even if the driver H's posture temporarily improves (for example, for 10 seconds), that time will not be counted.

[0070] The posture identification unit 303 has, for example, a sixth learning system, which learns in advance the correlation between an image of the driver H (an image including at least one of a head image, a hand image, a foot image, and a body image) and a likelihood score that the driver H will perform an undesirable operation, and outputs a likelihood score when an image of the driver H (an image including at least one of a head image, a hand image, a foot image, and a body image) is input. In this case, the posture identification unit 303 determines whether the posture of the driver H is disturbed based on the output likelihood score. Note that the learning method of the sixth trained model is not particularly limited and may be a deep neural network. [Explanation of symbols]

[0071] S Weight Change System H Driver 1a Reach forklift 1b Countertop forklift 10 wheels 11 Body 12 Driver's seat 13 Head guard 13a Post 13b Grip 14 pedals 15 Lever 16 Mast 17. Fork 18 Camera 19 Speaker 21 Weight change section 30 Control Unit 302 Skeleton Information Acquisition Unit 303 Posture identification part 308 Change Notice Section 310 Change Order Department

Claims

1. 1. A system for use in a forklift, comprising: a weight changing unit that changes the weight of the pedal and / or lever of the forklift; a posture identification unit that identifies whether the driver's posture is disturbed; a change command unit that controls the weight change unit to change the weight of the pedal and / or the lever when the driver's posture is disturbed.

2. a foot camera for capturing an image of the driver's right foot, left foot, or both feet to generate a foot image; The weight change system according to claim 1, wherein the posture identification unit identifies whether the driver's both feet, right foot, left foot, or both feet are in a predetermined position based on the foot image, and identifies whether the driver's posture is disturbed.

3. A head camera is further provided to capture an image of the driver's head and generate a head image. The weight change system according to claim 1 , wherein the posture identification unit identifies whether the driver's head is in a predetermined position based on the head image, and identifies whether the driver's posture is irregular.

4. A hand camera is further provided to capture an image of the driver's right hand, left hand, or both hands to generate a hand image. The weight change system according to claim 1 , wherein the posture specifying unit specifies whether the driver's hands are in a predetermined position based on the hand image, and determines whether the driver's posture is irregular.

5. a body camera that captures an image of at least a portion of the driver's body to generate a body image; a skeletal information acquisition unit that acquires skeletal information of the driver from the body image, The weight change system according to claim 1 , wherein the posture specifying unit specifies whether the driver's posture is irregular based on the skeletal information acquired by the skeletal information acquiring unit.

6. The system further includes a trained model that learns in advance a correlation between an image of the driver and a likelihood score of the driver performing an undesirable operation, and outputs the likelihood score when an image of the driver is input. The weight change system according to claim 1 , wherein the posture identification unit identifies whether the driver's posture is irregular based on the output likelihood score.

7. The weight change system described in any one of claims 1 to 6, wherein the posture identification unit further determines whether the driver's posture is disturbed based on whether the forklift is moving forward, backward, or loading and unloading when determining whether the driver's posture is disturbed.

8. The weight change system according to any one of claims 1 to 6, wherein the posture identification unit further identifies the driver's posture as being disturbed only when the driver's posture is substantially disturbed for a predetermined period of time.

9. The weight change system according to any one of claims 1 to 6, further comprising a change notice unit that notifies the driver that the weight of the pedal and / or the lever will be changed.

10. A forklift equipped with the weight changing system according to any one of claims 1 to 6.

11. A program used in a weight change system including a weight change unit that changes the weight of a pedal and / or lever of a forklift, and a computer, The computer, Identifying whether the driver's posture is disturbed; a weight change program that controls the weight change unit to change the weight of the pedal and / or the lever when the driver's posture is disturbed.

Images