JP6816933B1 - Cargo handling vehicle that can be switched between unmanned driving and manned driving - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a cargo handling vehicle capable of smoothly turning a curve by unmanned operation. SOLUTION: The cargo handling vehicle according to the present invention has a load, a direction of the curve, a degree of the curve, and a width of a passage when the cargo handling vehicle enters the curve when traveling on a curve by manned operation. When the model generation means 11 that generates the estimation model 17 by performing machine learning using the teacher data as to how much the vehicle has shifted from the center of direction and the curve to be traveled next are determined by unmanned driving, the load and By applying the direction of the curve and the degree of the curve to the estimation model 17, it is optimal to estimate how much the cargo handling vehicle should be shifted from the center of the width direction of the passage when entering the curve. The shift amount estimation means 12 and the steering control means 14 that controls the steering device 21 so that the shift amount when the cargo handling vehicle enters the curve matches the estimated amount are provided. [Selection diagram] Fig. 2

Description

The present invention relates to a cargo handling vehicle capable of switching between unmanned operation and manned operation, and more particularly to a cargo handling vehicle capable of smoothly turning a curve by unmanned operation.

Conventionally, a work vehicle capable of switching between manned operation operated by an operator and unmanned operation not operated by an operator has been used in various fields (see, for example, Patent Document 1). One of them is the field of distribution warehouses. In this field, for example, an operation is performed in which normal cargo handling work is left to unmanned operation and only emergency cargo handling work is performed by manned operation by an operator.

FIG. 5 shows a cargo handling vehicle 100 traveling by unmanned operation on a traveling route including a passage P1 and a passage P2 intersecting at right angles. The cargo handling vehicle 100 usually confirms its own position based on a relative distance to a reference object, etc., and has the lowest possibility of deviating from the passages P1 and P2, that is, the width of the passage P1. It is configured to travel along the widthwise centerline C2 of the direction centerline C1 and the passage P2. The broken line T in FIG. 5 is a traveling locus of the cargo handling vehicle 100.

JP-A-2010-222108

However, in order to smoothly turn a curve (for example, the joint of the passages P1 and P2) while maintaining a certain traveling speed, the cargo handling vehicle 100 has a width direction center line (for example, the passage P1) of the passage (for example, the passage P1). It may be preferable to travel along the width direction center line C1), but to travel at a position shifted from the width direction center line to either the left or right direction.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a cargo handling vehicle capable of smoothly turning a curve by unmanned operation.

As a result of diligent studies to solve the above problems, the present inventor may refer to how much the cargo handling vehicle shifts from the center line in the width direction of the passage and enters the curve during manned operation by the operator. I found that.

That is, the cargo handling vehicle that can be switched between unmanned operation and manned operation according to the present invention is (1) when traveling on a curve by manned operation, the load, the direction of the curve, the degree of the curve, and the curve. A model generation means that generates an estimated model by performing machine learning using the teacher data as to how much the cargo handling vehicle has shifted from the center of the width direction of the passage when entering the vehicle, and (2) next by unmanned driving. Once the curve to be driven is determined, the load, the direction of the curve, and the degree of the curve are applied to the estimation model to determine how much the cargo handling vehicle shifts from the center of the width direction of the passage when entering the curve. Optimal shift amount estimating means for estimating whether or not it is better, and (3) Steering control means for controlling the steering device so that the shift amount when the cargo handling vehicle enters the curve matches the estimated amount. It is characterized by having.

According to the present invention, it is possible to provide a cargo handling vehicle capable of smoothly turning a curve by unmanned operation.

It is a perspective view which shows the appearance of the cargo handling vehicle which concerns on embodiment of this invention. It is a block diagram which shows the structure of the cargo handling vehicle which concerns on embodiment of this invention. It is a figure which shows one example (A) and another example (B) of machine learning performed while the cargo handling vehicle which concerns on embodiment of this invention is traveling by manned driving. It is a figure which shows an example of traveling by unmanned operation of the cargo handling vehicle which concerns on embodiment of this invention. It is a figure which shows an example of running by unmanned operation of a conventional cargo handling vehicle.

Hereinafter, examples of the cargo handling vehicle according to the present invention will be described with reference to the accompanying drawings.

FIG. 1 shows a cargo handling vehicle 1 according to an embodiment of the present invention. As shown in the figure, the cargo handling vehicle 1 includes a vehicle body 2 on which the operator is boarding, a mast 3 provided in front of the vehicle body 2, and a pair of left and right forks 4 and 4 that move up and down along the mast 3. I have. Further, the vehicle body 2 includes an accelerator lever 5, a cargo handling command lever group 6, a steering wheel 7, a drive wheel 8, and a control device 10 operated by an operator. The drive wheels 8 also serve as steering wheels.

The cargo handling vehicle 1 can be switched between manned operation that requires the operator to operate the accelerator lever 5, the cargo handling command lever group 6, and the steering 7, and unmanned operation that does not require the operator's operation. These two operations may be switched by operating a switch (not shown), or may be automatically switched based on whether or not the operator is on board.

FIG. 2 shows a block diagram of the cargo handling vehicle 1. As shown in the figure, the cargo handling vehicle 1 includes a cargo handling device 20, a steering device 21, a vehicle position measuring device 22, a model generating means 11 constituting the control device 10, an optimum shift amount estimating means 12, and a traveling route determination. It includes means 13, steering control means 14, layout storage unit 15, and model storage unit 16.

The cargo handling device 20 includes the mast 3 and forks 4, 4 described above, and a load sensor that detects the load of the load 30 acting on the forks 4, 4. The cargo handling device 20 outputs a signal corresponding to the detected load.

The steering device 21 includes the above-mentioned drive wheels 8 (steering wheels) and a driving device that operates the driving wheels 8 under the control of the steering control means 14.

The vehicle position measuring device 22 measures its own position based on a relative distance to one or more reference objects, and outputs a signal corresponding to the measured position. The vehicle position measuring device 22 is, for example, a laser scanner device that projects laser light to the surroundings while rotating and receives reflected laser light from a reflector provided on the wall surface of a warehouse where the cargo handling vehicle 1 performs cargo handling work. May include.

The layout storage unit 15 stores the layout in the warehouse where the cargo handling vehicle 1 performs cargo handling work.

When the model generating means 11 travels on a curve by manned operation, the load, the direction of the curve, the degree of the curve, and how much the cargo handling vehicle 1 is from the center of the width direction of the passage when entering the curve. Machine learning is performed using the shift data as teacher data, and an estimation model 17 is generated in the model storage unit 16. Here, the model generating means 11 can specify the current load based on the signal output by the cargo handling device 20. Further, the model generating means 11 determines the direction and degree of the curve based on the layout stored in the layout storage unit 15 and the signal output by the vehicle position measuring device 22, and the cargo handling vehicle 1 when entering the curve. It is possible to identify how much the shift was.

For example, when the cargo handling vehicle 1 loaded with the load 30 by manned operation by the operator O turns a curve along the trajectory T shown in FIG. 3A, the model generating means 11 sets the load “W1” of the load 30 and the curve. "Left" which is the direction of, "90 °" which is the degree of the curve, and "+ D1" which is the amount of shift from the width direction center line C of the passage P of the cargo handling vehicle 1 when entering the curve. The estimation model 17 is updated by machine learning using the above as teacher data.

Further, when the cargo handling vehicle 1 loaded with the load 30 by manned operation by the operator O turns a curve along the trajectory T shown in FIG. 3 (B), the model generation means 11 sets the load “W2” of the load 30 and the curve. "Right" which is the direction of, "30 °" which is the degree of the curve, and "-D2" which is the amount of shift from the width direction center line C of the passage P of the cargo handling vehicle 1 when entering the curve. The estimation model 17 is updated by machine learning using and as teacher data.

Every time the cargo handling vehicle 1 turns a curve in manned operation, the model generation means 11 updates the estimation model 17 by machine learning.

The traveling route determining means 13 determines a traveling route for performing a predetermined cargo handling operation based on the layout stored in the layout storage unit 15 and the signal output by the vehicle position measuring device 22 during unmanned operation. , The signal regarding the curve to be traveled next is output to the optimum shift amount estimation means 12. The signal output by the traveling route determining means 13 includes a signal relating to the direction of the curve to be traveled next and a signal relating to the degree of the curve.

When the optimum shift amount estimating means 12 receives a new signal regarding the curve to be traveled next from the traveling route determining means 13, the optimum shift amount estimating means 12 applies the load, the direction of the curve, and the degree of the curve to the estimation model 17. , Estimate how much the cargo handling vehicle 1 should be shifted from the center in the width direction of the passage when entering the curve. Further, the optimum shift amount estimating means 12 outputs a signal regarding the estimated optimum shift amount toward the steering control means 14. Here, the optimum shift amount estimation means 12 can specify the current load based on the signal output by the cargo handling device 20 as in the model generation means 11.

The steering control means 14 controls the steering device 21 based on the layout stored in the layout storage unit 15, the signal output by the vehicle position measuring device 22, and the signal output by the optimum shift amount estimating means 12. More specifically, the steering control means 14 steers the steering device 21 so that the shift amount of the cargo handling vehicle 1 when entering the curve to be traveled next matches the optimum shift amount estimated by the optimum shift amount estimating means 12. To control.

For example, when the cargo handling vehicle 1 is determined to turn left at a 90 ° curve in front when the cargo handling vehicle 1 is in the position shown in FIG. 4A, the optimum shift amount estimating means 12 loads the load 30. "W3", "left" which is the direction of the curve, and "90 °" which is the degree of the curve are applied to the estimation model 17, and the optimum shift amount is estimated to be "+ D3". Then, in response to this, the steering control means 14 controls the steering device 21, and as a result, the shift amount of the cargo handling vehicle 1 when entering the curve becomes + D3 (see FIG. 4B).

As described above, when the cargo handling vehicle 1 according to the present embodiment enters the curve by unmanned operation with reference to the traveling route unconsciously determined by the skilled operator O according to the load of the load 30 and the degree of the curve. The optimum shift amount is estimated, and the vehicle travels according to the estimated result. Therefore, according to the cargo handling vehicle 1 according to the present embodiment, it is possible to turn a curve by unmanned driving with the same smoothness as manned driving by a skilled operator O.

Although the embodiment of the cargo handling vehicle according to the present invention has been described above, the configuration of the present invention is not limited to this, and it goes without saying that various modifications can be considered.

1 Cargo handling vehicle 2 Vehicle body 3 Mast 4 Fork 5 Accelerator lever 6 Cargo handling command lever group 7 Steering 8 Drive wheels 10 Control device 11 Model generation means 12 Optimal shift amount estimation unit 13 Travel route determination means 14 Steering control means 15 Layout storage unit 16 Model storage 17 Estimated model 20 Cargo handling device 21 Steering device 22 Vehicle position measuring device 30 Load O operator

Claims (1)

It is a cargo handling vehicle that can be switched between unmanned driving and manned driving.
When traveling on a curve by the manned operation, the load, the direction of the curve, the degree of the curve, and how much the cargo handling vehicle was shifted from the center of the width direction of the passage when entering the curve. A model generation means that generates an estimated model by performing machine learning using the heel as teacher data,
When the curve to be driven next is determined by the unmanned driving, the load, the direction of the curve, and the degree of the curve are applied to the estimation model, so that the cargo handling vehicle enters the curve when the cargo handling vehicle enters the passage. Optimal shift amount estimation means for estimating how much the shift should be from the center in the width direction,
Steering control means that controls the steering device so that the shift amount when the cargo handling vehicle enters the curve matches the estimated amount.
A cargo handling vehicle characterized by being equipped with.

Images