JPH0497097A - Automatic direction control method of small-diameter tunnel robot using fuzzy control - Google Patents

Abstract

PURPOSE:To lighten the burden of an operator by controlling the head angle of the front end of a robot while the robot is pushed in and propelled in a non-earth moving type and using fuzzy control conducting the correction of the direction. CONSTITUTION:The system of a tunnel robot is composed of a tunnel robot body 1 having a head-angle correcting function, a buried pipe 2, a pipe pushing device 3 pushing in the buried pipe, a hydraulic device 4 and a console panel 5. A multiple fuzzy control regulation is used for the determination method of a head angle as the control input of directional control. When a deviation x=xo and the angle of deviation y=yo are input to the multiple fuzzy control regulation, an algebraical product-addition-centroid method and other non-fuzzy changing techniques are used for the head angle (z) of a control input. Accordingly, execution having excellent quality can be conducted.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is a small-diameter tunnel robot that uses fuzzy control to correct the direction by controlling the Herad angle at the tip of the robot while propelling it with no displacement. This invention relates to an automatic direction control method.

[Conventional technology]

Figure 1 shows the system configuration of the tunnel robot. This system consists of a tunnel robot main body 1 with a head angle correction function, a buried pipe 2, a push pipe device 3 for pushing the buried pipe, a hydraulic device 4, and an operation panel 5. The buried pipes 2 are pushed in one by one using hydraulic pressure by the pushing pipe device 3. At this time, the operator 6 successively corrects the head angle and performs direction control so as to follow the planned line.

This direction control currently depends on the experience and knowledge of the operator 6. 7 is the surface of the earth. Figure 2 defines the head angle and pitching angle of this tunnel robot. FIG. 3 shows head angle-pitching angle variation characteristics obtained from actual construction data. The amount of change in pitching angle varies considerably even when the head angle is the same, and the variation varies depending on the head angle. For this reason, even if one wishes to correct a certain angle or direction, it is difficult to know how much the head angle should be corrected, making direction control extremely difficult.

[Problem to be solved by the invention]

This was developed in view of the above circumstances, and the fuzzy control method automates the direction control of small-diameter tunnel robots, which conventionally relied on the operator's experience and intuition, using fuzzy control that uses the operator's experience and knowledge as control rules. The purpose of this study is to provide an automatic direction control method for small-diameter tunnel robots.

[Means and effects for solving the problems] In order to solve the above problems, the present invention provides the following: multiple fuzzy control rules rlF x is A AND y is B THE
The antecedent part X of N z is Cl is the deviation of the small diameter tunnel robot body from the planned line, the antecedent part y is the deviation angle from the planned line (pitching angle - the slope of the planned line), and the antecedent part A is the fuzzy A set, a fuzzy set showing y as the antecedent part B, a fuzzy set showing the head angle of the small-diameter tunnel robot as the consequent part Z, and a fuzzy set showing Z as the consequent part C, and calculate the deviation and bias using a predetermined fuzzy control rule. When the absolute values of the deviations and declination values at the vertices of the triangle of the fuzzy set of angles are A1, B1, C1, ... Z1, 0 in descending order, Al-B1>B1-C1>...>= 71, the values of A1, B1, C1,...Zl are the declination angle A1 is 4.0 to 1.4 [deg], and Bl is 2.2 to 0.75.
(deg), C1 is 1, 1 to 0.3 (deg)
, DI is determined in the range of 0.1 (deg), and the multiple fuzzy control rule is applied to the deviation X=X0, argument y=y
When O is input, the angle Z is determined by human control using the "algebraic product-addition-centroid method" and other defuzzification methods.

The difference from the prior art is that by using the directional control method of the present invention, it is possible to automate the directional control of a small diameter tunnel robot, which was conventionally performed based on the operator's experience and intuition.

〔Example〕

In this example, directional control using the input head angle generated by the fuzzy control method shown in the claims is simulated and evaluated to demonstrate the effectiveness of this control method.

The simulator for direction correction is a dynamic model [
2 (1)] and the formula for calculating the pitching angle and place 1 of the robot [
2 (2), (3)]. The direction control simulation is performed as follows. First, the head angle is determined using fuzzy control rules.

Next, substituting the do angle into the dynamic model of Ni (11) and calculating the amount of direction correction. Then, equations (2) and (3
) to calculate the pitching angle and position of the robot.

The dynamic model of this system is expressed as a stochastic model in which the direction correction angle can be expressed by the sum of probability distribution terms and the chronological order of pitching angle changes that approximately represent the head angle and robot posture. The parameters a,,,b,, are estimated by the least squares method.

This model identification is described in Japanese Patent Application No. 1-277200.

Simulator Δθ, (k)=a, Δθp(k-1)+−+ a,,
Δθp(k-n)+boθh(k)+blθh(k-1
)+-+b,, θh(k-n)+e(k) (1) θJ
k)”θ, (k-1)+Δθ, (k)
(2) Y(k)=Y(k-1)+Lsin(θ
,, (k) ) (3) Define each parameter in FIG. The lower trajectory is the planned line, and the upper trajectory is the robot's trajectory. The position of the design line at stroke k is Y a (k), and the slope of the design line is θa (
k), the position of the robot is Y (k), the pitching angle of the robot is Δθ1, (k), the amount of change in pitching angle is Δθ
, (k), Let L be the length of one stroke. Also,
In the dynamic model of equation (2), e (k) is the residual and n is the order of the model.

A block diagram of the fuzzy control simulator is shown in FIG.

The simulation results using parameter estimates for the Okayama area are shown below. All planned lines were horizontal lines with initial positions and angles of O. FIG. 6 shows the results of fuzzy control simulation (fuzzy control rule [1]). In this simulation, multiple fuzzy control rules rIF x is A AND y is B THE
The antecedent part X of N z is Cl is the deviation of the small diameter tunnel robot body from the planned line, the antecedent part y is the deviation angle from the planned line, the antecedent part A is a fuzzy set indicating X, and the antecedent part B is y. The head angle of the small diameter tunnel robot is taken as the consequent part 2, the fuzzy set shown is 2 as the consequent part C, and the deviation X=X0 and the deviation angle y=Y are set in the multiple fuzzy control rule.

In the fuzzy control method, in which the head angle 2, which is the control input, is determined using a defuzzification method based on the "algebraic product-addition single centroid method" when , the set of multiple fuzzy control rules is determined as shown in Fig. 7. The fuzzy control rule [1] is used.

However, PH:positive hugePB:p
ositive b,ig PM:positive medianPS:pos
itive smallZO:zer.

NH: negative huge NB: negative big NM: negative medianNS: neg
It was set as active small.

The shape of the fuzzy set is a triangle, and it is set as shown in Figure 8, and according to fuzzy rule [1], the value of the deviation at both ends of the triangle of the fuzzy set is PH = 500 Cmm) , PB=l 60 [I
IIIlf, PM-30 (IIm), PS=2
(IIm), Z○=0 [fllI11], N5=PS
, NM=-PMXNB=-PB.

NH--P) (As the value of both ends of the apex and base of the triangle of the fuzzy set, PH=2.1 (deg), PB=0.75 (
deg), PM-〇, 3 (deg), PS=0
．． 1 (deg), ZO=0 (deg), N5=-PS, NM--PM, NB=-PB, NH=
-PH As the head angle values of the apex and base of the fuzzy set triangle, PB= 1.5 (deg), PM= 1.0 (
deg), PS=0.5 (deg), ZO=O(
degE, N5=-PS, NM=-PM, NB=-P
B, and as shown in the claims, PH-PB1>PB-PM>PN-PS>
=lPS1. INH-NB1>INB-NM>
It is selected so that INM-NS l >= INs I.

Further, the residual e (k) was approximated by a normal distribution with an average value of 0 [deg] and a standard deviation of 0.137 [deg]. The initial position and angle are 500 (sa+) and 0, respectively.
(deg). The planned line has an initial deviation of 0 (R1111
) as the horizontal line.

As can be seen from FIG. 6, very good control is achieved. When I investigated the control input head angle, I found that
Until a predetermined distance, the head angle is fixed at the maximum downward position of -1.5 degrees to approach the planned line at maximum speed, and then the head angle is gradually controlled upward to match the angle of the planned line. It can be seen that the control is being performed in an optimal manner. In addition, (initial deviation -400 [, mmJ) Figure 12 shows the fuzzy control simulation (initial deviation
300 [mm)), Figure 13 shows the fuzzy control simulation (initial deviation -
300(sm)), Figure 14 shows the fuzzy control simulation (initial deviation
200 [ma+)), Figure 17 shows the fuzzy control simulation (initial deviation -
100[mn+)), Figure 18 shows the fuzzy control simulation (initial deviation
50 (+am) ), and FIG. 19 is a fuzzy control simulation (initial deviation -50 [1II11]).

It can be seen that even when the initial deviation is changed as shown in FIGS. 8 to 19, the convergence is fast and the design line is followed well.

This fuzzy control rule has an initial deviation of ±500Cats.
If it is within J, almost optimal control can be performed in all cases.

Using the above fuzzy rules, we found the declination angles PH, PB, PM, and PS that provide good control in each district with different N(fi. These are shown in Figure 20.PH of each district in Figure 20 , PB, PM, and PS, the initial deviation is ±
If it is within 500 [113, the declination angle PH, PB in Figure 6
It was confirmed through simulation that even more optimal control can be performed than using , PM, and PS.

From the above, the PH of the declination angle is 4.0 to 1.4 Cdeg)
, PB is 2.2-0.75 (deg), PM is 1.
1 to 0.3 (deg), PS is 0.1 (deg)
) can be selected within the range.

〔Effect of the invention〕

As explained above, according to the present invention, the direction control of a small-diameter tunnel robot, which was conventionally performed based on the operator's experience and intuition, uses the fuzzy control rules of the claims, and the PH of the declination angle is 4.0. ~1.4 (deg), P
B is 2.2~0.75 (deg), PM is 1.1~
0.3 (deg), PS is 0.1 (deg)
By selecting and controlling within this range, the burden on the operator is reduced and high-quality construction is possible.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing the system configuration of the tunnel robot. Fig. 2 is an explanatory diagram showing the definition of head angle and pitching angle. Fig. 3 is a characteristic diagram showing the head angle-pitching angle variation characteristic. Fig. 4 is an explanatory diagram showing the definition of each parameter, Fig. 5 is a block diagram of the fuzzy control simulator, Fig. 6 is a characteristic diagram of fuzzy control simulation (fuzzy control rule [1]), and Fig. 7 is a fuzzy control rule An explanatory diagram showing [1], Fig. 8 is an explanatory diagram showing a fuzzy set, and Fig. 9 is an explanatory diagram showing a fuzzy control simulation (initial deviation -5
00(m+i)), and Figure 10 shows the fuzzy control simulation (initial deviation
400 (+n+)), Figure 11 is a characteristic diagram of fuzzy control simulation (initial deviation -400 (n+m)), Figure 12 is a characteristic diagram of fuzzy control simulation (initial deviation 300 (mm)), Figure 13 shows the fuzzy control simulation (initial deviation -
300 (nm)), and Figure 14 shows the fuzzy control simulation (initial deviation
200(mn+)), Fig. 15 is a characteristic diagram of fuzzy control simulation (initial deviation -200fm:l), Fig. 16 is a characteristic diagram of fuzzy control simulation (initial deviation
100 (nu++)), Fig. 17 is a characteristic diagram of fuzzy control simulation (initial deviation -100 [w+m)), Fig. 18 is a characteristic diagram of fuzzy control simulation (initial deviation 50 (mn+3)), Fig. 19 is fuzzy control simulation (initial deviation −50(+
am) ), FIG. 20 is an explanatory diagram showing the deflection angle to be controlled in each district using fuzzy rules. 1...Robot body, 2...Buried pipe, 3...Push pipe device, 4...Hydraulic device, 5...Operation panel, 6...Operator,...Ground surface.

Claims (1)

[Scope of Claim] Multi-fuzzy control regarding a method for determining the head angle, which is a control input for direction control of a small-diameter tunnel robot, which controls the head angle of the tip of the robot and corrects the direction while pushing and propelling the robot without soil removal. Rule “IF x is A AND y is B THE
The antecedent part x of "N z is C" is the deviation of the small diameter tunnel robot body from the planned line, the antecedent part y is the deviation angle from the planned line, the antecedent part A is a fuzzy set indicating x, and the antecedent part B is y , the head angle of the small diameter tunnel robot as the consequent part z, and the fuzzy set showing z as the consequent part C, and using a predetermined fuzzy control rule, calculate the vertices of the triangle of the fuzzy set of the deviation and declination angle. The absolute values of the deviation and declination values are A1, B1, C1 in descending order.
,... When Z1 is 0, A1-B1>B1-C1>
...>=Z1, A1, B1, C1,...
・The value of Z1 is the argument angle A1 is 4.0 to 1.4 [deg], B
1 is 2.2 to 0.75 [deg], and C1 is 1.1 to 0.
3 [deg], D1 is determined in the range of 0.1 [deg], and the deviation x=x0, deviation angle y=
A small-diameter tunnel robot using fuzzy control characterized in that when y0 is input, the head angle z, which is a control input, is determined using the "algebraic product-addition-centroid method" and other defuzzification methods. automatic direction control method.