CN103381603A - Autonomous obstacle crossing programming method of deicing and line inspecting robot for high-voltage transmission line - Google Patents

Abstract

The invention discloses an autonomous obstacle crossing programming method of a deicing and line inspecting robot for a high-voltage transmission line. The method comprises the following steps: step 1, detecting environment information by utilizing a laser radar mounted at the tail end of a mechanical arm, so as to obtain a robot movement ahead obstacle signal; step 2, according to a difference value between the current position and the expected position of the mechanical arm and the obstacle signal in the current condition, programming out a fuzzy programmed angle of the movement ahead mechanical arm by utilizing a fuzzy planner; step 3, performing online optimization to the fuzzy programmed angle by utilizing the particle swarm optimization, so as to obtain a particle swarm fuzzy programmed angle of the movement ahead mechanical arm; step 4, obtaining control moments of all joints by utilizing a neural network self-adaptive controller, and guiding the mechanical arm to act. By adopting the fuzzy programming method, an obstacle crossing programming decision can be made in real time according to the current condition of the deicing and line inspecting robot, and the inaccuracy and hysteretic nature of information perception can be overcome; meanwhile, by adopting the particle swarm optimization, the fuzzy programmed angle can be optimized online, so that the track can be smoother and the redundancy can be smaller.

Description

The active obstacle planing method of deicing high-voltage power transmission line inspection robot

Technical field

The present invention is mainly concerned with the independent navigation field of industrial special robot, particularly the active obstacle planing method of deicing high-voltage power transmission line inspection robot.

Background technology

Ultra-high-tension power transmission line is being undertaken the important task of electric power transfer, and its safe and reliable operation is directly connected to the stable sound development of a national economy.Be the operation that guarantees that it is safe and reliable, need to carry out regularly detecting and safeguarding.Maked an inspection tour by tower with ascender line by two people per month by rules requirement ultra-high-tension power transmission line, annual human and material resources expense is higher.And in the winter time under abominable hazardous environment, ultra-high-tension power transmission line accumulated snow and icing that freezing sleet can cause usually cause line tripping, insulator arc-over, broken string, the accidents such as bar and communication disruption of falling.China is one of ultra-high-tension power transmission line Harm Accident probability of happening higher country, alleviate ultra-high-tension power transmission line accumulated snow icing problem in winter for ensureing transmission of electricity safety, usually take two kinds of measures: great current deicing technology and the deicing of manually reaching the standard grade, the former causes serious energy loss, and not only labour intensity is large for the latter, operating efficiency low but also have larger danger.Therefore, the deicing that replaces manually reaching the standard grade of a kind of safe, economical, efficient deicer of exploitation extremely has realistic meaning.

The deicing inspection robot has line walking and two kinds of functions of deicing concurrently, compare with the Traditional Man deicing of reaching the standard grade, the work of deicing inspection robot need not to stop supply of electric power, security performance high, can work continuously, not affect operation of power networks, have boundless application prospect.When the deicing inspection robot can't harm the deicing operation in adverse circumstances, the aspects such as its structure, function, power, communication there is special requirement.In addition, realize that the deicing inspection robot carries out on a large scale, for a long time online deicing operation in the bad weather environment, just must make it possess the function of autonomous online walking and obstacle detouring.The autonomy-oriented of deicing inspection robot and intelligent concentrated reflection are accurately cognitive disorders thing and steadily obstacle detouring in particular surroundings.

As plan objects, deicing inspection robot mechanical arm be one become when having, the nonlinear system of the Multiinputoutput of close coupling, its Consideration is more, programmed decision-making is very complicated.In prior art, the deicing inspection robot is usually operated on the abominable ultra-high-tension power transmission line of weather conditions, outdoor destructuring environment has diversity, complexity, randomness and uncertainty, scenery has different scenes under different environment, as indefinite in the staggered confusion in shaft tower place, fluctuating in ultra-high-tension power transmission line, electrical equipment random distribution, electrical equipment and line ice coating exist each other and block.The dynamic change of illumination in addition,, background and weather all has a huge impact the normal operation of deicing inspection robot.Traditional visual servo is controlled the working condition that planning strategy is difficult to deal with complexity like this, and sensor-based method for planning track adopts repeatedly fitting of a polynomial usually, although the method can generate level and smooth planned trajectory, but barrier on the path is not taken into account, do not had the barrier of keeping away obstacle crossing function.Therefore, obtain ambient condition information by Laser Radar Scanning, design a kind of simple and reliable, real-time good, the deicing high-voltage power transmission line inspection robot obstacle detouring planing method of being convenient to realize, can deal with the multiclass complex environment is to guarantee the normal effectively key technology difficult problem of work of deicing inspection robot.

Summary of the invention

the invention provides the active obstacle planing method of deicing high-voltage power transmission line inspection robot, its purpose is to overcome the deficiencies in the prior art, adopt the online obstacle detouring planing method of particle cluster algorithm and fuzzy logic, realize in accessible situation, desired trajectory effectively being followed the tracks of, guarantee the accessibility in each orientation of deicing inspection robot mechanical arm, when the deicing inspection robot has the danger of the barrier of chance, can implement effectively to keep away barrier according to human expert's experience, obstacle detouring, thereby reach the active obstacle planing method that has than the deicing high-voltage power transmission line inspection robot of high-intelligentization.

The active obstacle planing method of deicing high-voltage power transmission line inspection robot comprises the steps:

Step 1: utilize the laser radar detection ambient condition information that is arranged on mechanical arm tail end, obtain the operation the place ahead obstacle signal B of robot;

Step 2: deicing robot observes by the camera that carries the position of moving the place ahead cable, and 30cm place, robot forearm the place ahead cable is set to the desired locations (x of mechanical arm tail end current time _g, y _g); Difference (e according to mechanical arm tail end current location and desired locations _x, e _y) and current obstacle signal B, utilize the fuzzy programming device to cook up the fuzzy programming angle in each joint of the front surface mechanical arm of traffic direction;

Step 3: can reach principle according to shortest path and target, utilize particle cluster algorithm to carry out on-line optimization to the fuzzy programming angle, obtain the population fuzzy programming expected angle in each joint of the front surface mechanical arm of traffic direction;

Step 4: with the input of each joint population fuzzy programming angle as neural network adaptive controller, neural network adaptive controller is exported each joint control moment τ, instructs the mechanical arm action.

Obtain the operation the place ahead obstacle signal B of robot in described step 1 and refer to according to the obstacle detouring judgement critical distance A that sets, judge whether the distance of the Obstacle Position that mechanical arm tail end current location and Airborne Lidar measure surpasses the obstacle detouring judgement critical distance A that sets:

e _xb=|x-x _b|(1)

e _yb=|y-y _b|(2)

Wherein, x _b, y _bBe the two-dimensional coordinate position of barrier, x, y is the mechanical arm tail end current location, works as e _xb≤ A and e _ybDuring≤A, judgement deicing inspection robot moves the place ahead has the barrier of chance dangerous, output obstacle signal B=1, otherwise B=0.

In described step 2 according to the difference (e of mechanical arm tail end current location and desired locations _x, e _y) and present case under obstacle signal B, each joint fuzzy programming angle of mechanical arm when utilizing the fuzzy programming device to cook up deicing robot to travel forward, detailed process is as follows:

1) determine fuzzy programming device input message, and the input data carried out Fuzzy processing:

With mechanical arm tail end control information (e _x, e _y) and obstacle signal B input as the fuzzy programming device:

$\left\{\begin{array}{c}{e}_x=x-x_g,e_x\∈[-\mathrm{0.4,0.36}]\\ e_y=y-y_g,e_y\∈[-\mathrm{0.4,0.56}]\end{array}\right.B\∈\left[\mathrm{0,1}\right]$

(x _g, y _g) be the desired locations that 30cm place, robot forearm the place ahead cable is set to the mechanical arm tail end current time;

Fuzzy programming device input message is carried out Fuzzy processing, mechanical arm tail end error of coordinate (e _x, e _y) corresponding 3 fuzzy language variablees { P}={ " bears " for N, Z, " zero ", " just " }, fuzzy language variable N, Z and P respectively with membership function μ _N, μ _ZAnd μ _PCorresponding one by one, the codomain scope of membership function is [0,1]; The input variable of membership function is mechanical arm tail end error of coordinate (e _x, e _y):

${\mathrm{\&mu;}}_N\left(e_x\right)=\left\{\begin{array}{cc}1& e_x\&le;-0.03\\ \frac{e_x}{-0.03}& -0.03<e_x\&le;0\\ 0& e_x>0\end{array}\right.---\left(3\right)$ ${\mathrm{\&mu;}}_N\left(e_y\right)=\left\{\begin{array}{cc}1& e_y\&le;-0.03\\ \frac{e_y}{-0.03}& -0.03<e_y\&le;0\\ 0& e_y>0\end{array}\right.---\left(6\right)$

${\mathrm{\&mu;}}_P\left(e_x\right)=\left\{\begin{array}{cc}0& e_x\&le;0\\ \frac{e_x}{0.03}& 0<e_x\&le;0.03\\ 1& e_x>0.03\end{array}\right.---\left(5\right)$ ${\mathrm{\&mu;}}_P\left(e_y\right)=\left\{\begin{array}{cc}0& e_y\&le;0\\ \frac{e_y}{0.03}& 0<e_y\&le;0.03\\ 1& e_y>0.03\end{array}\right.---\left(8\right)$

Choose the membership function value and be non-zero fuzzy language variable as the fuzzy language variable of input message;

2) obtain output fuzzy language variable according to fuzzy rule;

According to fuzzy language variable and the obstacle signal B of input message, inquiry fuzzy rule design table, as shown in table 1, obtain exporting accordingly fuzzy language variable { None, N, P}, wherein, None represents that mechanical arm is failure to actuate, and N represents the mechanical arm operating angle for negative, and P represents that the mechanical arm operating angle is for just;

The design of table 1 fuzzy rule

3) output fuzzy language variable is carried out fuzzy reasoning, obtain membership function value corresponding to output fuzzy language variable; Adopt the Mamdani reasoning, the membership function that output fuzzy language variable is corresponding is as follows:

${\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f1}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}\&le;10)\end{array}\right.---\left(9\right)$ ${\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f2}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f2}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}\&le;10)\end{array}\right.---\left(11\right)$

${\mathrm{\&mu;}}_P\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}<0)\\ \frac{{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}}{10}& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}\&le;10)\end{array}\right.---\left(10\right)$ ${\mathrm{\&mu;}}_P\left({\mathrm{\&theta;}}_{f2}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}<0)\\ \frac{{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}}{10}& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}\&le;10)\end{array}\right.---\left(12\right)$

With e _xAnd e _yCorresponding membership function value under each self-corresponding fuzzy language variable is respectively to e _xAnd e _yThe union of corresponding all membership function values is tried to achieve first and is occured simultaneously and the second common factor; The first common factor, the second common factor and B are asked the 3rd common factor, obtain an exact value; Occur simultaneously respectively to two μ that export with the 3rd _N(θ _f1) and μ _P(θ _f1) cut, obtain the membership function value of output fuzzy language variable;

4) adopt the sharpening interface to obtain output angle;

Because initial conditions may take a plurality of input fuzzy language variablees simultaneously, according to the derivation of fuzzy rule, will form a plurality of Output rusults, therefore need to carry out precision to the result of fuzzy reasoning;

At first get that in each Output rusults, the maximum membership degree functional value in two fuzzy programming angles compares, select the Output rusults at the place of membership function value maximum to export as final result, if in each Output rusults process relatively, the maximum membership degree function that a plurality of results occur is maximum and equal side by side, selects delivery to stick with paste less the exporting as final result of language amount in Output rusults; Then will export membership function value substitution formula 9-12 in the fuzzy language result function of inverting, obtain the two actual output angles in joint; The fuzzy language amount is pressed N, None, P ascending order.

It is as follows that particle cluster algorithm in described step 3 carries out the on-line optimization process to the fuzzy programming angle:

1): optimization object is selected: self-correcting parameter k=(k is set ₁, k ₂) k ₁, k ₂∈ [0,2] makes the fuzzy programming expectation angle after optimization be (k ₁θ _f1, k ₂θ _f2);

2) update mode is determined: adopt the elementary particle group to optimize update mode, be expressed as:

v _ij(n+1)=wv _ij(n)+r ₁(n)c ₁[pbest _ij(n)-x _ij(n)]+r ₂(n)c ₂[gbest _ij(n)-x _ij(n)](13)

x _ij(n+1)=x _ij(n)+v _ij(n+1)(14)

Wherein: i is for initializing correction parameter group number, and j is the dimension of every group of correction parameter, v _ij(n) diverse vector of the j dimension that is correction parameter group i in the n time iteration, x _ij(n) be the numerical value of correction parameter group i j dimension in the n time, r ₁, r ₂Be independent random function, c ₁, c ₂Be acceleration weight, c ₁, c ₂All value is that 1.2, w is inertia weight;

Inertia weight w upgrades by the mode of linear decrease, is shown below:

$w=w_{\max }-n\frac{w_{\max }-w_{\min }}{n_{\max }}---\left(15\right)$

Wherein: n is the current iteration number of times, n _maxBe total iterations, w _maxBe 0.9, w _minBe 0.4;

3) fitness function design:

Min:f(t)=s ₁f ₁(t)+s ₂f ₂(t)（16）

Wherein, s ₁Be 0.4, s ₂Be 0.6;

At any t constantly, the distance of mechanical arm tail end and initial point position is the integration minimum in time, fitness function f ₁(t) be:

$\mathrm{Min}:f_1\left(t\right)\underset{0}{\overset{t}{\&Integral;}}\sqrt{{[x\left(t\right)-x_0]}^2+{[y\left(t\right)-y_0]}^2}\mathrm{dt}---17$

Impact point attracts to be presented as that between mechanical arm tail end and impact point, air line distance is minimum, fitness function f ₂(t) be;

$\mathrm{Min}:f_2\left(t\right)=\sqrt{{[x\left(t\right)-x_g]}^2+{[y\left(t\right)-y_g]}^2}---\left(18\right)$

Wherein: x _o, y _oBe mechanical arm tail end initial point position, x (t), y (t) are t mechanical arm tail end position constantly, x _gAnd y _gExpression impact point position;

4) population on-line optimization:

Step1: population initializes, and produces initial value and the diverse vector of correction parameter group, with (the k of current time t ₁θ _f1, k ₂θ _f2) 16 calculating of substitution fitness function formula;

Step2: at iterations n _maxIn, according to formula 13, formula 14, formula 15 is upgraded correction parameter group numerical value.Each iteration is all selected local optimum parameter group pbest and the parameter group gbest of global optimum;

Step3: after iteration is completed, with the correction parameter (k of the parameter group gbest of global optimum output as moment t ₁, k ₂).The excitation mechanical arm is by (k ₁θ _f1(t), k ₂θ _f2(t)) action;

Step4: the mechanical arm tail end sensor obtains next t+1 terminal position constantly, turns Step1.

Neural network adaptive controller in described step 4 comprises proportion control-K _pS and neural network control

Described neural network adaptive controller is shown below:

$\mathrm{\&tau;}=-K_{p}S+{\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(19\right)$

$\stackrel{\stackrel{\&CenterDot;}{^}}{W}=-\mathrm{\&Gamma;g}\left(\mathrm{\&chi;}\right)S^T---\left(20\right)$

Wherein: τ is control moment, K _pWith Γ be symmetric positive definite matrix, K _pFor the proportion control parameter is set to $\left[\begin{array}{cc}5& 0\\ 0& 5\end{array}\right],$ Γ is set to 8 * 8 unit matrix, g (χ)=[g ₁(χ), g ₂(χ) ..., g _N(χ)] ^T, g _j(χ) (1≤j≤r) is the RBF vector,

Be the connection weight value matrix of hidden layer to output layer,

Initial value is made as 8 dimension matrixes in [1,1] interior random distribution;

S is the synovial membrane face

K θ wherein _fBe population fuzzy programming angle, k θ _f=[k ₁θ _f1, k ₂θ _f2], θ is the current angle of mechanical arm, θ=[θ ₁, θ ₂], Λ is symmetric positive definite matrix $\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right];$

Adaptive neural network is shown below:

$\hat{h}={\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(21\right)$

Wherein:

Be the neutral net input,

Be the first joint planning angular speed,

Be the first joint planning angular acceleration,

Be second joint planning angular speed,

Be second joint planning angular acceleration, θ ₁Be the first joint action angle, Be the first joint action angular speed, θ ₂Be the second joint operating angle, Be second joint operating angle speed,

Be the connection weight value matrix of hidden layer to output layer,

Initial value is made as 8 * 8 unit matrix, and the attitude different according to mechanical arm and each joint angles desired value real-time update are to adapt to different demands for control; G (χ)=[g ₁(χ), g ₂(χ) ..., g _N(χ)] ^TBe RBF vector, g _j(χ) (1≤j≤r) has following form:

$g_j\left(\mathrm{\&chi;}\right)=\exp (-\frac{{\left|\right|\mathrm{\&chi;}-{\mathrm{\&mu;}}_j\left|\right|}^2}{{\mathrm{\&sigma;}}_j^2})---\left(22\right)$

Wherein: r is the number of hidden layer neuron; μ _jCentered by vector, σ _jBe the width vector;

The output of adaptive neural network

The target of approaching be compensation term, compensation term is shown below:

$M\left(\mathrm{\&theta;}\right)[{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)]+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})[{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)]+G\left(\mathrm{\&theta;}\right);---\left(23\right)$

In adaptive neural network, hidden layer number r is 20, center vector μ _jBe set to the random function of even distribution [10,10], width vector σ _jAll be set to 2.

The output of neutral net is the compensation term matrix in each joint, and neutral net is input as

Namely that current angle information (angle and angular speed) according to population fuzzy programming angle information (angular speed and angular acceleration) and each joint goes the real-time compensation term of approaching

$M\left(\mathrm{\&theta;}\right)[{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)]+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})[{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)]+G\left(\mathrm{\&theta;}\right).$

In control law

Synovial membrane face S is comprised of e, and S is tending towards 0 and just can causes e to be tending towards 0, just can guarantee tracking performance in theory, and this provides in stability proof part.In this control law, first half is proportion control, and latter half is the compensation term that neutral net is approached, if the perfection of neutral net convergence approaches, this complication system just becomes simple proportion control.In formula

Be weight matrix, be real-time update, provide the formula of renewal

S is the synovial membrane face, is approximate of error,

Can be according to current synovial membrane face S and input and output

Real-time update.Whole process is a dynamic neural network control process.

Beneficial effect

The active obstacle planing method of deicing high-voltage power transmission line inspection robot provided by the invention, adopt laser radar detection deicing inspection robot ambient condition information, Changes in weather when having overcome the work of deicing inspection robot, illumination condition is changeable, the deficiencies such as cabinet shake can effectively detect the barrier on ultra-high-tension power transmission line; Compare with in the past traditional visual servo obstacle detouring strategy, the Fuzzy Programming that the present invention adopts can be according to deicing inspection robot the present situation, effectively follow the tracks of in real time, the obstacle detouring programmed decision-making, overcome inaccuracy and the hysteresis quality of perception end obtaining information, make the deicing inspection robot have better independence and intelligent; Adopt particle swarm optimization algorithm to carry out real-time optimization to fuzzy programming device planning angle, in the situation that guarantee that impact point can reach, realized that mechanical arm tail end goes through shortest path, effectively reduced the work energy consumption of deicing inspection robot, improved the flying power of deicing inspection robot; Adopt neural network adaptive controller that deicing inspection robot mechanical arm is implemented to control, this controller can the real-time online self study, has good adaptivity, particularly this quasi-nonlinear control object of multi-joint mechanical arm has been embodied good control performance, can realize the effective tracking to various types of signal, satisfy deicing inspection robot active obstacle demand.

Description of drawings

Fig. 1 is the fuzzy obstacle detouring Structure Planning of the population schematic diagram that the present invention adopts;

Fig. 2 is the deicing inspection robot frame for movement schematic diagram that the present invention adopts;

Fig. 3 is deicing inspection robot obstacle detouring schematic diagram;

Fig. 4 is fuzzy logic planner internal structure schematic diagram of the present invention;

Fig. 5 is deicing inspection robot mechanical arm tail end confusion region schematic diagram;

Fig. 6 is particle cluster algorithm on-line optimization schematic flow sheet of the present invention;

Fig. 7 is neural network adaptive controller cut-away view of the present invention;

Fig. 8 is deicing inspection robot mechanical arm eight directional reachability test figure, wherein, figure (a) is the top impact point, figure (b) is the front upper place impact point, and figure (c) is the place ahead impact point, and figure (d) is the front lower place impact point, figure (e) is the below impact point, figure (f) is back lower place impact point, and figure (g) is the rear area target point, figure (h is) back upper place impact point;

Fig. 9 is deicing inspection robot mechanical arm continuous signal tracing figure;

Figure 10 is deicing inspection robot mechanical arm obstacle performance resolution chart.

The specific embodiment

The present invention will be further described with concrete case study on implementation below with reference to accompanying drawing.

As shown in Figure 1, the present invention is the active obstacle planing method of deicing high-voltage power transmission line inspection robot, comprises the steps:

Step 1: utilize the laser radar detection ambient condition information that is arranged on mechanical arm tail end, obtain the operation the place ahead obstacle signal B of robot;

Step 2: deicing robot observes by being arranged on the camera on cabinet the position of moving the place ahead cable, 30cm place, robot forearm the place ahead cable is set to the desired locations (x of mechanical arm tail end current time _g, y _g); Difference (e according to mechanical arm tail end current location and desired locations _x, e _y) and present case under obstacle signal B, utilize the fuzzy programming device to cook up the fuzzy programming angle in each joint of surface mechanical arm before traffic direction;

Step 3: can reach principle according to shortest path and target, utilize particle cluster algorithm to carry out on-line optimization to the fuzzy programming angle, obtain the population fuzzy programming expected angle in each joint of the front surface mechanical arm of traffic direction;

Step 4: with the input of each joint population fuzzy programming angle as neural network adaptive controller, neural network adaptive controller output obtains each joint control moment τ, thereby instruct each joint action of mechanical arm, realize normal walking and obstacle detouring action on deicing inspection robot line.

In this example, each arm has two joints, and neural network adaptive controller output obtains the control moment (τ in two joints on each arm ₁, τ ₂).

Can run into barriers such as stockbridge damper, strain clamp, insulator when as shown in Figures 2 and 3, the deicing inspection robot is reached the standard grade work.According to the mechanical mechanism design of deicing inspection robot, the obstacle judgement link in above-mentioned steps (2) arranges as follows:

The deicing inspection robot is without the environmental knowledge of priori, adopt laser radar that the mechanical arm tail end surrounding environment is surveyed, to the memoryless function of the environmental information that detects, arbitrary moment, mechanical arm can only detect centered by its terminal position, radius is the information in the zone of d, if find that obstacle returns to corresponding complaint message.The obstacle detouring demand concrete according to the deicing inspection robot arranges as follows:

The scope that the detection of mechanical arm is set is 0.03 meter, and during less than 0.06 meter, judgement deicing inspection robot has the barrier of chance dangerous when the distance of the coordinate of mechanical arm tail end and barrier.All the other situations are all safety.

e _xb=|x-x _b|(1)

e _yb=|y-y _b|(2)

Wherein, x _b, y _bBe the two-dimensional coordinate position of barrier, x, y is the mechanical arm tail end current location, works as e _xb≤ 0.06 and e _yb≤ 0.06 o'clock, judgement deicing inspection robot moves the place ahead had the barrier of chance dangerous, output obstacle signal B=1, otherwise B=0.

As shown in Figure 3 and Figure 4, in above-mentioned steps (2), fuzzy programming device partial design in Fuzzy particle swarm artificial obstacle detouring planning strategy:

1. determine fuzzy programming device input message, and the input data carried out Fuzzy processing:

Select mechanical arm tail end control information (e _x, e _y) and obstacle judgement information B as the input of fuzzy programming device, next operating angle (θ constantly of mechanical arm each joint _f1, θ _f2) export as the fuzzy programming device.Because deicing inspection robot mechanical arm is subject to environmental constraints,

Its end coverage has certain limitation:

$\left\{\begin{array}{c}x\&Element;[-\mathrm{0.2,0.56}]\\ y\&Element;\left[\mathrm{0,0.96}\right]\end{array}\right.---\left(3\right)$

The initial point of considering deicing inspection robot mechanical arm tail end is (0.2,0.4) in the position of two-dimensional coordinate system, and therefore, the margin of error domain of mechanical arm tail end is:

$\left\{\begin{array}{c}{e}_x=x-x_g,e_x\&Element;[-\mathrm{0.4,0.36}]\\ e_y=y-y_g,e_y\&Element;[-\mathrm{0.4,0.56}]\end{array}\right.---\left(4\right)$

Design obstacle judgement information B is switching variable, and domain is B ∈ [0,1].

For the consideration of security, the output angle of fuzzy programming device is unsuitable excessive.Therefore, design fuzzy programming device output language variable field is:

u ₁,u ₂∈[-10,10](5)

Fuzzy programming device input message is carried out Fuzzy processing, mechanical arm tail end error of coordinate (e _x, e _y) corresponding 3 fuzzy language variablees { P}={ " bears " for N, Z, " zero ", " just " }, fuzzy language variable N, Z and P respectively with membership function μ _N, μ _ZAnd μ _PCorresponding one by one, the codomain scope of membership function is [0,1]; The input variable of membership function is mechanical arm tail end error of coordinate (e _x, e _y):

${\mathrm{\&mu;}}_N\left(e_x\right)=\left\{\begin{array}{cc}1& e_x\&le;-0.03\\ \frac{e_x}{-0.03}& -0.03<e_x\&le;0\\ 0& e_x>0\end{array}\right.---\left(6\right)$ ${\mathrm{\&mu;}}_N\left(e_y\right)=\left\{\begin{array}{cc}1& e_y\&le;-0.03\\ \frac{e_y}{-0.03}& -0.03<e_y\&le;0\\ 0& e_y>0\end{array}\right.---\left(9\right)$

${\mathrm{\&mu;}}_p\left(e_x\right)=\left\{\begin{array}{cc}0& e_x\&le;0\\ \frac{e_x}{0.03}& 0<e_x\&le;0.03\\ 1& e_x>0.03\end{array}\right.---\left(8\right)$ ${\mathrm{\&mu;}}_p\left(e_y\right)=\left\{\begin{array}{cc}0& e_y\&le;0\\ \frac{{\mathrm{<figure-callout\; id="0"\; label="e\; y"\; filenames="2013102690280100003DEST\_PATH\_IMAGE003.png,HDA00003437580300032.png"\; state="\{\{state\}\}">e</figure-callout>}}_y}{0.03}& 0<e_y\&le;0.03\\ 1& e_y>0.03\end{array}\right.---\left(11\right)$

Choose the membership function value and be non-zero fuzzy language variable as the fuzzy language variable of input message;

Next moment operating angle (θ of each joint of mechanical arm _f1, θ _f2) corresponding two fuzzy language collection { N, P}={ " bears ", " just " }, membership function adopts ramp function, is expressed as:

${\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f1}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}\&le;10)\end{array}\right.---\left(12\right)$

${\mathrm{\&mu;}}_p\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}<0)\\ \frac{{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}}{10}& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}\&le;10)\end{array}\right.---\left(13\right)$

Wherein, θ _f2The same θ of membership function design _f1

2) obtain output fuzzy language variable according to fuzzy rule;

Fuzzy language variable and obstacle signal B according to input message, inquiry fuzzy rule design table, as shown in table 1, obtain exporting accordingly fuzzy language variable { None, N, P}, wherein, None represents that mechanical arm is failure to actuate, and N represents the mechanical arm operating angle for negative, and P represents that the mechanical arm operating angle is for just;

The design of table 1 fuzzy rule

3) output fuzzy language variable is carried out fuzzy reasoning, obtain membership function value corresponding to output fuzzy language variable; Adopt the Mamdani reasoning, the membership function that output fuzzy language variable is corresponding is as follows:

${\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f1}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}\&le;10)\end{array}\right.---\left(14\right)$ ${\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f2}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f2}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}\&le;10)\end{array}\right.---\left(16\right)$

${\mathrm{\&mu;}}_P\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}<0)\\ \frac{{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}}{10}& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA00003437580300023.png"\; state="\{\{state\}\}">f</figure-callout>}1}\&le;10)\end{array}\right.---\left(15\right)$ ${\mathrm{\&mu;}}_P\left({\mathrm{\&theta;}}_{f2}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}<0)\\ \frac{{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}}{10}& (0\&le;{\mathrm{\&theta;}}_{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA00003437580300032.png"\; state="\{\{state\}\}">f</figure-callout>}2}\&le;10)\end{array}\right.---\left(17\right)$

With e _xAnd e _yCorresponding membership function value under each self-corresponding fuzzy language variable is respectively to e _xAnd e _yThe union of corresponding all membership function values is tried to achieve first and is occured simultaneously and the second common factor; The first common factor, the second common factor and B are asked the 3rd common factor, obtain an exact value; Occur simultaneously respectively to two μ that export with the 3rd _N(θ _f1) and μ _P(θ _f1) cut, obtain the membership function value of output fuzzy language variable;

4) adopt the sharpening interface to obtain output angle;

Because initial conditions may take a plurality of input fuzzy language variablees simultaneously, according to the derivation of fuzzy rule, will form a plurality of Output rusults, therefore need to carry out precision to the result of fuzzy reasoning.

At first get that in each Output rusults, the maximum membership degree functional value in two fuzzy programming angles compares, select the Output rusults at the place of membership function value maximum to export as final result, if in each Output rusults process relatively, the maximum membership degree function that a plurality of results occur is maximum and equal side by side, selects delivery to stick with paste less the exporting as final result of language amount in Output rusults; Then will export membership function value substitution formula 14-17 in the fuzzy language result function of inverting, obtain the two actual output angles in joint; The fuzzy language amount is pressed N, None, P ascending order.

The below carries out analytic explanation with one of the situation that may run in the real work of deicing inspection robot:

If mechanical arm tail end error (e _x, e _y) be (0.02 ,-0.02), obstacle judgement B=1 be danger.

Input obfuscation: with (e _x, e _y) substitution membership function formula 3-8 separately, obtain e _xMembership function value in fuzzy language N situation is 0.667, and the membership function value in fuzzy language Z situation is 0.333, and the membership function value in fuzzy language P situation is 0.Due to e _yWith e _xError is identical, and membership function is identical, so e _yMembership function value in fuzzy language N situation is 0.667, and the membership function value in fuzzy language Z situation is 0.333, and the membership function value in fuzzy language P situation is 0.

Fuzzy rule instructs: because the membership function value in fuzzy language P situation is 0, mamdani the reasoning results afterwards is 0, so do not consider.Therefore, consider the different fuzzy language values that output variable takies, (e _x, e _y) four kinds of combinations are arranged, be respectively (N, N), (N, Z), (Z, N), (Z, Z).According to the fuzzy rule in table 1, obtain output variable (θ _f1, θ _f2) four kinds of fuzzy language results are arranged, be respectively (N, None), (None, P), (None, P), (None, None).

Fuzzy reasoning: take the first situation as example, according to Mamdani reasoning, (e _x, e _y) being in (N, N) situation lower time, the membership function value of input is (0.667,0.667), B=1 during reasoning, at first will input e _xMembership function value μ when delivery is stuck with paste language amount N _N(e _x)=0.667 and e _xAll membership function μ _N(e _x) ∪ μ _Z(e _x) ∪ μ _P(e _xThe first common factor is made in)=1, obtains α _ex=0.667, in like manner, e _yWith e _yAll membership functions are done second and are occured simultaneously, and can obtain α _ey=0.667, in addition, the height α of B _BBe 1, with α _ex, α _eyAnd α _BMake the 3rd common factor, obtain α _ex∩ α _ey∩ α _B=0.667, then the 3rd exact value that occurs simultaneously is cut output fuzzy membership function formula 9-12, obtain exporting the exact value at each fuzzy programming angle.This process is known μ _N(θ _f1), μ _P(θ _f1), μ _N(θ _f2) and μ _P(θ _f2) be 0.667, to the formula 9-12 function of inverting, obtain the first joint θ _f1The output fuzzy language is that angle value under N is that angle value under-6.67, P is 6.67, second joint θ _f2The output fuzzy language is that angle value under N is that angle value under-6.67, P is 6.67.According to the fuzzy rule guiding, the first situation is output as (N, None) when being output as (N, N), therefore, and output variable (θ _f1, θ _f2) the membership function value be (0.667,0), output angle corresponds to (6.67,0).

In like manner can be input as (N, Z), (Z, N), the output language variable during (Z, Z) situation is respectively (None, P), (None, P) (None, None), its output variable (θ _f1, θ _f2) the membership function value be respectively (0,0.333), (0,0.333), (0,0).

Sharpening interface: adopt the SOM clarification method, in these Output rusults, get comparing of membership function maximum in two fuzzy programming angles, select the result at place of membership function maximum as final output.In this example, the situation that output function is maximum and equate does not appear in the membership function value of corresponding output (N, None) maximum (0.667〉0.333=0.333〉0) when being input as (N, N).Therefore get and be input as conduct output in (N, N) situation, therefore the output fuzzy language of mechanical arm is (N, None), corresponding output angle determined value is (6.67,0).

As shown in Figure 6, in above-mentioned steps (3), in Fuzzy particle swarm artificial obstacle detouring planning strategy, population on-line optimization partial design is as follows:

1. optimization object is selected: by self-correcting parameter k=(k is set after the fuzzy programming device ₁, k ₂) k ₁, k ₂∈ [0,2] utilizes particle cluster algorithm to carry out on-line optimization to it, makes the fuzzy programming expectation angle after optimization be (k ₁θ _f1, k ₂θ _f2).

2. update mode is determined: adopt the elementary particle group to optimize update mode, be expressed as:

v _ij(n+1)=wv _ij(n)+r ₁(n)c ₁[pbest _ij(n)-x _ij(n)]+r ₂(n)c ₂[gbest _ij(n)-x _ij(n)](18)

x _ij(n+1)=x _ij(n)+v _ij(n+1)(19)

Wherein: i is for initializing correction parameter group number, and j is the dimension of every group of correction parameter, v _ij(n) diverse vector of the j dimension that is correction parameter group i in the n time iteration, x _ij(n) be the numerical value of correction parameter group i j dimension in the n time, r ₁, r ₂Be independent random function, c ₁, c ₂Be acceleration weight, c ₁, c ₂All value is that 1.2, w is inertia weight;

Inertia weight w upgrades by the mode of linear decrease, is shown below:

$w=w_{\max }-n\frac{w_{\max }-w_{\min }}{n_{\max }}---\left(20\right)$

Wherein: n is the current iteration number of times, n _maxBe total iterations.w _maxBe 0.9, w _minBe 0.4.

3. fitness function design:

In order to reduce the redundancy of mechanical arm tail end track, designed the fitness function that comprehensive path and impact point attract, in the situation that guarantee that mechanical arm realizes that path planning is the shortest can arrive impact point.

Mechanical arm can be expressed as through shortest path: at t arbitrarily constantly, the distance of mechanical arm tail end and initial point position is the integration minimum in time, and fitness function is arranged:

$\mathrm{Min}:f_1\left(t\right)\underset{0}{\overset{t}{\&Integral;}}\sqrt{{[x\left(t\right)-x_0]}^2+{[y\left(t\right)-y_0]}^2}\mathrm{dt}---\left(21\right)$

Wherein: x _o, y _oBe mechanical arm tail end initial point position, x (t), y (t) are t mechanical arm tail end position constantly.

Impact point attracts to be presented as that between mechanical arm tail end and impact point, air line distance is minimum.

$\mathrm{Min}:f_2\left(t\right)=\sqrt{{[x\left(t\right)-x_g]}^2+{[y\left(t\right)-y_g]}^2}---\left(22\right)$

Therefore, the comprehensive fitness degree function is the weighting that shortest path and impact point attract, and is expressed as:

Min:f(t)=s ₁f ₁(t)+s ₂f ₂(t)(23)

For guaranteeing to realize that path planning is the shortest in situation that target can reach.Therefore the weight factor of Offered target point attraction is larger.Make s ₁Be 0.4, s ₂Be 0.6.

4. population on-line optimization step design:

According to fuzzy programming output and particle group optimizing mode, the online design Optimization Steps is as follows:

Step1: t at a time, the fuzzy programming device is cooked up fuzzy operating angle (θ according to the working environment of mechanical arm _f1(t), θ _f2(t)), simultaneously, population initializes, and produces initial value and the diverse vector of correction parameter group.

Step2: each correction parameter group forms (k by two correction parameters ₁(t), k ₂(t)), in conjunction with present Fuzzy operating angle (θ _f1(t), θ _f2(t)), substitution fitness function formula 18 is calculated.

Step3: at iterations n _maxIn, according to formula 18, formula 19, formula 20 is upgraded correction parameter group numerical value.Each iteration is all selected local optimum parameter group pbest and the parameter group gbest of global optimum.

Step4: after iteration is completed, with the correction parameter (k of the parameter group gbest of global optimum output as moment t ₁, k ₂).The excitation mechanical arm is by (k ₁θ _f1(t), k ₂θ _f2(t)) action.

Step5: the mechanical arm tail end sensor obtains next t+1 terminal position constantly, turns Step1.

As shown in Figure 7, in above-mentioned steps (4), the neural network adaptive controller partial design is as follows:

The output of adaptive neural network

Can be expressed as:

$\hat{h}={\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(24\right)$

Wherein: Be the neutral net input,

Be the connection weight value matrix of hidden layer to output layer, the attitude that they can be different according to mechanical arm and each joint angles desired value real-time update are to adapt to different demands for control; G (χ)=[g ₁(χ), g ₂(χ) ..., g _N(χ)] ^TBe the RBF vector.g _j(χ) (1≤j≤N) has following form

$g_j\left(\mathrm{\&chi;}\right)=\exp (-\frac{{\left|\right|\mathrm{\&chi;}-{\mathrm{\&mu;}}_j\left|\right|}^2}{{\mathrm{\&sigma;}}_j^2})---\left(25\right)$

Wherein: N is the number of hidden layer neuron; μ _jCentered by vector, σ _jBe the width vector.

Definition: e=k θ _f-θ and sliding-mode surface

Λ is symmetric positive definite matrix.The control inputs of mechanical arm can be expressed as:

$\mathrm{\&tau;}=-K_{p}S+{\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(26\right)$

$\stackrel{\stackrel{\&CenterDot;}{^}}{W}=-\mathrm{\&Gamma;g}\left(\mathrm{\&chi;}\right)s^T---\left(27\right)$

Wherein: K _PWith Γ be symmetric positive definite matrix.

2. the stability of a system proves:

The kinetics equation of mechanical arm can be expressed as:

$\mathrm{\&tau;}=M\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})\stackrel{\&CenterDot;}{\mathrm{\&theta;}}+G\left(\mathrm{\&theta;}\right)---\left(28\right)$

Wherein, M (θ) is inertia matrix,

Be centripetal acceleration coefficient and Corioli's acceleration coefficient, G (θ) is the gravity item.Consider

Therefore,

$S=\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f+\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)---\left(29\right)$

$\stackrel{\&CenterDot;}{S}=\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f+\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)---\left(30\right)$

In conjunction with the Neural Network Adaptive Control input of design, formula (28) can be expressed as:

$M\left(\mathrm{\&theta;}\right)[\stackrel{\&CenterDot;}{S}-{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)]+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})[S-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)]+G\left(\mathrm{\&theta;}\right)=\mathrm{\&tau;}$

$M\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;}{S}+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})S=\mathrm{\&tau;}-h$ (31)

Wherein, h is $h=M\left(\mathrm{\&theta;}\right)[{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)]+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})[{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)]+G\left(\mathrm{\&theta;}\right),$ Generally, h exports with neural network adaptive controller

And evaluated error

Sum represents.

$h=\hat{h}+\tilde{h}={\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)+{\tilde{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(32\right)$

Convolution (24) and formula (26), the output of Neural Network Adaptive Control can be expressed as:

$\mathrm{\&tau;}=-k_{p}s+\hat{h}---\left(33\right)$

With formula (31) and formula (32) substitution formula (33),

$M\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;}{S}+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})S=-k_{p}S+{\tilde{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(34\right)$

Get Li Yapu with regard to function:

$V=\frac{1}{2}{S}^{T}M\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;}{S}+\frac{1}{2}\mathrm{tr}\left(\tilde{W}{\mathrm{\&Gamma;}}^{-1}\tilde{W}\right)---\left(35\right)$

Differentiate obtains:

$\stackrel{\&CenterDot;}{V}=S^{T}M\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;}{S}+\frac{1}{2}{S}^T\stackrel{\&CenterDot;}{M}\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;}{S}+\mathrm{tr}\left(\tilde{W}{\mathrm{\&Gamma;}}^{-1}\stackrel{\stackrel{\&CenterDot;}{~}}{w}\right)$

$=S^T(\mathrm{\&tau;}-f-C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})S)+\frac{1}{2}{S}^T\stackrel{\&CenterDot;}{M}\left(\mathrm{\&theta;}\right)\stackrel{\&CenterDot;}{S}+\mathrm{tr}\left(\tilde{W}{\mathrm{\&Gamma;}}^{-1}\stackrel{\stackrel{\&CenterDot;}{~}}{W}\right)$ (36)

Utilize

Skew symmetry know

Formula (32) and formula (33) substitution formula (36) can be got:

$\stackrel{\&CenterDot;}{V}=S^T(\mathrm{\&tau;}-h)+\mathrm{tr}\left(\tilde{W}{\mathrm{\&Gamma;}}^{-1}\stackrel{\stackrel{\&CenterDot;}{~}}{W}\right)$

$=S^T[\mathrm{\&tau;}-\hat{h}-{\tilde{W}}^{T}g\left(\mathrm{\&chi;}\right)]+\mathrm{tr}\left(\tilde{W}{\mathrm{\&Gamma;}}^{-1}\stackrel{\stackrel{\&CenterDot;}{~}}{W}\right)$

$=-S^T{k}_{p}S-S^T{\tilde{W}}^{T}g\left(\mathrm{\&chi;}\right)+\mathrm{tr}\left(\tilde{W}{\mathrm{\&Gamma;}}^{-1}\stackrel{\stackrel{\&CenterDot;}{~}}{W}\right)$

$=-S^T{k}_{p}S$

(37)

Due to

To bear semidefinite, and K _pBe positive definite, work as

The time, S ≡ 0 substitution formula (29), as can be known

By the LaSalle theorem as can be known, the deicing inspection robot forearm overall situation is progressive stable.Namely from arbitrary initial conditions Set out, θ → θ is all arranged _d,

The output of neutral net is the compensation term matrix in each joint, and neutral net is input as

Namely that current angle information (angle and angular speed) according to population fuzzy programming angle information (angular speed and angular acceleration) and each joint goes the real-time compensation term of approaching

$M\left(\mathrm{\&theta;}\right)[{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)]+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})[{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)]+G\left(\mathrm{\&theta;}\right).$

In control law

S is comprised of e, and S is tending towards 0 and just can causes e to be tending towards 0, just can guarantee tracking performance in theory, and this provides in stability proof part before.In this control law, first half is simple proportion control, and latter half is the compensation term that neutral net is approached, if the neutral net perfection approaches, this complicated system just can become simple proportion control.

Being weight matrix, is real-time update, provides the formula of renewal

S can find out it is an approximate product of error, Can be according to error current S and input and output

Real-time update.A dynamic neural network control process that Here it is.

Below with a concrete application example, operation of the present invention is elaborated, the mechanical arm active obstacle planing method of this patent is applied to the deicing high-voltage power transmission line inspection robot and reaches the standard grade in work, mainly from accessible track following with meet barrier obstacle detouring planning two aspects and embody its validity.Specifically arrange as follows:

1. the neural network adaptive controller parameter is set, and it is 20 that hidden layer number N is set, width vector μ _jBe set to the random function of even distribution [10,10], center vector σ _jUnification is set to 2, connects the weights initial value and is arranged on [1,1] interior random distribution.

2. deicing inspection robot mechanical arm tail end initial point coordinate is set is (0.2,0.4), particle group optimizing partly arranges correction parameter (k ₁, k ₂) scope is [± 0.5, ± 1.5], the particle dimension is 2, particle position initialization scope is [2,2].In order to satisfy deicing inspection robot trajectory planning real-time demand, it is 5 that the population iterations is set, and population is 30, acceleration weight c1=c2=1.5.The design mechanical arm moves to eight directions, each orientation assigned address is respectively: the top---and (0.2,0.5), the front upper place---(0.3,0.5),---(0.3,0.4), front lower place---(0.3,0.3), below---(0.2, the place ahead, 0.3), the back lower place---(0.18,0.38), rear---(0.18,0.4) and back upper place---(0.18,0.5).Arrange in the mechanical arm motion space and do not contain barrier, namely the deicing inspection robot works under safe mode.

As shown in Figure 8, provided deicing inspection robot mechanical arm in accessible situation to the result of eight orientation reachability tests on every side, marked starting point and impact point in each task.Can find out just that from figure the arrival impact point that mechanical arm tail end can be level and smooth in each task has satisfied the accessibility requirement of the accessible track following of deicing inspection robot mechanical arm.

3. deicing inspection robot mechanical arm is set continuous signal is followed the tracks of, particle group optimizing partly arranges the same, and the design tracking signal is to be the center of circle with (0.31,0.4), and initial point is (0.21,0.4), and radius is 0.1 circumference, is expressed as:

$\left\{\begin{array}{cc}{x}_{\mathrm{signal}}\left(t\right)=-0.1\cos t+0.31& t\&GreaterEqual;0\\ y_{\mathrm{signal}}\left(t\right)0.1\sin t+0.4& t\&GreaterEqual;0\end{array}\right.---\left(38\right)$

As shown in Figure 9, provided the deicing inspection robot mechanical arm tracking situation to continuous signal in accessible situation, in two curves in Fig. 9,---expression continuous signal variation track, the pursuit path of------expression deicing inspection robot mechanical arm tail end, as can be seen from the figure, continuous signal is without the mechanical arm tail end initial point, mechanical arm can followed the tracks of continuous signal under the effect of population fuzzy programming device, and error is gradually zero, keeps stable.Therefore illustrate that this population fuzzy programming strategy can guarantee the good tracking performance of mechanical arm.

As shown in Figure 9, provided the deicing inspection robot mechanical arm tracking situation to continuous signal in accessible situation, in two curves in Fig. 9,---expression continuous signal variation track, the pursuit path of------expression deicing inspection robot mechanical arm tail end, as can be seen from the figure, continuous signal is without the mechanical arm tail end initial point, mechanical arm can followed the tracks of continuous signal under the effect of population fuzzy programming device, and error is gradually zero, keeps stable.Therefore illustrate that this population fuzzy programming strategy can guarantee the good tracking performance of mechanical arm.

4. the obstacle performance of this planning strategy is tested, deicing inspection robot mechanical arm tail end starting point is set is (0.2,0.4), be arranged on (0.35,0.4) located obstacle, the expanded radius of circle of this obstacle is 0.3, the detection range d=0.3 of mechanical arm tail end.The front arm tip terminal point is (0.5,0.4).Particle group optimizing partly arranges the same

As shown in figure 10, two-dimensional space information and the deicing inspection robot mechanical arm obstacle detouring geometric locus of barrier on the original state of deicing inspection robot mechanical arm and final state, ultra-high-tension power transmission line and circuit have been provided,------expression mechanical arm tail end variation track.As can be seen from the figure, in the situation that provide target location and obstacle location information, the obstacle detouring action can be made decisions on one's own, be planned to deicing inspection robot mechanical arm, has higher independence and intelligent.

Claims (5)

1. the active obstacle planing method of deicing high-voltage power transmission line inspection robot, is characterized in that, comprises the steps:

Step 1: utilize the laser radar detection ambient condition information that is arranged on mechanical arm tail end, obtain the operation the place ahead obstacle signal B of robot;

Step 2: deicing robot observes by the camera that carries the position of moving the place ahead cable, and 30cm place, robot forearm the place ahead cable is set to the desired locations (x of mechanical arm tail end current time _g, y _g); Difference (e according to mechanical arm tail end current location and desired locations _x, e _y) and current obstacle signal B, utilize the fuzzy programming device to cook up the fuzzy programming angle in each joint of the front surface mechanical arm of traffic direction;

Step 3: can reach principle according to shortest path and target, utilize particle cluster algorithm to carry out on-line optimization to the fuzzy programming angle, obtain the population fuzzy programming expected angle in each joint of the front surface mechanical arm of traffic direction;

Step 4: with the input of each joint population fuzzy programming angle as neural network adaptive controller, neural network adaptive controller is exported each joint control moment τ, instructs the mechanical arm action.

2. the active obstacle planing method of deicing high-voltage power transmission line inspection robot according to claim 1, it is characterized in that, obtain the operation the place ahead obstacle signal B of robot in described step 1 and refer to according to the obstacle detouring judgement critical distance A that sets, judge whether the distance of the Obstacle Position that mechanical arm tail end current location and Airborne Lidar measure surpasses the obstacle detouring judgement critical distance A that sets:

e _xb=|x-x _b| (1)

e _yb=|y-y _b|(2)

Wherein, x _b, y _bBe the two-dimensional coordinate position of barrier, x, y is the mechanical arm tail end current location, works as e _xb≤ A and e _ybDuring≤A, judgement deicing inspection robot moves the place ahead has the barrier of chance dangerous, output obstacle signal B=1, otherwise B=0.

3. the active obstacle planing method of deicing high-voltage power transmission line inspection robot according to claim 2, is characterized in that, in described step 2 according to the difference (e of mechanical arm tail end current location and desired locations _x, e _y) and present case under obstacle signal B, each joint fuzzy programming angle of mechanical arm when utilizing the fuzzy programming device to cook up deicing robot to travel forward, detailed process is as follows:

1) determine fuzzy programming device input message, and the input data carried out Fuzzy processing:

With mechanical arm tail end control information (e _x, e _y) and obstacle signal B input as the fuzzy programming device:

$\left\{\begin{array}{c}{e}_x=x-x_g,e_x\&Element;[-\mathrm{0.4,0.36}]\\ e_y=y-y_g,e_y\&Element;[-\mathrm{0.4,0.56}]\end{array}\right.B\&Element;\left[\mathrm{0,1}\right]$

(x _g, y _g) be the desired locations that 30cm place, robot forearm the place ahead cable is set to the mechanical arm tail end current time;

Fuzzy programming device input message is carried out Fuzzy processing, mechanical arm tail end error of coordinate (e _x, e _y) corresponding 3 fuzzy language variablees { P}={ " bears " for N, Z, " zero ", " just " }, fuzzy language variable N, Z and P respectively with membership function μ _N, μ _ZAnd μ _PCorresponding one by one, the codomain scope of membership function is [0,1]; The input variable of membership function is mechanical arm tail end error of coordinate (e _x, e _y):

${\mathrm{\&mu;}}_N\left(e_x\right)=\left\{\begin{array}{cc}1& e_x\&le;-0.03\\ \frac{e_x}{-0.03}& -0.03<e_x\&le;0\\ 0& e_x>0\end{array}\right.---\left(3\right){\mathrm{\&mu;}}_N\left(e_y\right)=\left\{\begin{array}{cc}1& e_y\&le;-0.03\\ \frac{e_y}{-0.03}& -0.03<e_y\&le;0\\ 0& e_y>0\end{array}\right.---\left(6\right)$

${\mathrm{\&mu;}}_p\left(e_x\right)=\left\{\begin{array}{cc}0& e_x\&le;0\\ \frac{e_x}{0.03}& 0<e_x\&le;0.03\\ 1& e_x>0.03\end{array}\right.---\left(5\right){\mathrm{\&mu;}}_p\left(e_y\right)=\left\{\begin{array}{cc}0& e_y\&le;0\\ \frac{e_y}{0.03}& 0<e_y\&le;0.03\\ 1& e_y>0.03\end{array}\right.---\left(8\right)$

Choose the membership function value and be non-zero fuzzy language variable as the fuzzy language variable of input message;

2) obtain output fuzzy language variable according to fuzzy rule;

According to fuzzy language variable and the obstacle signal B of input message, inquiry fuzzy rule design table, as shown in table 1, obtain exporting accordingly fuzzy language variable { None, N, P}, wherein, None represents that mechanical arm is failure to actuate, and N represents the mechanical arm operating angle for negative, and P represents that the mechanical arm operating angle is for just;

The design of table 1 fuzzy rule

3) output fuzzy language variable is carried out fuzzy reasoning, obtain membership function value corresponding to output fuzzy language variable; Adopt the Mamdani reasoning, the membership function that output fuzzy language variable is corresponding is as follows:

${\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f1}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{f1}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{f1}\&le;10)\end{array}\right.---\left(9\right){\mathrm{\&mu;}}_N\left({\mathrm{\&theta;}}_{f2}\right)=\left\{\begin{array}{cc}\frac{{\mathrm{\&theta;}}_{f2}}{-10}& (-10\&le;{\mathrm{\&theta;}}_{f2}<0)\\ 0& (0\&le;{\mathrm{\&theta;}}_{f2}\&le;10)\end{array}\right.---\left(11\right)$

${\mathrm{\&mu;}}_p\left({\mathrm{\&theta;}}_{f1}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{f1}<0)\\ \frac{{\mathrm{\&theta;}}_{f1}}{10}& (0\&le;{\mathrm{\&theta;}}_{f1}\&le;10)\end{array}\right.---\left(10\right){\mathrm{\&mu;}}_p\left({\mathrm{\&theta;}}_{f2}\right)=\left\{\begin{array}{cc}0& (-10\&le;{\mathrm{\&theta;}}_{f2}<0)\\ \frac{{\mathrm{\&theta;}}_{f2}}{10}& (0\&le;{\mathrm{\&theta;}}_{f2}\&le;10)\end{array}\right.---\left(12\right)$

With e _xAnd e _yCorresponding membership function value under each self-corresponding fuzzy language variable is respectively to e _xAnd e _yThe union of corresponding all membership function values is tried to achieve first and is occured simultaneously and the second common factor; The first common factor, the second common factor and B are asked the 3rd common factor, obtain an exact value; Occur simultaneously respectively to two μ that export with the 3rd _N(θ _f1) and μ _P(θ _f1) cut, obtain the membership function value of output fuzzy language variable;

4) adopt the sharpening interface to obtain output angle;

At first get that in each Output rusults, the maximum membership degree functional value in two fuzzy programming angles compares, select the Output rusults at the place of membership function value maximum to export as final result, if in each Output rusults process relatively, the maximum membership degree function that a plurality of results occur is maximum and equal side by side, selects delivery to stick with paste less the exporting as final result of language amount in Output rusults; Then will export membership function value substitution formula 9-12 in the fuzzy language result function of inverting, obtain the two actual output angles in joint; The fuzzy language amount is pressed N, None, P ascending order.

4. the active obstacle planing method of deicing high-voltage power transmission line inspection robot according to claim 3, is characterized in that, it is as follows that the particle cluster algorithm in described step 3 carries out the on-line optimization process to the fuzzy programming angle:

1): optimization object is selected: self-correcting parameter k=(k is set ₁, k ₂) k ₁, k ₂∈ [0,2] makes the fuzzy programming expectation angle after optimization be (k ₁θ _f1, k ₂θ _f2);

2) update mode is determined: adopt the elementary particle group to optimize update mode, be expressed as:

vi _j(n+1)=wvi _j(n)+r ₁(n)c ₁[pbesti _j(n)-xi _j(n)]+r ₂(n)c ₂[g ^besti _j(n)-xi _j(n)] (13)

xi _j(n+1)=xi _j(n)+vi _j(n+1) (14)

Wherein: i is for initializing correction parameter group number, and j is the dimension of every group of correction parameter, v _ij(n) diverse vector of the j dimension that is correction parameter group i in the n time iteration, x _ij(n) be the numerical value of correction parameter group i j dimension in the n time, r ₁, r ₂Be independent random function, c ₁, c ₂Be acceleration weight, c ₁, c ₂All value is that 1.2, w is inertia weight;

Inertia weight w upgrades by the mode of linear decrease, is shown below:

$w=w_{\max }-n\frac{w_{\max }-w_{\min }}{n_{\max }}---\left(15\right)$

Wherein: n is the current iteration number of times, n _maxBe total iterations, w _maxBe 0.9, w _minBe 0.4;

3) fitness function design:

Min:f(t)=s ₁f ₁(t)+s ₂f ₂(t) （16）

Wherein, s ₁Be 0.4, s ₂Be 0.6;

At any t constantly, the distance of mechanical arm tail end and initial point position is the integration minimum in time, fitness function f ₁(t) be:

$\mathrm{Min}:f_1\left(t\right)\underset{0}{\overset{t}{\&Integral;}}\sqrt{{[x\left(t\right)-x_0]}^2+{[y\left(t\right)-y_0]}^2}\mathrm{dt}---\left(17\right)$

Impact point attracts to be presented as that between mechanical arm tail end and impact point, air line distance is minimum, fitness function f ₂(t) be;

$\mathrm{Min}:f_2\left(t\right)=\sqrt{{[x\left(t\right)-x_g]}^2+{[y\left(t\right)-y_g]}^2}---\left(18\right)$

Wherein: x _o, y _oBe mechanical arm tail end initial point position, x (t), _y(t) be t mechanical arm tail end position constantly, x _gAnd y _gExpression impact point position;

4) population on-line optimization:

Step1: population initializes, and produces initial value and the diverse vector of correction parameter group, with (the k of current time t ₁θ _f1, k ₂θ _f2) 16 calculating of substitution fitness function formula;

Step2: at iterations n _maxIn, according to formula 13, formula 14, formula 15 is upgraded correction parameter group numerical value.Each iteration is all selected local optimum parameter group pbest and the parameter group gbest of global optimum;

Step3: after iteration is completed, with the correction parameter (k of the parameter group gbest of global optimum output as moment t ₁, k ₂).The excitation mechanical arm is by (k ₁θ _f1(t), k ₂θ _f2(t)) action;

Step4: the mechanical arm tail end sensor obtains next t+1 terminal position constantly, turns Step1.

5. the active obstacle planing method of deicing high-voltage power transmission line inspection robot according to claim 4, is characterized in that, the neural network adaptive controller in described step 4 comprises proportion control-K _pS and neural network control

Described neural network adaptive controller is shown below

$\mathrm{\&tau;}=-K_{p}S+{\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(19\right)$

$\stackrel{\stackrel{\&CenterDot;}{^}}{W}=-\mathrm{\&Gamma;g}\left(\mathrm{\&chi;}\right)S^T---\left(20\right)$

Wherein: τ is control moment, K _pWith Γ be symmetric positive definite matrix, K _pFor the proportion control parameter is set to $\left[\begin{array}{cc}5& 0\\ 0& 5\end{array}\right],$

Be set to 8 * 8 unit matrix, g (χ)=[g ₁(χ), g ₂(χ) ..., g _N(χ)] ^T, g _j(χ) (1≤j≤r) is the RBF vector,

Be the connection weight value matrix of hidden layer to output layer,

Initial value is made as 8 * 8 matrixes in [1,1] interior random distribution;

S is the synovial membrane face

E=k θ _f-θ, wherein k θ _fBe population fuzzy programming angle, k θ _f=[k ₁θ _f1, k ₂θ _f2], θ is the current angle of mechanical arm, θ=[θ ₁, θ ₂], Λ is symmetric positive definite matrix $\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right];$

Adaptive neural network is shown below:

$\hat{h}{\hat{W}}^{T}g\left(\mathrm{\&chi;}\right)---\left(21\right)$

Wherein:

Be the neutral net input,

Be the first joint planning angular speed,

Be the first joint planning angular acceleration,

Be second joint planning angular speed,

Be second joint planning angular acceleration, θ ₁Be the first joint action angle,

Be the first joint action angular speed, θ ₂Be the second joint operating angle, Be second joint operating angle speed,

Be the connection weight value matrix of hidden layer to output layer,

Initial value is made as 8 dimension unit matrixs, and the attitude different according to mechanical arm and each joint angles desired value real-time update are to adapt to different demands for control; G (χ)=[g ₁(χ), g ₂(χ) ..., g _N(χ)] ^TBe RBF vector, g _j(χ) (1≤j≤r) has following form:

$g_j\left(\mathrm{\&chi;}\right)=\exp (-\frac{{\left|\right|\mathrm{\&chi;}-{\mathrm{\&mu;}}_j\left|\right|}^2}{{\mathrm{\&sigma;}}_j^2})---\left(22\right)$

Wherein: r is the number of hidden layer neuron; μ _jCentered by vector, σ _jBe the width vector;

The output of adaptive neural network

The target of approaching be compensation term, compensation term is shown below:

$M\left(\mathrm{\&theta;}\right)[{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\stackrel{\&CenterDot;}{\mathrm{\&theta;}}-{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f)]+C(\mathrm{\&theta;},\stackrel{\&CenterDot;}{\mathrm{\&theta;}})[{\stackrel{\&CenterDot;}{\mathrm{\&theta;}}}_f-\mathrm{\&Lambda;}(\mathrm{\&theta;}-{\mathrm{\&theta;}}_f)]+G\left(\mathrm{\&theta;}\right);---\left(23\right)$ In adaptive neural network, hidden layer number r is 20, center vector μ _jBe set to the random function of even distribution [10,10], width vector σ _jAll be set to 2.

Images