Disclosure of Invention
According to the invention, according to the imperfection and the deficiency of the prior background technology, the inclusion control of output feedback of the multi-double-link mechanical arm under a directed communication network is realized based on the active disturbance rejection and inversion technology, an extended state observer is designed in each sub-controller by utilizing the input and the output of a follower to reconstruct the system state of the double-link mechanical arm, and unknown dynamics are compensated in real time through the estimation of the extended state quantity, so that the designed inclusion controller of the output feedback has disturbance rejection. Further, the derivative of the virtual control law in the sub-controllers is estimated by means of a tracking differentiator, so that the problem of computation complexity explosion is overcome. The controller of the multi-double-connecting-rod mechanical arm based on active disturbance rejection and inversion technology only depends on the output position, and the problems of unknown dynamics in a system, simplified complex derivation operation and the like can be effectively solved.
In order to achieve the technical purpose, the invention adopts the following technical scheme:
multiple-double-connection based on output positionThe rod mechanical arm comprises a controller, N double-link mechanical arms containing unknown dynamics are regarded as followers, a networked system formed by connecting the N double-link mechanical arms with M leaders through a one-way topological graph is used as a controlled object, the leaders form a static convex hull, the controllers are designed by utilizing the positions output by
joints 1 and
joints 2 of the followers, so that the output of the followers is converged in the static convex hull, the controller structure of the ith (i is not less than 1 and not more than N) follower comprises a1 st sub-controller and a2 nd sub-controller, and the input ends of the 1 st and the 2 nd sub-controllers and a directed graph
The output ends of the 2 sub-controllers are all connected with the input end of the ith follower,
the 1 st sub-controller comprises a first extended state observer unit, an error surface s
i,1,1The device comprises an arithmetic unit, a first nonlinear arithmetic unit, a first comparator unit, a first tracking differentiator unit and a second nonlinear arithmetic unit; two input ends of the first extended state observer unit are respectively the output y of the
ith follower joint 1
i,1And the output u of the second non-linear operation unit
i,1(ii) a Error surface s
i,1,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,1,dAdjacent communication a
ijAnd the output of the first extended state observer unit
The input ends of the first nonlinear operation units are directed graphs respectively
Middle observer state
Output y of jth follower (j is more than or equal to 1 and less than or equal to N)
j,1Adjacent communication a
ijOutput y of i-
th follower joint 1
i,1Output of the first extended state observer unit
Sum error surface s
i,1,1An output of the arithmetic unit; the first comparator unit being an error surface s
i,1,2The input end of which is the output of the first extended state observer unit
And the output x of the first non-linear operation unit
i,1,2,d(ii) a The input end of the first tracking differentiator unit is the output x of the first nonlinear operation unit
i,1,2,d(ii) a The input end of the second nonlinear operation unit is a directed graph
Intermediate adjacent communication a
ijOutput y of i-
th follower joint 1
i,1Error surface s
i,1,1Output of arithmetic unit, output v of first tracking differentiator unit
i,1,2An output s of the first comparator unit
i,1,2And the output of the first extended state observer unit
The 2 nd sub-controller comprises a second extended state observer unit, an error surface s
i,2,1The device comprises an arithmetic unit, a third nonlinear arithmetic unit, a second comparator unit, a second tracking differentiator unit and a fourth nonlinear arithmetic unit; two input ends of the second extended state observer unit are respectively the output y of the
ith follower joint 2
i,2And the output u of the fourth non-linear operation unit
i,2(ii) a Error surface s
i,2,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,2,dAdjacent communication a
ijAnd the output of the second extended state observer unit
The input ends of the third nonlinear operation units are directed graphs respectively
Middle observer state
Output y of
jth follower joint 2
j,2Adjacent communication a
ijOutput y of the i-
th follower joint 2
i,2The output of the second extended state observer unit
Sum error surface s
i,2,1An output of the arithmetic unit; the second comparator unit being an error surface s
i,2,2The input end of which is the output of the second extended state observer unit
And the output x of the third non-linear operation unit
i,2,2,d(ii) a The input end of the second tracking differentiator unit is the output x of the third nonlinear operation unit
i,2,2,d(ii) a The input end of the fourth nonlinear operation unit is a directed graph
Intermediate adjacent communication a
ijOutput y of the i-
th follower joint 2
i,2Error surface s
i,2,1Output of arithmetic unit, output v of second tracking differentiator unit
i,2,2The output s of the second comparator unit
i,2,2And the output of the second extended state observer unit
The multi-double-link mechanical arm based on the output position comprises a controller, a multi-agent network formed by connecting follower groups consisting of N double-link mechanical arms containing unknown dynamic states and M leaders through a one-way topological graph is considered, and each follower is communicated with at least one leader; the information communication between them can be made by directed graphs
Is shown in which
Is a set of nodes, n
iIndicating a double link arm i, n
jA double link robot arm j is shown,
is a collection of edges that are to be considered,
indicating that agent j can directly obtain the information of agent i;
is a contiguous matrix, a
ijThe definition is as follows:
the contiguous set of nodes i is defined as
Directed graph
Laplacian matrix of
The definition is as follows:
laplace matrix
Wherein D ═ diag [ D [ ]
1,…,d
N]As directed graphs
The degree matrix of (c) is,
the follower has at least one adjacent node, and the leader has no adjacent node, the Laplace matrix can
And (3) decomposition:
the mathematical model of the ith double-link mechanical arm in the follower is:
in the formula, q
i,1Is the angular position, q, of the ith double link
robot arm joint 1
i,2Is the angular position, q, of the ith double link
robot arm joint 2
i=[q
i,1 q
i,2]
TIs the connection angular position;
respectively, connecting angular velocity and acceleration, tau
i=[τ
i,1 τ
i,2]
TIs a control input, the system parameters are selected as follows:
In the formula,
a
3=m
2l
1l
c2cos(δ),a
4=m
2l
1l
c2sin(δ),I
1viscous coefficient of friction for turning the
joint 1 to
2Viscous coefficient of friction to which the
joint 2 turns,/
1Is the length of the
rod 1, m
1Is the mass of the
rod 1, m
2Is the mass of the
rod 2, delta is the angle between the
initial joint 2 and the longitudinal axis, l
c1Is the distance of the
joint 1 from the center of mass of the
bar 1, l
c2Distance of
joint 2 to the center of mass of
rod 2, further, let [ x [ ]
i,1,1 x
i,2,1]
T=[q
i,1 q
i,2]
T,
[u
i,1 u
i,2]
T=[τ
i,1 τ
i,2]
TThe equation of state for the ith follower can be found:
in the formula, y
i,1、y
i,2Output of the i-
th follower joint 1 and
joint 2, W, respectively
i,11=a
1+2a
2cos(y
i,1)+2a
4sin(y
i,2),
And
the inner unknown dynamic part of the ith follower dual link robotic arm.
The design method of the multi-double-connecting-rod mechanical arm containing controller based on the output position comprises the following steps:
A. design of the 1 st sub-controller:
a1, design of first extended state observer unit:
the two inputs of the first extended state observer unit are divided into the output y of the
ith follower joint 1
i,1And the output u of the second non-linear operation unit
i,1The output signal of the first extended state observer unit is obtained by calculation of the following formula
And
and
estimated signals of the first, second and third order state variables of the i-th follower's
joint 1, respectively:
wherein T is time, Ti,1,1、Ti,1,2And Ti,1,3All are the gains of the extended state observer, riE (1, ∞) is the parameter to be designed, function gi(τ) is fal composed of linear and fractional power functionsi(τ,δi,j1) function, j ═ 1,2,3, expressed as:
wherein, deltai,jE (0,1) is an adjustable parameter, sign () is a symbolic function, and the expression is as follows:
a2, error surface si,1,1Designing an arithmetic unit:
error surface s
i,1,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,1,dAdjacent communication a
ijAnd the output of the first extended state observer unit
Obtaining an error surface s by calculation of the following formula
i,1,1Output s of arithmetic unit
i,1,1:
A3, design of first nonlinear arithmetic unit:
the input ends of the first nonlinear operation units are directed graphs respectively
Middle observer state
Output y of
jth follower joint 1
j,1Adjacent communication a
ijOutput y of i-
th follower joint 1
i,1Output of the first extended state observer unit
Sum error surface s
i,1,1The output of the operation unit is calculated by the following formula to obtain the virtual control law x
i,1,2,d:
In the formula, T
i,1,1And T
j,1,1In order to extend the gain of the state observer,
r
i∈(1,∞)、r
j∈(1,∞)、q
i,1,1parameters to be designed are more than 0;
a4, design of first comparator unit:
the input of the first comparator unit is the output of the first extended state observer unit
And the output x of the first non-linear operation unit
i,1,2,dAn error surface s is obtained by calculation of the following formula
i,1,2:
A5, design of the first tracking differentiator unit:
the input end of the first tracking differentiator is the output x of the first nonlinear operation uniti,1,2,dThe output v of the tracking differentiator is obtained by calculation of the following formulai,1,2:
In the formula, vi,1,2Is the first non-linear operation unit output xi,1,2,dIs estimated, α ∈ (0,1) is the filter factor, λ > 0 is the speed factor;
design of a6 and second nonlinear operation unit:
the input end of the second nonlinear operation unit is a directed graph
Intermediate adjacent communication a
ijI th follower joint 1Output y
i,1Error surface s
i,1,1Output of arithmetic unit, output v of first tracking differentiator unit
i,1,2An output s of the first comparator unit
i,1,2And the output of the first extended state observer unit
The control law u is obtained by calculation of the following formula
i,1:
In the formula, T
i,1,2In order to extend the gain of the state observer,
r
i∈(1,∞)、q
i,2,1parameters to be designed are more than 0;
B. design of the 2 nd sub-controller:
b1, design of second extended state observer unit:
the two inputs of the second extended state observer unit are divided into the output y of the
ith follower joint 1
i,2And the output u of the fourth non-linear operation unit
i,2The output signal of the second extended state observer unit is obtained by calculation of the following formula
And
and
estimated signals of the first, second and third order state variables of the i-th follower's
joint 2, respectively:
in the formulaT is time, Ti,2,1、Ti,2,2And Ti,2,3All are the gains of the extended state observer, riE (1, infinity) is a parameter to be designed; b2, error surface si,2,1Designing an arithmetic unit:
error surface s
i,2,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,2,dAdjacent communication a
ijAnd the output of the second extended state observer unit
Obtaining an error surface s by calculation of the following formula
i,2,1Output s of arithmetic unit
i,2,1:
B3, design of third nonlinear operation unit:
the input ends of the third nonlinear operation units are directed graphs respectively
Middle observer state
Output y of jth follower joint 2
j,2Adjacent communication a
ijOutput y of the i-th follower joint 2
i,2The output of the second extended state observer unit
Sum error surface s
i,2,1The output of the operation unit is calculated by the following formula to obtain the virtual control law x
i,2,2,d:
In the formula, T
i,2,1And T
j,2,1Are the gains of the extended state observer,
r
i∈(1,∞)、r
j∈(1,∞)、q
i,2,1parameters to be designed are more than 0;
b4, design of second comparator cell:
the input of the second comparator unit is the output of the second extended state observer unit
And the output x of the third non-linear operation unit
i,2,2,dAn error surface s is obtained by calculation of the following formula
i,2,2:
B5, design of second tracking differentiator unit:
the input end of the second tracking differentiator is the output x of the third nonlinear operation uniti,2,2,dThe output v of the tracking differentiator is obtained by calculation of the following formulai,2,2:
In the formula, vi,2,2Is the third non-linear operation unit output xi,2,2,dIs estimated, α ∈ (0,1) is the filter factor, λ > 0 is the speed factor;
b6, design of fourth nonlinear operation unit:
the input end of the fourth nonlinear operation unit is a directed graph
Adjacent communication a
ijOutput y of the i-th follower joint 2
i,2Error surface s
i,2,1Output of arithmetic unit, output v of second tracking differentiator unit
i,2,2The output s of the second comparator unit
i,2,2And the output of the second extended state observer unit
The control law u is obtained by calculation of the following formula
i,2:
In the formula, T
i,2,2In order to extend the gain of the state observer,
r
i∈(1,∞)、q
i,2,2more than 0 is a parameter to be designed;
to this end, the control input u of the ith follower is obtainedi,1And ui,2。
Compared with the prior art, the technical scheme of the invention has the beneficial effects that:
according to the invention, according to the imperfection and the deficiency of the prior background technology, the inclusion control of output feedback of the multi-double-link mechanical arm under a directed communication network is realized based on the active disturbance rejection and inversion technology, an extended state observer is designed in each sub-controller by utilizing the input and the output of a follower to reconstruct the system state of the double-link mechanical arm, and unknown dynamics are compensated in real time through the estimation of the extended state quantity, so that the designed inclusion controller of the output feedback has disturbance rejection. Further, the derivative of the virtual control law in the sub-controllers is estimated by means of a tracking differentiator, so that the problem of computation complexity explosion is overcome. The controller of the multi-double-connecting-rod mechanical arm based on active disturbance rejection and inversion technology only depends on the output position, and the problems of unknown dynamics in a system, simplified complex derivation operation and the like can be effectively solved.
(1) The invention relates to a multi-double-link mechanical arm containing controller based on an output position, aiming at the difficult problem that part of system states of followers in the multi-double-link mechanical arm are unavailable, an extended state observer is designed through the input and the output of the followers to reconstruct the system state of each double-link mechanical arm, and the multi-double-link mechanical arm containing controller for outputting feedback is designed by utilizing the system states observed in real time.
(2) The invention relates to a multi-double-connecting-rod mechanical arm containing controller based on an output position, which solves the problem of complexity in calculation by utilizing the characteristic that a tracking differentiator can effectively estimate a complex nonlinear function derivative aiming at the problem of complex derivation of a virtual control law in the containing controller, and effectively avoids the complexity of design of the containing controller for output feedback of the multi-double-connecting-rod mechanical arm
(3) The invention relates to a multi-double-link mechanical arm containing controller based on an output position, aiming at the influence of unknown dynamics existing in the multi-double-link mechanical arm on the design of the containing controller, by means of the characteristic that an extended state observer does not depend on an accurate double-link mechanical arm model, and the unknown dynamics are compensated in real time through the estimation of an extended state quantity, so that the designed containing controller outputting feedback has the immunity.
Drawings
FIG. 1 is a schematic diagram of an ith follower output feedback structure including a controller according to the present invention;
FIG. 2 is a schematic view of the structure of the ith dual link robotic arm of the follower of the present invention;
FIG. 3 is a one-way topological diagram between a follower dual-link robotic arm and a leader of the present invention;
FIG. 4 is a diagram of a convex hull formed by a controller according to the present invention that converges follower output to a leader;
FIG. 5. Inclusion error e in the output of the joint 1 of four followers in the present invention i11,2,3, 4;
FIG. 6. Inclusion error e in the output of the joint 2 of four followers in the present invention i21,2,3, 4;
FIG. 7 eight control laws u for four followers in the present invention ij1,2,3,4, j is 1, 2;
FIG. 8a is a graph of the observed effect of a first state in a first extended state observer of a second follower of the present invention;
FIG. 8b is a graph of the observed effect of a second state in the first extended state observer of a second follower of the present invention;
FIG. 9a is a graph of the observed effect of a first state in a second extended state observer of a second follower of the present invention;
FIG. 9b is a graph of the observed effect of a second state in a second extended state observer of a second follower of the present invention;
FIG. 10a shows the parameter r in the present inventioniA plot of the absolute value of the estimated error for the first state in the first extended state observer for the second follower at 1,2,3,4 changes;
FIG. 10b shows the parameter r in the present inventioniA plot of the absolute value of the estimated error for the second state in the first extended state observer for the second follower at 1,2,3,4 changes;
FIG. 11a shows the parameter r in the present inventioniA plot of the absolute value of the first state estimate error in the second extended state observer of the second follower at 1,2,3,4 changes;
FIG. 11b shows the parameter r in the present inventioniA second plot of absolute state estimation error in a second extended state observer of a second follower at 1,2,3,4 changes;
FIG. 12a controller parameter q in the present inventioni,2,1、qi,2,2The first follower's first comparison graph contains the absolute value of the error when i changes to 1,2,3, 4;
FIG. 12b illustrates the controller parameter q of the present inventioni,2,1、qi,2,2The second of the first followers contains a comparison graph of absolute error values when i is changed to 1,2,3, 4.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings and simulations.
The multi-double-link mechanical arm based on the output position comprises a controller, N double-link mechanical arms containing unknown dynamics are regarded as followers, a networked system formed by connecting the N double-link mechanical arms with M leaders through a unidirectional topological graph is used as a controlled object, the leaders form a static convex hull, the controller is designed by utilizing the positions of respective joint 1 and joint 2 outputs of the followers, the output of the followers is converged in the static convex hull, the controller structure of the ith (i is more than or equal to 1 and less than or equal to N) follower comprises a1 st sub-controller and a2 nd sub-controller, and the input ends of the 1 st and the 2 nd sub-controllers and the directed graph are connected
The output ends of the 2 sub-controllers are all connected with the input end of the ith follower,
the 1 st sub-controller comprises a first extended state observer unit, an error surface s
i,1,1The device comprises an arithmetic unit, a first nonlinear arithmetic unit, a first comparator unit, a first tracking differentiator unit and a second nonlinear arithmetic unit; two input ends of the first extended state observer unit are respectively the output y of the ith follower joint 1
i,1And the output u of the second non-linear operation unit
i,1(ii) a Error surface s
i,1,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,1,dAdjacent communication a
ijAnd the output of the first extended state observer unit
The input ends of the first nonlinear operation units are directed graphs respectively
Middle observer state
Output y of jth follower (j is more than or equal to 1 and less than or equal to N)
j,1Adjacent communication a
ijOutput y of i-th follower joint 1
i,1Output of the first extended state observer unit
Sum error surface s
i,1,1An output of the arithmetic unit; the first comparator unit being an error surface s
i,1,2The input end of which is the output of the first extended state observer unit
And the output x of the first non-linear operation unit
i,1,2,d(ii) a The input end of the first tracking differentiator unit is the output x of the first nonlinear operation unit
i,1,2,d(ii) a The input end of the second nonlinear operation unit is a directed graph
Intermediate adjacent communication a
ijOutput y of i-th follower joint 1
i,1Error surface s
i,1,1Output of arithmetic unit, output v of first tracking differentiator unit
i,1,2An output s of the first comparator unit
i,1,2And the output of the first extended state observer unit
The 2 nd sub-controller comprises a second extended state observer unit, an error surface s
i,2,1The device comprises an arithmetic unit, a third nonlinear arithmetic unit, a second comparator unit, a second tracking differentiator unit and a fourth nonlinear arithmetic unit; two input ends of the second extended state observer unit are respectively the output y of the ith follower joint 2
i,2And the output u of the fourth non-linear operation unit
i,2(ii) a Error surface s
i,2,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,2,dAdjacent communication a
ijAnd the output of the second extended state observer unit
The input ends of the third nonlinear operation units are directed graphs respectively
Middle observer state
Output y of jth follower joint 2
j,2Adjacent communication a
ijOutput y of the i-th follower joint 2
i,2The output of the second extended state observer unit
Sum error surface s
i,2,1An output of the arithmetic unit; the second comparator unit being an error surface s
i,2,2The input end of which is the output of the second extended state observer unit
And the output x of the third non-linear operation unit
i,2,2,d(ii) a The input end of the second tracking differentiator unit is the output x of the third nonlinear operation unit
i,2,2,d(ii) a The input end of the fourth nonlinear operation unit is a directed graph
Intermediate adjacent communication a
ijOutput y of the i-th follower joint 2
i,2Error surface s
i,2,1Output of arithmetic unit, output v of second tracking differentiator unit
i,2,2The output s of the second comparator unit
i,2,2And a second extended state observer unitOutput of
The multi-double-link mechanical arm based on the output position comprises a controller, a multi-agent network formed by connecting follower groups consisting of N double-link mechanical arms containing unknown dynamic states and M leaders through a one-way topological graph is considered, and each follower is communicated with at least one leader; the information communication between them can be made by directed graphs
Is shown in which
Is a set of nodes, n
iIndicating a double link arm i, n
jA double link robot arm j is shown,
is a collection of edges that are to be considered,
indicating that agent j can directly obtain the information of agent i;
is a contiguous matrix, a
ijThe definition is as follows:
the contiguous set of nodes i is defined as
Directed graph
Laplacian matrix of
The definition is as follows:
laplace matrix
Wherein D ═ diag [ D [ ]
1,…,d
N]As directed graphs
The degree matrix of (c) is,
the follower has at least one adjacent node, and the leader has no adjacent node, the Laplace matrix can
And (3) decomposition:
many double-link arms based on output position contain the controller, and the mathematical model of the ith double-link arm in the follower is:
in the formula, q
i,1Is the angular position, q, of the ith double link robot arm joint 1
i,2Is the angular position, q, of the ith double link robot arm joint 2
i=[q
i,1 q
i,2]
TIs the connection angular position;
respectively, connecting angular velocity and acceleration, tau
i=[τ
i,1 τ
i,2]
TControl input is performed, and system parameters are selected as follows:
in the formula,
a
3=m
2l
1l
c2cos(δ),a
4=m
2l
1l
c2sin(δ),I
1viscous coefficient of friction for turning the joint 1 to
2Viscous coefficient of friction to which the joint 2 turns,/
1Is the length of the
rod 1, m
1Is the mass of the
rod 1, m
2Is the mass of the
rod 2, delta is the angle between the
initial joint 2 and the longitudinal axis, l
c1Is the distance of the joint 1 from the center of mass of the
bar 1, l
c2Distance of joint 2 to the center of mass of
rod 2, further, let [ x [ ]
i,1,1 x
i,2,1]
T=[q
i,1 q
i,2]
T,
[u
i,1 u
i,2]
T=[τ
i,1 τ
i,2]
TThe equation of state for the ith follower can be found:
in the formula, y
i,1、y
i,2Output of the i-th follower joint 1 and joint 2, W, respectively
i,11=a
1+2a
2cos(y
i,1)+2a
4sin(y
i,2),
And
the inner unknown dynamic part of the ith follower dual link robotic arm.
The design method of the multi-double-connecting-rod mechanical arm containing controller based on the output position comprises the following steps:
A. design of the 1 st sub-controller:
a1, design of first extended state observer unit:
the two inputs of the first extended state observer unit are divided into the output y of the ith follower joint 1
i,1And the output u of the second non-linear operation unit
i,1The output signal of the first extended state observer unit is obtained by calculation of the following formula
And
and
estimated signals of the first, second and third order state variables of the i-th follower's joint 1, respectively:
wherein T is time, Ti,1,1、Ti,1,2And Ti,1,3All are the gains of the extended state observer, riE (1, ∞) is the parameter to be designed, function gi(τ) is fal composed of linear and fractional power functionsi(τ,δi,j1) function, j ═ 1,2,3, expressed as:
wherein, deltai,jE (0,1) is a tunable parameter, sign (. cndot.) is a sign functionThe expression is:
a2, error surface si,1,1Designing an arithmetic unit:
error surface s
i,1,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,1,dAdjacent communication a
ijAnd the output of the first extended state observer unit
Obtaining an error surface s by calculation of the following formula
i,1,1Output s of arithmetic unit
i,1,1:
A3, design of first nonlinear arithmetic unit:
the input ends of the first nonlinear operation units are directed graphs respectively
Middle observer state
Output y of jth follower joint 1
j,1Adjacent communication a
ijOutput y of i-th follower joint 1
i,1Output of the first extended state observer unit
Sum error surface s
i,1,1The output of the arithmetic unit is calculated byThe calculation of the formula yields the virtual control law x
i,1,2,d:
In the formula, T
i,1,1And T
j,1,1In order to extend the gain of the state observer,
r
i∈(1,∞)、r
j∈(1,∞)、q
i,1,1parameters to be designed are more than 0;
a4, design of first comparator unit:
the input of the first comparator unit is the output of the first extended state observer unit
And the output x of the first non-linear operation unit
i,1,2,dAn error surface s is obtained by calculation of the following formula
i,1,2:
A5, design of the first tracking differentiator unit:
the input end of the first tracking differentiator is the output x of the first nonlinear operation uniti,1,2,dThe output v of the tracking differentiator is obtained by calculation of the following formulai,1,2:
In the formula, vi,1,2Is the first non-linear operation unit output xi,1,2,dIs estimated, α ∈ (0,1) is the filter factor, λ > 0 is the speed factor;
design of a6 and second nonlinear operation unit:
the input end of the second nonlinear operation unit is a directed graph
Intermediate adjacent communication a
ijOutput y of i-th follower joint 1
i,1Error surface s
i,1,1Output of arithmetic unit, output v of first tracking differentiator unit
i,1,2An output s of the first comparator unit
i,1,2And the output of the first extended state observer unit
The control law u is obtained by calculation of the following formula
i,1:
In the formula, T
i,1,2In order to extend the gain of the state observer,
r
i∈(1,∞)、q
i,2,1parameters to be designed are more than 0;
B. design of the 2 nd sub-controller:
b1, design of second extended state observer unit:
the two inputs of the second extended state observer unit are divided into the output y of the ith follower joint 1
i,2And the output u of the fourth non-linear operation unit
i,2The output signal of the second extended state observer unit is obtained by calculation of the following formula
And
and
estimated signals of the first, second and third order state variables of the i-th follower's joint 2, respectively:
wherein T is time, Ti,2,1、Ti,2,2And Ti,2,3All are the gains of the extended state observer, riE (1, infinity) is a parameter to be designed;
b2, error surface si,2,1Designing an arithmetic unit:
error surface s
i,2,1The input ends of the arithmetic units are directed graphs respectively
Middle observer state
Output y of kth leader double link robot arm
k,2,dAdjacent communication a
ijAnd the output of the second extended state observer unit
Obtaining an error surface s by calculation of the following formula
i,2,1Output s of arithmetic unit
i,2,1:
B3, design of third nonlinear operation unit:
the input ends of the third nonlinear operation units are directed graphs respectively
Middle observer state
Output y of jth follower joint 2
j,2Adjacent communication a
ijOutput y of the i-th follower joint 2
i,2The output of the second extended state observer unit
Sum error surface s
i,2,1The output of the operation unit is calculated by the following formula to obtain the virtual control law x
i,2,2,d:
In the formula, T
i,2,1And T
j,2,1Are the gains of the extended state observer,
r
i∈(1,∞)、r
j∈(1,∞)、q
i,2,1parameters to be designed are more than 0;
b4, design of second comparator cell:
the input of the second comparator unit is the output of the second extended state observer unit
And the output x of the third non-linear operation unit
i,2,2,dAn error surface s is obtained by calculation of the following formula
i,2,2:
B5, design of second tracking differentiator unit:
the input end of the second tracking differentiator is the output x of the third nonlinear operation uniti,2,2,dThe output v of the tracking differentiator is obtained by calculation of the following formulai,2,2:
In the formula, vi,2,2Is the third non-linear operation unit output xi,2,2,dIs estimated, α ∈ (0,1) is the filter factor, λ > 0 is the speed factor;
b6, design of fourth nonlinear operation unit:
the input end of the fourth nonlinear operation unit is directed graph G adjacent communication a
ijOutput y of the i-th follower joint 2
i,2Error surface s
i,2,1Output of arithmetic unit, output v of second tracking differentiator unit
i,2,2The output s of the second comparator unit
i,2,2And the output of the second extended state observer unit
The control law u is obtained by calculation of the following formula
i,2:
In the formula, T
i,2,2In order to extend the gain of the state observer,
r
i∈(1,∞)、q
i,2,2more than 0 is a parameter to be designed;
to this end, the control input u of the ith follower is obtainedi,1And ui,2。
Examples
The invention selects the state equation of the ith double-link mechanical arm system in the follower:
in the formula, y
i,1、y
i,2Connecting angular position, u, output for i-th follower joint 1 and joint 2
i,1、u
i,2Control law, W, input for i-
th follower joints 1 and 2
i,11=a
1+2a
2cos(y
i,1)+2a
4sin(y
i,2),
I1 is the viscous molar movement to which the
joint 1 is transferredCoefficient of friction of 0.12, I
2The viscous friction coefficient of the joint 2 is 0.25 m
1The mass of the
rod 1 is 1kg, m
2The mass of the
rod 2 is 2kg, I
c2The distance from the joint 2 to the mass center of the
rod 2 is 0.6m,
a
4=m
2l
1l
c2sin(δ),l
1is a
rod 1 having a length of 1m, I
c1The distance from the joint 1 to the barycenter of the
rod 1 is 0.5m, and delta is the included angle between the
initial joint 2 and the longitudinal axis pi/6 rad. The initial values of the four follower system states are respectively:
the outputs of the three leaders are: y is
5,d=[0.8,2]
T、y
6,d=[1,4]
T、y
7,d=[1.6,3]
T。
Considering a multi-agent network consisting of 4 followers and 3 leaders, with each follower having communication with at least one leader, the communication topology of which is shown in fig. 3, wherein 1,2,3,4 are the numbers of four followers, and 5, 6, 7 are the numbers of three leaders, a laplacian matrix is further obtained
From the laplace matrix, we know that:
in the embodiment, the purpose of system control is to effectively solve the problems of unknown dynamic and partial system state failure existing in a multi-double-link mechanical arm system when the output positions of four followers are controlled under a one-way topological diagram by the aid of the controller designed hereinCan obtain and virtually control the difficult problem of complex derivation in the inversion design, so that the output y of the followeri=[yi,1,yi,2]TCan converge within the static convex hull formed by the leader outputs.
For this system, the controller comprising the ith dual link robot arm of the follower can be designed according to fig. 1 as follows:
the following parameters were selected depending on the repeated simulation experiments to determine the controller parameters: r isi=3、Ti,1,1=10、Ti,1,2=20、Ti,1,3=50、Ti,2,1=10、Ti,2,2=20、Ti,2,3=50、qi,1,1=1.1、qi,1,2=15、qi,2,1=1.1、qi,2,2In fig. 3 to 7, the initial value of each state in the extended state observer is the same as the initial value of the observed system, and the initial values of the remaining extended state observer and tracking differentiator are all 0.
As shown in FIGS. 3-7, the output positions of the follower groups consisting of four two-link arms are designed to include the control law ui,j(j-1, 2) into the convex hull formed by the leader, four followers' inclusion ei,j(j 1,2) the error converges to a very small neighborhood of zero. As can be seen in FIGS. 8a-9 b: in the case where the extended state observers are both 0, the two extended state observers of the second follower can effectively estimate the state in the system. As can be seen in FIGS. 10a-11 b: in the extended state, the observer is initialized to 0 and divides the parameter riExcept for the case where all parameters are unchanged, the observer follows the expansion of the parameter r in the stateiAs the value becomes larger, the absolute value of the estimation error of the two extended state observers in the second follower becomes smaller. As can be seen from fig. 12a and 12 b: in-division controller gain qi,2,1、qi,2,2With all other parameters being constant, with the controller gain qi,2,1、qi,2,2When it becomes largeThe absolute value of the first follower including the error becomes smaller. The simulation shows that the controller of the multi-double-connecting-rod mechanical arm based on the active disturbance rejection and inversion technology only depends on the output position, and the problems of unknown dynamics in a system, simplified complex derivation operation and the like can be effectively solved.