CN114474051B - A Personalized Gain Teleoperation Control Method Based on Operator Physiological Signals - Google Patents

Abstract

The present invention is a personalized gain teleoperation control method based on the operator's physiological signal, which belongs to the technical field of human-machine hybrid intelligence; firstly, a dynamic model representing the motion characteristics of the teleoperation robot is established, and the operator is represented by collecting electromyographic signals Part of the operating intention, combined with the design of the fixed-time uncertain observer and the fixed-time controller, enables the slave-end robot to track to the desired trajectory of the master-end within a fixed time under the condition of combining part of the operator's operating intention. The invention designs the observer and the controller by defining the tracking error including the self-time delay, improves the transparency of the system, and reduces the adverse influence of the transmission time delay on the tracking effect. By designing a personalized gain control method combined with the operator's physiological signal, the electromyographic signal is used as a representation to reflect the operator's operating intention, and the control gain in the controller is dynamically adjusted to improve the effect of teleoperation trajectory tracking.

Description

A personalized gain teleoperation control method based on operator's physiological signals

Technical Field

The invention belongs to the technical field of human-machine hybrid intelligence, and in particular relates to a personalized gain teleoperation control method based on operator physiological signals.

Background Art

Teleoperation is a technology that can transmit the operation information of the local operator to the remote robot for execution. Since it expands the limitations of local space and distance, teleoperation technology has also received great attention and application in recent years, such as the da Vinci surgical robot, space robotic arms, and underwater robotic arms. However, due to the inherent reasons of the distance span, the teleoperation system also brings the problem of operation delay. The existing teleoperation control methods are difficult to achieve fixed-time observation of uncertain information and fixed-time convergence of the system in the presence of delay, which brings challenges to some teleoperation tasks that need to meet speed and accuracy. In addition, the existing control methods of teleoperation systems are generally designed with fixed gains, without fully considering the operator's operating habits and intentions.

Generally, a higher gain in the controller will result in a larger control input and better transient performance, but will generally result in problems such as control input jitter. A lower gain in the controller will result in a smaller control input, but the transient performance is generally poor. When different operators operate or perform different tasks, in order to obtain good control performance, the existing fixed-gain controllers often need to be adjusted multiple times to achieve a better control effect. The process is cumbersome and it is difficult to achieve personalized and efficient control effects. As a human physiological signal that is easy to collect, electromyographic signals can be generated about 30 to 150ms before the limbs move, and can to a certain extent reflect the operator's operating intentions. Applying electromyographic signals to the personalized gain adjustment of the controller can to a certain extent combine the advantages of the high-gain and low-gain parts of the controller, thereby improving the control effect.

Summary of the invention

Technical issues to be solved:

In order to avoid the shortcomings of the prior art, the present invention proposes a personalized gain teleoperation control method based on the operator's physiological signals. First, a dynamic model representing the motion characteristics of the teleoperated robot is established, and the operator's partial operation intention is characterized by electromyographic signal acquisition. Combined with the fixed-time uncertain observer and fixed-time controller design, the slave robot can track the master end's expected trajectory within a fixed time in combination with part of the operator's operation intention. The present invention designs a personalized gain control method combined with the operator's physiological signals, uses electromyographic signals as a representation, reflects the operator's operation intention, and dynamically adjusts the control gain in the controller to achieve the effect of improving teleoperation trajectory tracking.

The technical solution of the present invention is: a personalized gain remote operation control method based on the operator's physiological signal, characterized by the following specific steps:

Step 1: Establish the dynamic model of the teleoperated robot, and obtain the dynamic model of the slave robot as follows:

分别为标称的惯性矩阵、科里奥利矩阵以及重力向量；不确定项ΔM_i，ΔC_i，Δg_i分别为惯性矩阵、科里奥利矩阵以及重力向量的不确定部分；

代表扰动向量，下标i＝m，s分别代表主端和从端；τ_s为从端机器人控制输入，τ_e为环境对从端机器人施加的力；Among them, the nominal model

are the nominal inertia matrix, Coriolis matrix and gravity vector respectively; the uncertainties ΔM _i , ΔC _i , Δg _i are the uncertain parts of the inertia matrix, Coriolis matrix and gravity vector respectively;

represents the disturbance vector, the subscript i=m, s represents the master and slave respectively; _τs is the control input of the slave robot, _τe is the force exerted by the environment on the slave robot;

Step 2: Design a fixed-time uncertain observer;

其中

为已知常值时延，δ(t)为存在上界的时变时延；Assume that the time required for the master robot to transmit information to the slave robot is T _m (t),

in

is a known constant delay, δ(t) is a time-varying delay with an upper bound;

The tracking error of the slave robot including the self-delay is defined as follows:

为引入的自时延部分；in,

is the self-delay part introduced;

The tracking error dynamics of the remote robot from the slave end are obtained as follows:

为包含不确定项的集总扰动；in,

is the lumped disturbance containing uncertain terms;

Combining the tracking error dynamics of the slave teleoperated robot (3) and the slave robot dynamics model (1), the following fixed-time uncertain observer is designed:

e_v，

的观测值，λ₁，λ₂，λ₃，λ₄，λ₅，λ₆，p为观测器参数，通过设计该固定时间不确定观测器，包含时延以及不确定项的集总干扰能够被在固定时间内进行观测，且观测误差为零，因此可以得到

Among them, z ₀ , z ₁ , z ₂ , z ₃ are e _p ,

e _v ,

, λ ₁ , λ ₂ , λ ₃ , λ ₄ , λ ₅ , λ ₆ , and p are observer parameters. By designing the fixed-time uncertain observer, the aggregate interference including delay and uncertainty can be observed within a fixed time, and the observation error is zero, so we can get

Step 3: Personalized gain controller control strategy based on physiological signals;

Combined with the observed values, the tracking error dynamics (3) of the slave teleoperated robot in step 2 is rewritten as follows:

in,

The non-singular terminal sliding surface is defined as:

Wherein, α ₁ , α ₂ , μ ₁ , μ ₂ are parameters in sliding mode design, sgn(*) ^a =sign(*)|*| ^a ;

Combining the tracking error model (2) and the non-singular terminal sliding surface (7), the following control law is obtained:

μ₃，μ₄为控制律中设计的参数，ρ为操作者操作时计算生成的肌电信号增益；in,

μ ₃ , μ ₄ are the parameters designed in the control law, and ρ is the myoelectric signal gain calculated and generated when the operator operates;

Step 4: Verify system stability through Lyapunov function.

A further technical solution of the present invention is: in the step 1, the Cartesian space dynamics model of the master end and the slave end teleoperated robot is considered as follows:

为机器人惯性矩阵，

为机器人科里奥利矩阵，

代表重力向量，

代表扰动向量，下标i＝m，s分别代表主端及从端；τ_m为主端机器人的控制输入，τ_h为操作员对主端机器人施加的力，τ_s为从端机器人控制输入，τ_e为环境对从端机器人施加的力；in,

is the robot inertia matrix,

For the robot Coriolis matrix,

represents the gravity vector,

represents the disturbance vector, the subscript i=m, s represents the master and slave respectively; _τm is the control input of the master robot, _τh is the force applied by the operator to the master robot, _τs is the control input of the slave robot, and _τe is the force applied by the environment to the slave robot;

The nominal model and uncertainty terms are introduced to represent the real robot model, which is expressed as follows:

分别为标称的惯性矩阵、科里奥利矩阵以及重力向量。不确定项ΔM_i，ΔC_i，Δg_i分别为惯性矩阵、科里奥利矩阵以及重力向量的不确定部分。Among them, the nominal model

are the nominal inertia matrix, Coriolis matrix and gravity vector respectively. The uncertainties ΔM _i , ΔC _i , Δg _i are the uncertain parts of the inertia matrix, Coriolis matrix and gravity vector respectively.

By substituting formula (10) into formula (9), the dynamic model of the slave robot is obtained as follows:

A further technical solution of the present invention is: in the step 4, the Lyapunov function is selected as follows:

V＝s ^T s (12)

By taking the derivative of the Lyapunov function of equation (12) and substituting it into the control law (8) and the sliding surface (7), we can obtain:

证明了公式(8)得到的包含自时延的控制系统的固定时间稳定特性；in

The fixed-time stability characteristic of the control system containing self-delay obtained by formula (8) is proved;

The stability of the tracking error between the slave and the master without self-delay is further proved. The tracking error between the slave and the master without self-delay is as follows:

x _s ^* (t)＝x _s (t)-x _m (t) (14)

The tracking error (2) including the self-delay can be rewritten as follows:

From (13), we can get that the error system with self-delay can be stable in fixed time and _ep = 0. Combined with equation (15), it is proved that _xs ^* (t) is bounded, that is, the slave robot can track the master reference trajectory in a fixed time. The error system without self-delay is stable and the tracking error is bounded as follows:

A further technical solution of the present invention is: the control strategy is simulated and verified on a three-joint robot arm, the initial position of the slave end is [0.6, 0.3, 0.2] rad, and the reference trajectory of the three joints of the master end is given as:

Beneficial Effects

The beneficial effects of the present invention are as follows: the present invention proposes a personalized gain teleoperation control method based on the operator's physiological signal, establishes a teleoperation robot dynamics model, designs a fixed-time uncertain observer and a personalized gain controller based on physiological signals. Compared with existing research, the method of the present invention has the following advantages:

(1) By defining the tracking error including self-delay to design the observer and controller, compared with the traditional error definition method that does not include self-delay, the transparency of the system is improved and the adverse impact of transmission delay on the tracking effect is reduced.

(2) By designing a fixed-time uncertain observer and a fixed-time control strategy, the fixed-time convergence of the observation value of the uncertain term and the teleoperation system is achieved. The upper bound of the convergence time of the system is only related to the controller parameters. Compared with the finite-time control method and the asymptotically stable control method, the tracking speed of the system is improved.

(3) By incorporating the physiological signal of electromyography into the controller design process, a personalized gain control strategy based on physiological signals is realized, and the advantages of high gain parameters and low gain parameters in the controller are combined. It can be seen from the tracking error in Figure 6 that the personalized gain control based on physiological signals can achieve better tracking effect, faster convergence time and no jitter phenomenon compared with the low gain parameter controller in Figure 2 and the high gain parameter controller in Figure 4. In addition, the introduction of electromyography signals in the controller parameters reflects the operation intentions of some operators to a certain extent, improving the intelligence of the teleoperation system.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a diagram showing the tracking effect of joint position trajectory under a low-gain parameter controller of the present invention;

FIG2 is a diagram of the joint position tracking error under the low gain parameter controller of the present invention;

FIG3 is a diagram showing the tracking effect of the joint position trajectory under the high gain parameter controller of the present invention;

FIG4 is a diagram of the joint position tracking error of the high gain parameter controller of the present invention;

FIG5 is a diagram showing the joint position trajectory tracking effect of the personalized gain parameter controller of the present invention;

FIG6 is a diagram of the joint position tracking error under the personalized gain parameter controller of the present invention;

FIG. 7 shows the personalized gain of myoelectric acquisition according to the present invention.

DETAILED DESCRIPTION

The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but should not be construed as limiting the present invention.

Step 1: Dynamic Model of Teleoperated Robot

Consider the Cartesian space dynamics model of the master and slave teleoperated robots as follows:

为机器人惯性矩阵，

为机器人科里奥利矩阵，

代表重力向量，

代表扰动向量，下标i＝m，s分别代表主端及从端。τ_m为主端机器人的控制输入，τ_h为操作员对主端机器人施加的力，τ_s为从端机器人控制输入，τ_e为环境对从端机器人施加的力。in,

is the robot inertia matrix,

For the robot Coriolis matrix,

represents the gravity vector,

represents the disturbance vector, the subscript i=m, and s represents the master and slave respectively. _τm is the control input of the master robot, _τh is the force applied by the operator to the master robot, _τs is the control input of the slave robot, and _τe is the force applied by the environment to the slave robot.

Since it is difficult to obtain an accurate dynamic model of the robot in actual situations, a nominal model and uncertainty terms are introduced to represent the real robot model. The corresponding expressions are as follows:

分别为标称的惯性矩阵、科里奥利矩阵以及重力向量。不确定项ΔM_i，ΔC_i，Δg_i分别为惯性矩阵、科里奥利矩阵以及重力向量的不确定部分。Among them, the nominal model

are the nominal inertia matrix, Coriolis matrix and gravity vector respectively. The uncertainties ΔM _i , ΔC _i , Δg _i are the uncertain parts of the inertia matrix, Coriolis matrix and gravity vector respectively.

By substituting equation (2) into equation (1), the dynamics of the slave robot can be obtained as follows:

Step 2: Fixed-time uncertain observer design

其中

为已知常值时延，δ(t)为存在上界的时变时延。Assume that the time required for the master robot to transmit information to the slave robot is T _m (t), and assume that the unknown delay T _m (t) consists of two parts:

in

is a known constant delay, and δ(t) is a time-varying delay with an upper bound.

The tracking error of the slave robot including the self-delay is defined as follows:

为引入的自时延部分。in,

This is the introduced self-delay part.

The corresponding tracking error dynamics can be obtained as follows:

为包含不确定项的集总扰动。in

is the lumped disturbance including uncertain terms.

Considering the tracking error dynamics (5) and the dynamics model (3) of the remote robot, the following fixed-time uncertain observer can be designed:

的观测值，λ₁，λ₂，λ₃，λ₄，λ₅，λ₆，p为观测器参数，通过设计该固定时间不确定观测器，包含时延以及不确定项的集总干扰能够被在固定时间内进行观测，且观测误差为零，因此可以得到

Among them, z ₀ , z ₁ , z ₂ , z ₃ are e _p ,

, λ ₁ , λ ₂ , λ ₃ , λ ₄ , λ ₅ , λ ₆ , and p are observer parameters. By designing the fixed-time uncertain observer, the aggregate interference including delay and uncertainty can be observed within a fixed time, and the observation error is zero, so we can get

Step 3: Personalized gain controller control strategy based on physiological signals

To facilitate subsequent design and analysis, the tracking error dynamics (5) in step 2 is rewritten as follows in combination with the observed values:

in,

The non-singular terminal sliding surface is defined as:

Wherein, α ₁ , α ₂ , μ ₁ , μ ₂ are parameters in sliding mode design, and sgn(*) ^a =sign(*)|*| ^a .

Combining the tracking error model (4) and the designed non-singular terminal sliding surface (9), the following control law can be obtained:

in,

Wherein μ ₃ and μ ₄ are the parameters designed in the control law, and ρ is the myoelectric signal gain calculated and generated when the operator operates.

Step 4: Determine system stability

The Lyapunov function is selected as follows:

V＝s ^T s (11)

By taking the derivative of the Lyapunov function of equation (11) and substituting it into the control law (10) and the sliding surface (9), we can obtain:

证明了公式(10)得到的包含自时延的控制系统的固定时间稳定特性。in

The fixed-time stability characteristic of the control system containing self-delay obtained by formula (10) is proved.

It will be further proved that the tracking error stability between the slave and the master without self-delay is as follows:

x _s ^* (t)＝x _s (t)-x _m (t) (13)

The tracking error (4) including the self-delay can be rewritten as follows

From (12), we can get that the error system with self-delay can be stable in fixed time and e _p = 0. Combined with equation (14), it can be proved that x _s ^* (t) is bounded, that is, the slave robot can track the master reference trajectory in a fixed time. The error system without self-delay is stable and the tracking error is bounded as follows:

Step 5: Simulation Verification

The control strategy was simulated and verified on a three-joint robot arm. The initial position of the slave end was [0.6, 0.3, 0.2] rad, and the reference trajectory of the three joints of the master end was given as:

Simulation Verification Results

As shown in Figure 1, the joint position trajectory tracking curve μ ₃ =μ ₄ =0.1, ρ = 0 when the gain is small and the electromyographic physiological signal is not considered, and the position trajectory tracking diagram is shown in Figure 2. It can be seen that under the condition of low controller gain, the tracking error tends to be stable in about 4 seconds. As shown in Figure 3, the joint position trajectory tracking curve μ ₃ =μ ₄ =5, ρ = 0 when the gain is large and the electromyographic physiological signal is not considered, and Figure 4 is the corresponding master-slave tracking error trajectory. It can be seen that under the condition of high controller gain, the transient performance of the system is improved, and it can be stable in about 1 second, but it will bring some jitter. Figure 5 is the joint position trajectory tracking curve μ ₃ =μ ₄ =0.1 when a lower gain is used but the electromyographic physiological signal is considered, and Figure 6 is the corresponding master-slave tracking error. It can be seen that the system can be stable in about 2 seconds, and there is no jitter in the steady state, and it combines the advantages of low gain controller parameters and high gain controller parameters. The change gain ρ is shown in FIG7 . It can be seen that when the error is large, the gain of the electromyographic signal is large, and when the tracking error is small, the gain of the electromyographic signal is correspondingly small.

Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, substitute and modify the above embodiments within the scope of the present invention without departing from the principles and purpose of the present invention.

Claims (4)

1. A personalized gain teleoperation control method based on operator physiological signals, characterized by the following specific steps: Step 1: Establish the dynamic model of the teleoperated robot, and obtain the dynamic model of the slave robot as follows:

Among them, the nominal model

are the nominal inertia matrix, Coriolis matrix and gravity vector respectively; the uncertainties ΔM _i , ΔC _i , Δg _i are the uncertain parts of the inertia matrix, Coriolis matrix and gravity vector respectively;

represents the disturbance vector, the subscripts i = m, s represent the master and slave respectively; _τs is the control input of the slave robot, and _τe is the force exerted by the environment on the slave robot; Step 2: Design a fixed-time uncertain observer; Assume that the time required for the master robot to transmit information to the slave robot is T _m (t),

in

is a known constant delay, δ(t) is a time-varying delay with an upper bound; The tracking error of the slave robot including the self-delay is defined as follows:

in,

is the self-delay part introduced; The tracking error dynamics of the remote robot from the slave end are obtained as follows:

in,

is the lumped disturbance containing uncertain terms; Combining the tracking error dynamics of the slave teleoperated robot (3) and the slave robot dynamics model (1), the following fixed-time uncertain observer is designed:

Among them, z ₀ , z ₁ , z ₂ , z ₃ are e _p ,

e _v ,

, λ ₁ , λ ₂ , λ ₃ , λ ₄ , λ ₅ , λ ₆ , and p are observer parameters. By designing the fixed-time uncertain observer, the aggregate interference including delay and uncertainty can be observed within a fixed time, and the observation error is zero, so we can get

Step 3: Personalized gain controller control strategy based on physiological signals; Combined with the observed values, the tracking error dynamics (3) of the slave teleoperated robot in step 2 is rewritten as follows:

in,

The non-singular terminal sliding surface is defined as:

Wherein, α ₁ , α ₂ , μ ₁ , μ ₂ are parameters in sliding mode design, sgn(*) ^a =sign(*)|*| ^a ; Combining the tracking error model (2) and the non-singular terminal sliding surface (7), the following control law is obtained:

in,

μ ₃ , μ ₄ are the parameters designed in the control law, and ρ is the myoelectric signal gain calculated and generated when the operator operates; Step 4: Verify system stability through Lyapunov function. 2. According to claim 1, the personalized gain teleoperation control method based on the operator's physiological signal is characterized in that: in the step 1, the Cartesian space dynamics model of the master-end and slave-end teleoperation robots is considered as follows:

in,

is the robot inertia matrix,

For the robot Coriolis matrix,

represents the gravity vector,

represents the disturbance vector, the subscript i=m, s represents the master and slave respectively; _τm is the control input of the master robot, _τh is the force applied by the operator to the master robot, _τs is the control input of the slave robot, and _τe is the force applied by the environment to the slave robot; The nominal model and uncertainty terms are introduced to represent the real robot model, which is expressed as follows:

Among them, the nominal model

are the nominal inertia matrix, Coriolis matrix and gravity vector respectively; the uncertainties ΔM _i , ΔC _i , Δg _i are the uncertain parts of the inertia matrix, Coriolis matrix and gravity vector respectively; By substituting formula (10) into formula (9), the dynamic model of the slave robot is obtained as follows:

3. The personalized gain teleoperation control method based on operator physiological signals according to claim 1 is characterized in that: in the step 4, the Lyapunov function is selected as follows: V＝s ^T s (12) By taking the derivative of the Lyapunov function of equation (12) and substituting it into the control law (8) and the sliding surface (7), we can obtain:

in

The fixed-time stability characteristic of the control system containing self-delay obtained by formula (8) is proved; The stability of the tracking error between the slave and the master without self-delay is further proved. The tracking error between the slave and the master without self-delay is as follows: x _s ^* (t)＝x _s (t)-x _m (t) (14) The tracking error (2) including the self-delay can be rewritten as follows:

From (13), we can get that the error system with self-delay can be stable in fixed time and _ep = 0. Combined with equation (15), it is proved that _xs ^* (t) is bounded, that is, the slave robot can track the master reference trajectory in a fixed time. The error system without self-delay is stable and the tracking error is bounded as follows:

4. According to claim 3, the personalized gain teleoperation control method based on the operator's physiological signal is characterized in that: the control strategy is simulated and verified on a three-joint robot arm, the initial position of the slave end is [0.6, 0.3, 0.2] rad, and the three-joint reference trajectory of the master end is given as:

Images