CN106527127A - Time delay teleoperation robot adaptive control method based on environmental impedance model - Google Patents

Abstract

The invention discloses an adaptive control method for a time-delay teleoperation robot based on an environmental impedance model, including a master-end robot, a master-end environmental geometric model, an environmental impedance model, a time-delay communication link, a slave-end environmental geometric model, and position correction calculation links, slave robots, and real environments. The environmental impedance model outputs the reference force signal, and the force signal generated by the contact between the slave robot and the real environment is input into the position correction calculation link for comparison operation, and the adjustable gain is output, and the slave environmental geometric model is adjusted according to the adjustable gain to obtain the actual position command Control the movement of the slave robot, continuously correct the position information of the slave robot to make it approach the virtual environment position information, form the master end to guide the slave end, and overcome the delay effect of adaptive control teleoperation, so that the system can run stably.

Description

A time-delay teleoperated robot adaptive control method based on environmental impedance model

technical field

The invention relates to an adaptive control method for a time-delay teleoperation robot based on an environment impedance model, and belongs to the technical field of robot control.

Background technique

Due to the harsh conditions of the dangerous environment, it is extremely dangerous and costly to complete the work tasks through the operator. Using robots to replace people to complete the work tasks is the main way to realize remote work. However, the existence of communication delay between the operator's location and the remote environment reduces the stability of the teleoperation system, reduces the transparency of the system, and causes great difficulties for the operator to operate the robot remotely.

In most teleoperation systems, we pre-set the tasks to be performed by the robot, and most of the environmental objects to be interacted with are known, such as space teleoperation pins such as holes, and recycling satellite fragments, all of which are structured Interaction objects can be defined as structured environments. Most of the current teleoperation works in a structured environment, which is characterized by a definite communication delay, and the dynamic properties, size, and shape of the environmental objects are known, and the real-time position of the environmental objects is mainly obtained by sensor detection, which is prone to measurement errors . In this case, because the operator operates according to the geometric model with position error, it is easy to make wrong judgment and operation.

The invention patent with the Chinese patent number 201010265872.2 discloses an adaptive control method for a teleoperated robot based on a master-slave reference model, including: a master-side loop, a slave-side loop, and a communication delay link. The master-side loop is composed of the operator and the master robot. It is composed of the main edge environment model, and the slave edge circuit is composed of environment, slave robot, slave edge environment model, model parameter correction module and simulation delay module. The reference force signal is provided from the edge environment model, compared with the force signal input from the environment feedback to the model parameter correction module, and the error is used to output the adjustable gain pc(τ), pc(τ) is output from the model parameter correction module, through communication After the time-delay link, adjust the environment model of the main side, and the environment model of the main side will provide the feedback force signal for the operator. The side environment model and the side environment model are constantly approaching the real environment model, forming a teleoperation that overcomes the influence of time delay, and enables the system to obtain stable control. However, the patented method mainly corrects the main side environment model based on the side environment model. It is effective for operations with small delays. As the delay increases, its effect will gradually weaken. In the case of large delays, even if the master-side environment model and the slave-side environment model continue to approach, it will still cause dangerous operations.

Contents of the invention

Purpose of the invention: In the time-delay teleoperation environment, the present invention proposes an adaptive control method for a time-delay teleoperation robot based on an environmental impedance model to achieve stable operation of the system, aiming at the error problem in the geometric modeling of the environment.

Technical scheme: in order to achieve the above object, the technical scheme adopted in the present invention is:

An adaptive control method for a time-delayed teleoperated robot based on an environmental impedance model, including a main-end robot (2), a main-end environmental geometric model (3), an environmental impedance model (4), a time-delay communication link (5), and a slave End environment geometric model (6), position correction calculation link (7), slave end robot (8), real environment (9), environmental impedance model (4) is consistent with real environment (9), master end environment geometric model (3 ) and the slave-end environmental geometric model (6) have a positional deviation, let t be the time variable of the control system, and T be the delay amount of the communication delay link, including the following steps:

Step 1, the operator (1) sends out the main-end position signal x _m (t) through the main-end robot (2), and the main-end position signal x _m (t) is respectively input into the main-end environmental geometric model (3), environmental impedance Model (4) and time-delay communication link (5), the environmental impedance model (4) produces the impedance model output force signal f _m (t), and the described environmental impedance model (4) is:

Among them, f _m (t) represents the output force signal of the impedance model, t is the time variable of the control system, k is the elastic coefficient of the environment model, x _m (t) represents the position signal of the main end, n is the surface geometry coefficient of the environment model, b is the model coefficient of the environment model, Denotes the derivative of x _m (t);

The impedance model output force signal f _m (t) forms a delayed 1T impedance model output force signal f _m (tT) after the time delay communication link (5) delays 1T and enters the position correction calculation link (7);

Step 2, the master-side position signal x _m (t) forms the slave-side position signal x _m (tT) after a time-delay communication link (5) delay of 1T, and the slave-side position signal x _m (tT) enters the slave-end environment The geometric model (6) forms the slave-end control position signal x _c (tT), and the slave-end environmental geometric model (6) is:

x _c (t) = _Re (t+T) x _m (t);

Among them, x _c (t) represents the control position signal of the slave end, and R _e (t+T) represents the adjustable gain;

The slave-end control position signal x _c (tT) is input to the slave-end robot (8), so that the slave-end robot (8) moves and generates an environmental position signal x _e (tT), which acts on the real environment (9) to generate a real environment Output force signal f _e (tT), the real environment is:

f _e (t) represents the real environment output force signal, x _e (t) represents the environment position signal;

The real environment output force signal f _e (tT) enters the position correction calculation link (7), and the position correction calculation link (7) outputs the force signal f _m (tT) and The real environment output force signal f _e (tT) outputs the adjustable gain R _e (tT), and the adjustable gain R _e (tT) is input into the slave-end environmental geometric model (6), and the slave-end environmental geometric model (6) is based on The adjustable gain R _e (tT) modifies the slave-end environmental geometric model (6), and uses the revised slave-end environmental geometric model (6) for the control of the next cycle;

Step 3, the operator (1) sends a new master-end position signal x _m (t-2T) through the master-end robot (2), and the new master-end position signal x _m (t-2T) is respectively input into the geometric model of the master-end environment (3), the environmental impedance model (4) and the time delay communication link (5), the environmental impedance model (4) produces a new impedance model output force signal f _m (t-2T);

Step 4, the new master-side position signal x _m (t-2T) forms a new slave-side position signal x _m (t-3T) after a time-delay communication link (5) delay of 1T, and the new environmental impedance model outputs The force signal f _m (t-2T) passes through the delay communication link (5) to form a new environmental impedance model output force signal f _m (t-3T) after a delay of 1T and enters the position correction calculation link (7). The edge position signal x _m (t-3T) enters the revised slave-end environmental geometric model (6) to form a new slave-end control position signal x _c (t-3T), and the new slave-end control position signal x _c (t- 3T) is input to the slave robot (8), so that the slave robot (8) moves and generates a new environment position signal x _e (t-3T), which acts on the real environment (9) to generate a new real environment output force signal f _e ( t-3T), the new real environment output force signal f _e (t-3T) enters the position correction calculation link (7), and the position correction calculation link (7) is based on the time delay 1T after the delay communication link (5) After the new environmental impedance model output force signal f _m (t-3T) and the new real environment output force signal f _e (t-3T) output a new adjustable gain R _e (t-3T), the new adjustable gain R _e (t-3T) inputs the revised slave-end environmental geometric model (6), and the revised slave-end environmental geometric model (6) corrects the new slave-end environmental geometric model according to the new adjustable gain R _e (t-3T) (6), using the revised new slave-end environment geometric model (6) for the control of the next cycle;

Step 5 returns to step 1;

By repeating the above steps, the environmental geometric model (6) at the slave end gradually approaches the environmental geometric model (3) at the master end, and adaptive control is realized.

The generation method of the adjustable gain R _e (t) is:

The real environment output force signal f _e (t), the environmental impedance model output force signal f _m (t), and the control rate of the position correction calculation link (7) is β represents the gain ratio, and e(t) represents the generalized error. The specific steps are as follows:

The first step is to set the gain rate β=0.01-2;

In the second step, the generalized error e(t) is obtained according to the real environment output force signal f _e (t) and the environmental impedance model output force signal f _m (t) input to the position correction calculation link (7);

e(t) = f _e (t) - f _m (t);

The third step is to use the control rate formula to calculate _Re (t).

Beneficial effects: Compared with the prior art, the present invention has the following beneficial effects:

(1) The present invention effectively overcomes the instability problem caused by large time delay by approaching the geometric model of the environment at the slave end to the geometric model of the environment at the master end.

(2) In the remote operating system, the present invention corrects the geometric model of the real environment through the difference of the force feedback signals of the master and slave ends, promotes the consistency of the master and slave end environments, and improves the control accuracy of the remote operating system.

(3) The control method adopted in the present invention has the advantages of small calculation amount and low algorithm complexity.

Description of drawings

Fig. 1 is a control diagram of the teleoperation robot system based on the environmental impedance model of the present invention.

detailed description

Below in conjunction with accompanying drawing and specific embodiment, further illustrate the present invention, should be understood that these examples are only for illustrating the present invention and are not intended to limit the scope of the present invention, after having read the present invention, those skilled in the art will understand various aspects of the present invention All modifications of the valence form fall within the scope defined by the appended claims of the present application.

An adaptive control method for a time-delayed teleoperated robot based on an environmental impedance model, as shown in Figure 1, including: an operator 1, a main-end robot 2, a main-end environmental geometric model 3, an environmental impedance model 4, and a time-delay communication link 5. Slave-end environmental geometric model 6, position correction calculation link 7, slave-end robot 8, real environment 9, environmental impedance model 4 is consistent with real environment 9, master-end environmental geometric model 3 and slave-end environmental geometric model 6 have position deviation , let t be the time variable of the control system, and T be the delay amount of the communication delay link.

In a structured environment, it is known that there is a delay in the delay communication link 5, and T is the delay amount of the delay communication link 5. In the simulation environment, T=1s. In a structured environment, the impedance model of the real environment is set in advance , an accurate environmental impedance model 4 can be established, while the master-end environmental geometric model 3 is based on the initial sensor detection setting, there are errors, and the slave-end environmental geometric model 6 is a real environmental geometric model. Therefore, the master-end environmental geometric model 3 and the slave-end There is a deviation in the environmental geometric model 6. During the control process, by continuously adjusting the adjustable gain parameters, the slave environmental geometric model 6 approaches the master environmental geometric model 3, and finally the master and slave geometric models are consistent with the real environmental model.

Specific steps are as follows:

Step 1, the operator 1 sends the main-end position signal x _m (t) through the main-end robot 2, and the main-end position signal x _m (t) is input into the main-end environmental geometric model 3, and then input into the environmental impedance model 4, the environmental impedance Model 4 generates the impedance model output force signal f _m (t), namely

Among them, f _m (t) represents the output force signal of the impedance model, t is the time variable of the control system, k is the elastic coefficient of the environment model, x _m (t) represents the position signal of the main end, n is the surface geometry coefficient of the environment model, b is the model coefficient of the environment model, Indicates the derivative of x _m (t), and the output force signal f _m (t) of the impedance model is delayed by 1T in the delay communication link 5 to form f _m (tT) and enters the position correction calculation link 7 .

Step 2, the master-side position signal x _m (t) forms the slave-side position signal x _m (tT)) after a delay of 1T through the delay communication link 5, and the slave-side position signal x _m (tT) enters the slave side The end environment geometric model 6 forms the slave end control position signal x _c (tT), namely

x _c (t) = _Re (t+T) x _m (t),

Among them, x _c (t) represents the control position signal of the slave end, and _Re (t+T) represents the adjustable gain.

The slave-end control position signal x _c (tT) is input to the slave-end robot 8, so that the slave-end robot 8 moves and generates an environment position signal x _e (tT), which acts on the real environment 9 to generate a real environment output force signal f _e ( tT), namely

Among them, f _e (t) represents the real environment output force signal, and x _e (t) represents the environment position signal.

The force signal f _e (tT) output by the real environment 9 enters the position correction calculation link 7, and the position correction calculation link 7 is based on the f _m (tT) after the delay communication link 5 and the force signal f _e ( tT) outputs the adjustable gain _Re (tT), the algorithm used to calculate the adjustable gain _Re (t) in the position correction link 7 is an adaptive law based on the gradient method, assuming the output force f _m of the environmental impedance model 4 ( The difference between t) and the force signal f _e (t) output by the real environment is e(t), and the adaptive law adopted is β is the adjustment rate, and its value range is 0.01-2. In the simulation, β=0.2, the specific steps are as follows:

The first step is to set the gain rate β, set β=0.01-2,

In the second step, according to the two inputs f _e (t) and f _m (t) of the position correction calculation link 7, the generalized error e(t) is obtained,

_e (t)=fe(t)-fm( _t ),

The third step is to use the control rate formula to calculate _Re (t).

The adjustable gain _Re (tT) is input into the slave-end environmental geometric model 6, and the slave-end environmental geometric model 6 is revised according to _Re (tT), and the slave-end environmental geometric model 6 is obtained to obtain a new slave-end environmental geometric model 6, and enter the next cycle control,

Step 3, the operator 1 sends a new master-end position signal x _m (t-2T) through the master-end robot 2, and the new master-end position signal x _m (t-2T) is input into the master-end environmental geometric model 3, and then input into the environment The impedance model 4, the environment impedance model 4 generates a new impedance model output force signal f _m (t-2T).

Step 4, the new master position signal x _m (t-2T) forms a new slave side position signal x _m (t-3T) after a delay of 1T through the delay communication link 5, and the new environmental impedance model outputs force The signal f _m (t-2T) forms a new environmental impedance model output force signal f _m (t-3T) after being delayed by 1T through the time-delay communication link 5 and enters the position correction calculation link 7, and the new slave side position signal x _m (t -3T), enter the new slave-end environmental geometric model 6 to form a new slave-end control position signal x _c (t-3T), and the new slave-end control position signal x _c (t-3T) is input to the slave-end robot 8, so that the slave The terminal robot 8 moves and generates a new environment position signal x _e (t-3T), which acts on the real environment 9 to generate a new real environment output force signal f _e (t-3T), and the new real environment output force signal f _e ( t-3T) enters the position correction calculation link 7, the position correction calculation link 7 outputs the force signal f _m (t-3T) and the new real environment output force signal f _e (t) based on the new environmental impedance model after the delay communication link 5 -3T) outputs a new adjustable gain _Re (t-3T), and the new adjustable gain _Re (t-3T) is input into the revised slave-end environmental geometric model 6, and the revised slave-end environmental geometric model 6 The revised slave-end environmental geometric model 6 is further corrected according to the new adjustable gain R _e (t-3T), and then the revised new slave-end environmental geometric model 6 is used for the control of the next cycle.

Step 5 returns to Step 1.

In this cycle, the environmental geometric model 6 at the slave end gradually approaches the environmental geometric model 3 at the master end, and adaptive control is realized.

The present invention adaptively adjusts the state of the geometric model of the real environment through the interactive force feedback of the real environment and the output force signal of the master-end environment impedance model (corrects the slave-end environment model according to the master-end model, and adjusts the command output according to the position correction), so that The master-slave environment eventually tends to be consistent, fully overcome the negative impact of large delays, and achieve stable and accurate operations.

The above is only a preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made. It should be regarded as the protection scope of the present invention.

Claims (2)

1. An adaptive control method for a time-delay teleoperated robot based on an environmental impedance model, including a main-end robot (2), a main-end environmental geometric model (3), an environmental impedance model (4), and a time-delay communication link (5) , the geometric model of the slave environment (6), the position correction calculation link (7), the slave robot (8), the real environment (9), the environmental impedance model (4) is consistent with the real environment (9), and the geometric model of the master environment (3) there is a position deviation with the slave-end environment geometric model (6), let t be the time variable of the control system, and T be the time delay amount of the communication time delay link, it is characterized in that, comprises the following steps: Step 1, the master-end robot (2) sends out the master-end position signal x _m (t), and the master-end position signal x _m (t) is respectively input into the master-end environmental geometric model (3), the environmental impedance model (4) and In the time-delay communication link (5), the environmental impedance model (4) generates an impedance model output force signal f _m (t), and the environmental impedance model (4) is: Among them, f _m (t) represents the output force signal of the impedance model, t is the time variable of the control system, k is the elastic coefficient of the environment model, x _m (t) represents the position signal of the main end, n is the surface geometry coefficient of the environment model, b is the model coefficient of the environment model, Denotes the derivative of x _m (t); The impedance model output force signal f _m (t) forms a delayed 1T impedance model output force signal f _m (tT) after the time delay communication link (5) delays 1T and enters the position correction calculation link (7); Step 2, the master-side position signal x _m (t) forms the slave-side position signal x _m (tT) after a time-delay communication link (5) delay of 1T, and the slave-side position signal x _m (tT) enters the slave-side environment The geometric model (6) forms the slave-end control position signal x _c (tT), and the slave-end environmental geometric model (6) is: x _c (t) = _Re (t+T) x _m (t); Among them, x _c (t) represents the control position signal of the slave end, and R _e (t+T) represents the adjustable gain; The slave-end control position signal x _c (tT) is input to the slave-end robot (8), so that the slave-end robot (8) moves and generates an environmental position signal x _e (tT), which acts on the real environment (9) to generate a real environment Output force signal f _e (tT), the real environment is: f _e (t) represents the real environment output force signal, x _e (t) represents the environment position signal; The real environment output force signal f _e (tT) enters the position correction calculation link (7), and the position correction calculation link (7) outputs the force signal f _m (tT) and The real environment output force signal f _e (tT) outputs the adjustable gain R _e (tT), and the adjustable gain R _e (tT) is input into the slave-end environmental geometric model (6), and the slave-end environmental geometric model (6) is based on The adjustable gain R _e (tT) modifies the slave-end environmental geometric model (6), and uses the revised slave-end environmental geometric model (6) for the control of the next cycle; Step 3, the master-end robot (2) sends out a new master-end position signal x _m (t-2T), and the new master-end position signal x _m (t-2T) is respectively input into the master-end environment geometric model (3), the environment The impedance model (4) and the time-delay communication link (5), the environmental impedance model (4) generates a new impedance model output force signal f _m (t-2T); Step 4, the new master-side position signal x _m (t-2T) forms a new slave-side position signal x _m (t-3T) after a time-delay communication link (5) delay of 1T, and the new environmental impedance model outputs The force signal f _m (t-2T) passes through the delay communication link (5) to form a new environmental impedance model output force signal f _m (t-3T) after a delay of 1T and enters the position correction calculation link (7). The side position signal x _m (t-3T) enters the revised slave-end environmental geometric model (6) to form a new slave-end control position signal x _c (t-3T), and the new slave-end control position signal x _c (t- 3T) is input to the slave robot (8), so that the slave robot (8) moves and generates a new environment position signal x _e (t-3T), which acts on the real environment (9), and generates a new real environment output force signal f _e ( t-3T), the new real environment output force signal f _e (t-3T) enters the position correction calculation link (7), and the position correction calculation link (7) is based on the time delay 1T after the delay communication link (5) After the new environmental impedance model output force signal f _m (t-3T) and the new real environment output force signal f _e (t-3T) output a new adjustable gain R _e (t-3T), the new adjustable gain R _e (t-3T) inputs the revised slave-end environmental geometric model (6), and the revised slave-end environmental geometric model (6) corrects the new slave-end environmental geometric model according to the new adjustable gain R _e (t-3T) (6), use the revised new slave-end environment geometric model (6) for the control of the next cycle; Step 5 returns to step 1; By repeating the above steps, the environmental geometric model (6) at the slave end gradually approaches the environmental geometric model (3) at the master end, and adaptive control is realized. 2. the time delay teleoperated robot adaptive control method based on environmental impedance model according to claim 1, is characterized in that: the generation method of described adjustable gain _Re (t) is: The real environment output force signal f _e (t), the environmental impedance model output force signal f _m (t), and the control rate of the position correction calculation link (7) is β represents the gain ratio, and e(t) represents the generalized error. The specific steps are as follows: The first step is to set the gain rate β=0.01-2; In the second step, the generalized error e(t) is obtained according to the real environment output force signal f _e (t) and the environmental impedance model output force signal f _m (t) input to the position correction calculation link (7); e(t) = f _e (t) - f _m (t); The third step is to use the control rate formula to calculate _Re (t).