Control system and control method

Abstract

To improve control stability and accuracy even when communication delays fluctuate.SOLUTION: A control system includes a prediction control unit that transmits a control command value to a control target via a communication network to perform control and uses a set including a predicted control command value for each of a plurality of points in the future than the transmission time as the control command value, a command delay measurement unit that measures the delay in reaching the control command value on the control target side of the communication network, and a command selection unit that selects the control command value from the set at a point in time corresponding to the arrival delay measured by the command delay measurement unit.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a control system and a control method. [Background technology]

[0002] In recent years, technological development of next-generation mobile communication systems is expanding from mobile applications to industrial applications. Three features are proposed for next-generation mobile communication systems: ultra-high speed and large capacity, multiple simultaneous connections, and ultra-reliable low latency. Low latency is particularly important for realizing remote motion control of IoT (Internet of Things) terminals.

[0003] Even if low latency is achieved, communication latency is not constant but constantly fluctuates, so technology that can accommodate fluctuations in communication latency is required. For example, Patent Document 1 proposes a control system that compensates for communication delays using a Smith predictor. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2022-9740 Summary of the Invention [Problem to be solved by the invention]

[0005] Telecommunications carriers have technology to predict the amount of communication delay including fluctuations, so they can apply the predicted communication delay to the Smith predictor of Patent Document 1. However, in the case of general companies, it is not possible to accurately predict communication delay including communication delay fluctuations. For this reason, in the technology of Patent Document 1, there is a risk that fluctuations in communication delays may cause control instability, and that steady-state deviations may occur due to the superposition of delay fluctuations and disturbances, etc. Although the technology of Patent Document 1 is a feedback control technology, instability and steady-state deviations associated with fluctuations in communication delays may also occur in the case of feedforward control.

[0006] Therefore, an object of the present invention is to stabilize control and improve accuracy even when fluctuations occur in communication delay. [Means for solving the problem]

[0007] In order to solve the above problem, a control system according to one embodiment of the present invention comprises a predictive control unit that performs control by transmitting a control command value to a controlled object side via a communication network, and uses as the control command value a set including control command values ​​predicted for each of a plurality of time points in the future from the transmission time, a command delay measurement unit that measures the arrival delay of the control command value on the controlled object side of the communication network, and a command selection unit that selects from the set the control command value at a time point corresponding to the arrival delay measured by the command delay measurement unit.

[0008] According to such a control system, the controlled object is controlled by a control command value corresponding to the measured arrival delay, so that even if there is a fluctuation in the delay, control can be achieved by an appropriate control command value, thereby stabilizing control and improving accuracy. In the above control system, it is desirable that the predictive control unit performs feedback control by reflecting the control result in the controlled object via the communication network, and further comprises a result delay measurement unit provided on the predictive control unit side of the communication network for measuring the arrival delay of the control result, and a state estimation unit for estimating a change in the control state in the controlled object due to the arrival delay measured by the result delay measurement unit from the control result and the control command value, and reflecting the change in the control state in the predictive control unit in the feedback control.

[0009] According to such a control system, the accuracy of estimating the control state is improved by measuring the arrival delay of the control result, and therefore the accuracy of calculating a control command value at a future point in time is also improved. In the control system, it is preferable that the command delay measurement unit measures the arrival delay by comparing a timestamp added to the control command value with the reception time. The use of the timestamp enables accurate and easy delay measurement.

[0010] In the control system, it is also preferable that the result delay measurement unit measures the arrival delay by comparing a timestamp added to the control result with the reception time. In addition, it is preferable that the control system further comprises a disturbance observer provided on the controlled object side of the communication network for compensating for disturbances to the controlled object and parameter errors of the controlled object. The disturbance observer suppresses disturbances and the like, improving the calculation accuracy of a control command value at a future time point.

[0011] In addition, it is preferable that the control system further comprises a tracking control unit provided on the controlled object side of the communication network for making the controlled object track the control command value. By providing a tracking control unit on the controlled object side of the communication network, the degree of freedom of the format of the control command value is increased. In the control system, when the command selection unit is unable to receive the set of control command values, it is preferable that the command selection unit selects a control command value from a set previously received. The set of previously received control command values ​​includes a control command value that is more appropriate for a future time than the selected control command value. Therefore, when a packet is lost, the previously received control command value is selected as a substitute for the control command value, thereby improving the accuracy of control.

[0012] In the control system, the predictive control unit may determine the control command value so as to reduce a deviation of a control state of the controlled object from a control target, thereby making it possible to make the state of the controlled object follow the control target. In addition, in the above control system, it is preferable that at least one of position, angle, force, and torque is used as the control target. Position, angle, force, and torque are control amounts that are practically required for the operation of a controlled object, and are suitable as control targets.

[0013] In the above control system, it is preferable that a command value indicating at least one of torque, force, current, and voltage is used as the control command value. Torque, force, current, and voltage are basic elements of control, including when controlling control amounts such as position and angle, and therefore are suitable as control command values.

[0014] In order to solve the above problem, a control method according to one embodiment of the present invention includes a predictive control step of transmitting a control command value to a controlled object side via a communication network to perform control, and using as the control command value a set including control command values ​​predicted for each of a plurality of time points in the future from a transmission time, a command delay measurement step of measuring an arrival delay of the control command value on the controlled object side of the communication network, and a command selection step of selecting from the set the control command value at a time point corresponding to the arrival delay measured by the command delay measurement step.

[0015] According to this control method, the controlled object is controlled by a control command value corresponding to the measured arrival delay, so that even if there is a fluctuation in the delay, control can be achieved by an appropriate control command value, thereby stabilizing control and improving accuracy. Effect of the Invention

[0016] According to the present invention, it is possible to stabilize control and improve accuracy even when fluctuations occur in communication delay. [Brief description of the drawings]

[0017] [Figure 1] FIG. 1 is a diagram showing a remote operation system to which a control system of the present invention is applied. [Diagram 2] FIG. 1 is a diagram illustrating a first comparative example. [Diagram 3] FIG. 11 is a diagram illustrating a second comparative example. [Figure 4] FIG. 2 is a diagram showing a control system in a remote control system of the present embodiment. [Diagram 5] 11 is a graph showing a schematic prediction of a state quantity. [Figure 6] FIG. 2 is a diagram illustrating an internal configuration of a prediction control unit. [Figure 7] FIG. 2 is a diagram illustrating an arrangement of manipulated variables. [Figure 8] 11 is a graph showing an example of an operation amount when a packet loss occurs; [Figure 9] FIG. 2 is a block diagram of a disturbance observer. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. However, in order to avoid the following description becoming unnecessarily redundant and to facilitate understanding by those skilled in the art, more detailed description than necessary may be omitted. For example, detailed description of already well-known matters and duplicated description of substantially the same configuration may be omitted. In addition, elements shown in the previously described figures may be appropriately referenced in the description of the later figures.

[0019] <Remote control system> FIG. 1 is a diagram showing a remote control system 1 to which a control system of the present invention is applied. The remote control system 1 includes an operating side device 10, a communication network 20, and a remote side device 30.

[0020] The operating side device 10 outputs a control command value for controlling the operation of the remote side device 30. The operating side device 10 includes, for example, a controller 11 that calculates a control command value based on a program or the like, and an operation device 12 that receives an operation by an operator and outputs a control command value corresponding to the operation. Specific shapes of the operation device 12 include, for example, a joystick 12a and a steering wheel 12b. The operating side device 10 may be a composite device in which the operation device 12 and the controller 11 are combined.

[0021] The operating side device 10 is given a control target for at least one of the position, angle, force, and torque in the remote side device 30, and outputs at least one control command value of force, torque, current, and voltage to realize the control target. Position, angle, force, and torque are control amounts required for practical purposes such as the operation of the remote side device 30, and are suitable as control targets. Furthermore, force, torque, current, and voltage are basic elements of control, including when controlling control amounts such as position and angle, and are therefore suitable as control command values. In the following, a representative example will be described in which a control target of angle and a command value of torque are adopted. Furthermore, the control command value may be referred to as an operation amount.

[0022] The communication network 20 is, for example, a network such as 5G or LPWA (Low Power Wide Area), and transmits a control command value (torque command) output by the operating side device 10 to the remote side device 30. The remote device 30 is a device that is remotely operated by a control command value. The remote device 30 includes, for example, a laboratory device 31 and a heavy machine 32, as well as a transport robot, a surgical robot, a machine tool, and an XY table. The remote device 30 includes an actuator, for example, a motor or a hydraulic device, and a mechanism that is driven by the actuator. The actuator of the remote device 30 includes a drive control unit that receives a control command value for a drive torque and generates a drive torque indicated by the control command value with the actuator. The drive control unit corresponds to an example of a tracking control unit that causes the output of the actuator to track the control command value.

[0023] In the remote side device 30, the angle, torque, etc. generated in the mechanism as a result of the drive by the actuator are detected by a sensor or calculated as a state quantity indicating the control state. The state quantity obtained in the remote side device 30 is returned to the operating side as a monitor value of the remote side device 30 via the communication network 20. When operating the remote side device 30 in the remote operation system 1, a communication delay for the outbound path of the communication network 20 occurs, and when monitoring the control state of the remote side device 30, a communication delay for the return path of the communication network 20 occurs.

[0024] <First Comparative Example> FIG. 2 is a diagram showing a first comparative example for the control system 50 (see FIG. 4) in the remote control system 1 of this embodiment, and FIG. 3 is a diagram showing a second comparative example. The control system 40 of the first comparative example executes feedback control.

[0025] The control system 40 of the first comparative example includes a position control unit 15 and an actuator model 16 in the operating side device 10, an actuator 35 and a mechanical part 36 in the remote side device 30, and a communication network 20. The control target in the control system 40 is the mechanical part 36, and as an example, the output angle θ of the mechanical part 36 is controlled.

[0026] The control system 40 has a control target for the mechanism portion 36, for example, a target angle value θ * The position control unit 15 receives the target angle θ * and the monitor value of the output angle θ in the mechanism portion 36 is input, and the position control portion 15 adjusts the current command value I * Calculate.

[0027] Current command value I * is converted into a manipulated variable u (e.g., a torque command value) via the actuator model 16 and input to the actuator 35 via the communication network 20. The manipulated variable u is subject to a communication delay 22 on the way thereto due to communication via the communication network 20, and the manipulated variable u arrives at the remote side with a delay of a delay time td2.

[0028] The actuator 35 outputs torque T corresponding to the operation amount u. * The disturbance torque Td is generated and is added to the output of the actuator 35. The disturbance torque Td includes an external force transmitted from the outside of the remote device 30 and a torque T * Response errors and so on are included.

[0029] A mechanical part 36 is driven by the torque of the actuator 35 to which the disturbance torque Td has been added, and the mechanical part 36 outputs an angle θ. The output angle θ of the mechanical part 36 is monitored by a sensor or the like, and the monitored value is fed back to the operation side via the communication network 20. A return communication delay 21 occurs in the monitored value of the output angle θ due to communication via the communication network 20, and the monitored value reaches the operation side with a delay of a delay time td1.

[0030] As described above, the monitored value of the fed back output angle θ and the target angle value θ * The difference between these is calculated and input to the position control unit 15. Although angle control is shown as an example above, the control in remote control system 1 may be control of angular velocity or angular acceleration, or may be two-variable control of angle and angular velocity.

[0031] A transfer function G(s) in the control system 40 of the comparative example shown in FIG.

number

[0032] In the first comparative example, since the denominator polynomial of the transfer function G(s) includes a communication delay term, the larger the communication delay amount, the more unstable the control becomes. In addition, there is a risk of deterioration of responsiveness and steady-state deviation occurring due to disturbance torque Td acting on the actuator 35 on the remote device 30 side and parameter errors that may be included in the control model C(s) on the operating side.

[0033] For this reason, in the second comparative example, the state on the remote side is predicted using a model of communication delay and a model of the controlled object, and the feedback loop is stabilized. <Second Comparative Example>

[0034] 3, a control system 41 of the second comparative example is provided with a Smith predictor 42 on the operation side. The function of the Smith prediction is expressed by the following equation (2).

number

[0035] That is, the Smith predictor 42 predicts the state of the remote side based on the nominal model Pn of the mechanical part 36 to be controlled and the model values ​​of the delay times td1 and td2 to compensate for the feedback control. The transfer function Gs(s) in the control system 41 of the second comparative example equipped with the Smith predictor 42 is expressed by the formula (4) when the condition of the following formula (3) is satisfied.

number

[0036]

number

[0037] Since the communication delay term is eliminated from the denominator polynomial of the transfer function Gs(s), the feedback control is stable as long as the condition of formula (3) is satisfied. However, since the communication delay times td1 and td2 include error parts tde1 and tde2, if there is no technology to accurately predict fluctuations at future times, the state prediction by the Smith predictor 42 will also be inaccurate, and there is a risk of control instability or steady-state deviation occurring. Therefore, in the remote control system 1 of this embodiment, a control system 50 described below is used.

[0038] <Present embodiment> FIG. 4 is a diagram showing a control system 50 in the remote control system 1 of this embodiment. The control system 50 in this embodiment includes a prediction control unit 51, a time stamp providing unit 52, and a delay measurement unit 53 in the operating side device 10. The control system 50 also includes an actuator 35, a mechanical part 36, a delay measurement unit 54, an optimal value selection unit 55, a disturbance observer 56, and a time stamp providing unit 57 in the remote side device 30. In the control system 50 in this embodiment, the operating side device 10 also controls the remote side device 30 via the communication network 20. The controlled object in the control system 50 is the mechanical part 36, and as described above, the output angle θ of the mechanical part 36 is controlled as an example.

[0039] As in the comparative example, the control system 50 in this embodiment has a target angle value θ as a control target for, for example, the output angle θ of the mechanical part 36. * is input, and each target value θ * is input as an array, for example. * For example, if the trajectory at a future time point is determined by automatic operation such as a program, each target value θ * In the case of manual operation, for example, each target value θ * may be input.

[0040] On the other hand, the state quantity θ on the remote side is fed back to the operating side device 10 via a communication delay 21 on the return path, so that the operating side device 10 obtains a state quantity θ-td1 that is the past by the delay time td1. A time stamp ts is added to the state quantity θ fed back to the operating side device 10 by a time stamp adding unit 57 of the remote side device 30. Then, a delay measuring unit 53 of the operating side device 10 measures the delay time td1 on the return path (i.e., the arrival delay of the state quantity θ) by comparing the time stamp ts added to the state quantity θ with the reception time indicated by the internal clock.

[0041] <Predictive Control> The prediction control unit 51 of the operation side device 10 predicts the state quantity θ of the controlled object from the past to the future, and calculates a future target value θ * The operation amount u at each future point in time based on the above is output, for example as an array.

[0042] FIG. 5 is a graph that illustrates a prediction of the state quantity θ, and FIG. 6 is a diagram illustrating the internal configuration of the prediction control unit 51. As shown in FIG. The horizontal axis in FIG. 5 indicates time, and the vertical axis indicates state quantity. With the current time set to "0", the prediction control unit 51 predicts each state quantity θ from a state quantity θ-td1 in the past by a delay time td1 on the return path to a state quantity θtd2 in the future by a delay time td2 on the outward path.

[0043] The prediction control unit 51 includes a state prediction unit 61 and a future control unit 62, and prediction from a past state quantity θ-td1 to a present state quantity θ0 is executed by the state prediction unit 61. That is, the state prediction unit 61 receives an input of the delay time td1 measured by the delay measurement unit 53 of the operation side device 10 and the fed back past state quantity θ-td1, and outputs a predicted value of the state quantity θ0 at each time point up to the present based on the past operation quantity u and the nominal model Pn(s) of the mechanical part 36.

[0044] On the other hand, prediction of the state quantity θ at each time point from the future immediately following the present to the future by the delay time td2 of the outward journey is executed by the future control unit 62. That is, the future control unit 62 calculates a predicted value of the current state quantity θ0 and a target value θ for the present and future. *Based on this, prediction is made up to the future state quantity θtd2. In addition, the future control unit 62 calculates the predicted state quantity θ and the target value θ * The manipulated variable u at a future point in time is calculated based on the difference between the current state quantity θ and the current state quantity θ. The manipulated variable u calculated by the future control unit 62 is transmitted to the remote device 30 for control, and is also stored in the state prediction unit 61 for use in predicting the past state quantity θ.

[0045] The prediction of the state quantity θ will be described in detail below. In the following description, it is assumed that the manipulated variable u is sent from the operating side device 10 to the remote side device 30 at each step of 1 ms in the control system 50 and used for control.

[0046] When the return path delay time td1 is, for example, 50 ms, the state prediction unit 61 uses each operation amount u from the operation amount u-50 of 50 steps ago (i.e., 50 ms ago) to the operation amount u0 of the immediately previous step to predict the state amount θ, as shown in the following equation (5).

number

[0047] As the predicted state quantity θ, each state quantity θ is predicted step by step by the state prediction unit 61 from the state quantity θ-50 50 steps ago, which corresponds to the state quantity θ-td1 fed back with a delay of the delay time td1, to the current state quantity θ0, as shown in the following formula (6). In other words, the state prediction unit 61 estimates a change in the control state of the controlled object for the measured delay time td1 from the past state quantity θ-td1 and the past manipulated variable u.

number

[0048] As a specific procedure, first, the state quantity θ-49 49 steps ago is calculated based on the state quantity θ-50 50 steps ago, the operation quantity u-50, and the nominal model Pn(s) of the mechanical part 36. Next, the state quantity θ-48 48 steps ago is calculated based on the state quantity θ-49 49 steps ago, the operation quantity u-49, and the nominal model Pn(s). This procedure is repeated to calculate each state quantity θ up to the current state quantity θ0. In this way, by making a prediction based on the delay time td1 measured when feeding back the state quantity θ, the state quantity θ can be predicted with high accuracy even if a fluctuation occurs in the delay time td1.

[0049] In the future control unit 62, as shown in the following equations (7) and (8), each manipulated variable u and each state variable θ are predicted for each of a plurality of time points in the future.

number

[0050]

number

[0051] The operation amount u and state amount θ required for control are the operation amount u and state amount θ that are in the future by the delay time td2 on the outbound route. However, since the delay time td2 on the outbound route is undetermined before the operation amount u is transmitted, the future control unit 62 calculates the operation amount u and state amount θ with a margin of error, for example, up to a maximum of 100 ms. As a specific procedure, first, the current state quantity θ0 calculated by the state prediction unit 61 and the current target value θ * The manipulated variable u+1 for one step in the future is calculated based on 0 and the control function C(s). Next, the state variable θ1 for one step in the future is calculated based on the manipulated variable u+1 for one step in the future, the current state variable θ0, and the nominal model Pn(s).

[0052] Furthermore, the state quantity θ1 and the target value θ *Based on +1 and the control function C(s), the manipulated variable u+2 two steps into the future is calculated. This procedure is repeated, and the manipulated variable u and the state variable θ are calculated alternately, one step at a time, up to 100 ms. Note that, although a fixed control function C(s) is used here as an example, the manipulated variable u may be calculated using optimal control such as model predictive control.

[0053] The operation amount u of each step up to 100 ms in the future calculated in this manner is transmitted, for example, as an array from the operating side device 10 to the remote side device 30. When transmitting, a timestamp ts is provided by the timestamp providing unit 52 to the array of the operation amount u.

[0054] FIG. 7 is a diagram showing a schematic arrangement of the manipulated variable u. 7 includes the manipulated variable u+1, which is the optimal value 1 ms after the transmission time, to the manipulated variable u+100, which is the optimal value 100 ms after the transmission time. A timestamp (e.g., ts2) indicating the transmission time is added to the manipulated variable u array.

[0055] When the sequence of the operation amounts u with the time stamps is received by the remote side device 30, the delay measurement unit 54 in Fig. 4 measures the delay time td2 of the outward path by comparing the time stamp with the reception time indicated by the internal clock. Then, the optimum value selection unit 55 selects the operation amount u at the time point corresponding to the delay time td2 from among the operation amounts u in the sequence. As shown in Fig. 7, for example, when the delay time td2 is 51 ms, the operation amount u+51, which is the optimum value 51 ms later, is selected. Note that, for example, the measured delay time td2 or the like is notified from the remote side device 30 to the operating side device 10, so that the prediction control unit 51 of the operating side device 10 can recognize which operation amount u has been selected from the sequence.

[0056] In this way, the manipulated variable u predicted for each future point in time is sent from the operating side to the remote side, and the manipulated variable u corresponding to the delay time td2 measured on the remote side is selected, thereby enabling accurate control even if the delay time td2 fluctuates. Note that the effects obtained by predicting and selecting the manipulated variable u can also be obtained when applied to feedforward control.

[0057] <Countermeasures against packet loss> If the delay time td2 fluctuates greatly, packet loss may occur, in which the sequence of the manipulated variable u itself is lost. When packet loss occurs, the optimum value selection unit 55 uses the sequence already received in the previous step to select the manipulated variable u corresponding to the next step of the selected manipulated variable u as a replacement for the lost manipulated variable. In the example shown in Fig. 7, when packet loss occurs at the next step (i.e., at the timing of timestamp ts3) after 51 ms, in which the manipulated variable u+51, which is the optimum value, is selected, the manipulated variable u+52 one step ahead is selected as a replacement.

[0058] Furthermore, if packet loss continues even at the timing of timestamp ts4, the manipulated variable u+53, which is two steps ahead, is selected as a substitute. Then, if the sequence of manipulated variables u is received at the timing of timestamp ts5, the manipulated variable u corresponding to the delay time td2 is selected by the optimum value selection unit 55 from the sequence of the new manipulated variables u. By repeating such a procedure, even if the sequence of the manipulated variables u is not received at the scheduled reception timing, a substitute manipulated variable u can be obtained.

[0059] FIG. 8 is a graph showing an example of the operation amount u when a packet loss occurs. The horizontal axis of FIG. 8 indicates the reception timing as a time stamp, and the vertical axis indicates the manipulated variable u. In the comparative example shown by the dotted line, an alternative manipulated variable u is not used, and the manipulated variable u selected at the timestamp ts2 before the packet loss occurs is held as the manipulated variable u at the timestamps ts3 and ts4 when the packet loss occurs. For this reason, the remote device 30 does not operate smoothly, and there is a risk of a large error occurring in response to the command.

[0060] In contrast to this, in the present embodiment shown by the solid line, at the times of timestamps ts3 and ts4 where packet loss occurs, the operation amount u predicted at the timestamp ts2 before the packet loss occurs is used as a substitute, so that the remote side device 30 operates smoothly and the error with the command is suppressed.

[0061] <Disturbance observer> The disturbance observer 56 of the remote device 30 shown in FIG. 4 improves prediction accuracy by reducing disturbance Td acting on the controlled object and parameter fluctuations. As described above, the state prediction in the predictive control unit 51 is performed based on the nominal model Pn(s) of the mechanical part 36 to be controlled. Therefore, when the viscous friction B changes due to the environmental temperature or the like, causing a parameter error of the model (i.e., a model error), there is a risk of disturbance in the transient response of the control. In addition, when a disturbance such as a disturbance torque Td occurs in the remote device 30, there is a risk of deviation in the prediction by the predictive control unit 51, and a steady-state deviation from the target value occurs in the control by the control system 50.

[0062] Therefore, in this embodiment, the disturbance observer 56 is used to reduce the disturbance Td and parameter fluctuations. Fig. 9 is a block diagram of the disturbance observer 56. In Fig. 9, the internal configuration of the disturbance observer 56 is shown, and the configuration of the peripheral portion of the disturbance observer 56 is also shown.

[0063] The disturbance torque estimated value T(^)d estimated by the disturbance observer 56 is calculated by the operation amount u selected by the optimal value selection unit 55. td2 is used to correct the torque command value represented by* is input to the actuator 35. The actuator 35 outputs a torque command value T * The torque command value T * is corrected by the disturbance torque estimate T(^)d, the additional disturbance torque Td is cancelled. Note that the disturbance observer 56 calculates the disturbance torque estimate T(^)d including the operation fluctuation due to parameter fluctuation of the mechanical part 36 as part of the disturbance torque Td, so the parameter fluctuation is also reduced by the correction using the disturbance torque estimate T(^)d.

[0064] Therefore, as a result of the correction using the disturbance torque estimate value T(^)d, the mechanical part 36 is driven with the accurate torque required for control, and an accurate output angle θ is obtained. From another point of view, the disturbance observer 56 has a role of correcting the behavior of the mechanical part 36 to the behavior of the nominal model Pn(s).

[0065] The nominal model Pn is composed of a nominal moment of inertia Jn and a viscous friction coefficient Bn, and the respective model errors ΔJ and ΔB have the relationship shown in Equation (9) below. J=Jn+ΔJ, B=Bn+ΔB ……(9) The disturbance observer 56 includes an inverse model 64 of the mechanism portion 36 and a low-pass filter 63. The disturbance observer 56 of this embodiment has a structure that does not use differentiation.

[0066] The low-pass filter 63 receives the torque command value T * is input to the inverse model 64, and the angle θ is input to the inverse model 64. Then, the value obtained by subtracting the output of the low-pass filter 63 from the output of the inverse model 64 becomes the disturbance torque estimated value T(^)d. It is preferable to use the recursive least squares method (RLS) or the fixed trace algorithm (FTA) to improve accuracy when determining the nominal parameters in the disturbance observer 56. RLS is an abbreviation for Recursive Least Squares. FTA is an abbreviation for Fixed Trace Algorithm.

[0067] The block diagram of the disturbance observer 56 shown in FIG. 9 can be expressed by the following equations (10) and (11), and the disturbance torque estimate T(^)d by the disturbance observer 56 is expressed by the following equation (12).

number

[0068]

number

[0069]

number

[0070] If the time constant τ is set small, the disturbance torque estimated value T(^)d can be estimated as including both the disturbance torque Td and the model errors ΔJ and ΔB of the controlled object 36. The estimated disturbance torque T(^)d is * By adding this to, the model error of the controlled object 36 and the disturbance torque Td are cancelled out.

[0071] It should be noted that (τs+1) in the denominators of the low-pass filter 63 and the inverse model 64 may be cubed instead of squared. [Explanation of symbols]

[0072] 1. Remote control system 10…Operation side device 11…Controller 12. Operation device 20…Communications Network 21,22…Communication delay 30...Remote device 35…Actuator 36…mechanism part 50…Control system 51...Prediction control unit 52, 57...Time stamp assignment section 53, 54…Delay measurement section 55…Optimum value selection section 56…Disturbance observer 61... State prediction unit 62…Future Control Unit 63...Low pass filter 64…Inverse model of the control object

Claims

1. a predictive control unit that transmits a control command value to a controlled object via a communication network to perform control, and uses, as the control command value, a set including control command values ​​predicted for each of a plurality of time points in the future from the transmission time; a command delay measurement unit that measures an arrival delay of the control command value on the controlled object side of the communication network; a command selection unit that selects, from the set, the control command value at a time point corresponding to the arrival delay measured by the command delay measurement unit; A control system comprising:

2. The prediction control unit performs feedback control by reflecting a control result of the control target via the communication network, a result delay measurement unit that measures an arrival delay of the control result on the prediction control unit side of the communication network; a state estimation unit that estimates a change in a control state of the control target due to the arrival delay measured by the result delay measurement unit from the control result and the control command value, and reflects the change in the control state in the feedback control of the prediction control unit; The control system of claim 1 further comprising:

3. 2. The control system according to claim 1, wherein the command delay measurement unit measures the arrival delay by comparing a timestamp added to the control command value with a reception time.

4. The control system according to claim 2 , wherein the result delay measurement unit measures the arrival delay by comparing a timestamp added to the control result with a reception time.

5. 2. The control system according to claim 1, further comprising a disturbance observer provided on the controlled object side of the communication network, for compensating for disturbances to the controlled object and parameter errors of the controlled object.

6. 2. The control system according to claim 1, further comprising a tracking control unit provided on the controlled object side of the communication network, the tracking control unit causing the controlled object to track the control command value.

7. The control system according to claim 1 , wherein, when the set of control command values ​​cannot be received, the command selection unit selects control command values ​​from a set that has been received in the past.

8. The control system according to claim 1 , wherein the predictive control unit determines the control command value so as to reduce a deviation of a control state of the controlled object from a control target.

9. The control system according to claim 8 , wherein at least one of a position, an angle, a force, and a torque is used as the control objective.

10. 2. The control system according to claim 1, wherein a command value indicating at least one of torque, force, current, and voltage is used as the control command value.

11. a predictive control step of transmitting a control command value to a controlled object side via a communication network to perform control, and using, as the control command value, a set including control command values ​​predicted for each of a plurality of time points in the future from the transmission time; a command delay measuring step of measuring an arrival delay of the control command value on the controlled object side of the communication network; a command selection step of selecting, from the set, the control command value at a time point corresponding to the arrival delay measured by the command delay measurement step; The control method includes:

Images