JP2017041223A - Feedback control simulation device, control device, feedback control simulation method, and feedback control simulation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a control signal for accurately controlling an object to be controlled without performing complicate calculation.SOLUTION: A feedback control simulation device 12 includes: a control model 16 corresponding to a control object 20; and a controller 18 for a control model for, on the basis of a target value r of an output signal yto be output from the control model 16 and an output signal youtput from the control model 16 in a state that a virtual disturbance dv is input to the control model 16, generating a control signal usuch that the output signal youtput from the control model 16 becomes the target value r, and for outputting the control signal uto the control model 16.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a feedback control simulation device, a control device, a feedback control simulation method, and a feedback control simulation program.

  Conventionally, there is known a feedback control simulation device that compares an output target value with an actual output value and performs control so that the output value automatically matches the target value (see, for example, Patent Documents 1 and 2). There is also known a feed-forward control device that predicts a disturbance that causes a change in output and cancels the disturbance in advance. Furthermore, an apparatus that combines feedback control and feedforward control is known (see, for example, Patent Documents 3 to 5).

  Feedback control and feedforward control are applied to technologies in various fields, for example, a vehicle driving support system. In recent years, driving support systems have been put to practical use such as following a preceding vehicle, avoiding a collision with a preceding vehicle, and maintaining a lane. As described above, the application range of driving assistance has been expanded, but it has not yet been realized that driving without accidents. Causes of accidents include oversight of danger and insufficient attention to sudden jumps. In order to cope with these causes, it is necessary to perform driving in consideration of a plurality of caution points in a wide range at the same time. For this reason, in order to consider a plurality of cautions, a control using an artificial potential method has been devised that can add up the elements to be dealt with. However, when considering fully automatic driving, which is the final goal of the driving support system, it is difficult to set a free route and speed with the artificial potential method that performs route planning.

Special table 2009-21334 gazette JP-T 2009-521751 JP 2010-282618 A JP 2010-218008 A JP 2014-13607 A

  For example, conventionally, two-step control is first performed: a step of calculating a trajectory for avoiding an obstacle by the potential method, and a step of determining an operation (such as a handle operation amount) that causes the control target to follow the calculated trajectory. There is a problem that calculation is complicated.

  The present invention provides a feedback control simulation device, a control device, a feedback control simulation method, and a feedback control simulation program capable of generating a control signal for accurately controlling a controlled object without performing complicated calculations. It is an object.

  In order to solve the above-described problem, a feedback control simulation apparatus according to a first aspect of the present invention provides a control model corresponding to a control target, a target value of a first output signal output from the control model, and the control model. Based on the second output signal output from the control model in a state where a virtual disturbance is input, a control signal such that the first output signal output from the control model becomes the target value. A control model controller that generates and outputs the control model to the control model.

  The invention according to claim 2 includes a control signal output unit for outputting the control signal output from the control model controller to the control target.

  According to a third aspect of the present invention, an equivalent virtual disturbance that is a signal added to the control signal to obtain a state equivalent to a state in which the virtual disturbance is input to the control model is generated based on the virtual disturbance. A virtual disturbance generation unit is provided, and the control signal output unit outputs a control signal obtained by adding the control signal and the equivalent virtual disturbance to the control target.

  A fourth aspect of the invention includes an output signal output unit for outputting the first output signal output from the control model to a control target controller that controls the control target.

  According to a fifth aspect of the present invention, the object to be controlled is a moving body, the target value is a target trajectory of the moving body, and the virtual disturbance virtualizes an obstacle that becomes an obstacle when the moving body moves. It is a disturbance, and the first output signal is an obstacle avoidance trajectory avoiding the obstacle.

  According to a sixth aspect of the present invention, the virtual disturbance is a disturbance obtained from a potential function.

  According to a seventh aspect of the invention, the moving body is a vehicle, and the potential function is a function that gives a repulsive force to the vehicle.

  According to an eighth aspect of the present invention, the potential function is a function set so that the repulsive force becomes maximum at an intersection entrance where the vehicle enters.

  According to a ninth aspect of the present invention, the movable body has two arms, and the potential function is set by regarding each arm as the obstacle.

  A control device according to a tenth aspect of the invention controls the control target based on the feedback control simulation device according to the fourth aspect and the first output signal output from the control model of the feedback control simulation device. A control target controller.

  In the invention described in claim 11, the controller to be controlled is an inverse model of the object to be controlled.

  The feedback control simulation method according to claim 12 is a method in which a computer inputs a target value of a first output signal output from a control model corresponding to a control target and a virtual disturbance to the control model. Based on the second output signal output from the control model, a control signal is generated such that the first output signal output from the control model becomes the target value and is output to the control model. Includes steps.

  A feedback control simulation program according to a thirteenth aspect of the present invention is a feedback control simulation program for causing a computer to function as each means constituting the feedback control simulation apparatus according to any one of the first to ninth aspects.

  According to the present invention, there is an effect that it is possible to generate a control signal for accurately controlling an object to be controlled without performing complicated calculations.

It is a block diagram of control device 10 concerning a 1st embodiment. It is a flowchart of the feedback control simulation process performed with a feedback control simulation apparatus. It is a block diagram of the feedback control simulation apparatus which concerns on 2nd Embodiment. It is a block diagram of the feedback control simulation apparatus which concerns on 3rd Embodiment. It is a block diagram of the feedback control simulation apparatus which concerns on 4th Embodiment. It is a figure for demonstrating the repulsion from an obstruction. It is a figure for demonstrating the potential field generate | occur | produced around an obstruction. It is a graph for demonstrating a potential function. It is a figure for demonstrating the simulation result of parking vehicle overtaking. It is a figure for demonstrating the simulation result of a no-signal intersection passing. It is a figure for demonstrating the concept of the module which concerns on 5th Embodiment. It is a block diagram of the feedback control simulation apparatus which concerns on 5th Embodiment. It is a figure for demonstrating the deviation suppression input in the collision avoidance of an arm. It is a figure for demonstrating the collision avoidance of an arm. It is a figure which shows an example of a potential field. It is a figure which shows the simulation result of the collision avoidance of an arm. It is a figure which shows the simulation result of the collision avoidance of an arm. It is a block diagram of the control apparatus 10 which concerns on 6th Embodiment. It is a block diagram of the feedback control simulation apparatus which concerns on 7th Embodiment. It is a block diagram of a control device concerning an 8th embodiment. It is a block diagram of a control device concerning a 9th embodiment.

  Hereinafter, an example of an embodiment of the present invention will be described with reference to the drawings.

  (First embodiment)

  In FIG. 1, the block diagram of the control apparatus 10 which concerns on this embodiment was shown. As shown in FIG. 1, the control device 10 includes a feedback control simulation device 12 and a control target controller 14.

  The feedback control simulation device 12 includes a control model 16 and a control model controller 18.

  The control model 16 is a control model corresponding to the control target 20, that is, a simulation model of the control target 20.

  As the control object 20, for example, a mobile object such as an automobile, a flying object (such as a drone), an industrial robot, or a humanoid robot is applied, but is not limited thereto. In the present embodiment, a case where the control target 20 is a moving body will be described.

Control model controller 18 is output from the first and the target value r of the output signal y _M as the output signal, the control model 16 in a state in which the input virtual disturbance dv the control model 16 which is output from the control model 16 the output signal y _m as a second output signal for feedback control was based on the output signal y _M output from the control model 16 generates a control signal u _M such as to follow the target value r Output to the control model 16. The output signal y _m for feedback control, not only the output signal y _M, a signal including observables other than the output signal y _M.

That is, the control model controller 18 controls performs feedback control on the control model 16 is a simulation model of the target 20, and generates and controls signals u _M as the output signal y _M to follow the target value r control Output to model 16.

Here, the target value r is, for example, a target trajectory of the control target 20 that is a moving body. Further, the virtual disturbance dv is a disturbance in which an obstacle that becomes an obstacle when the control target 20 moves is assumed as an example. The output signal y _M is a obstacle avoidance route that avoids obstacles. Note that the disturbance obtained from the potential function can be applied to the virtual disturbance. However, the disturbance is not limited to this as long as the obstacle is an obstacle that becomes an obstacle when the control target 20 moves.

The feedback control simulation apparatus 12 includes a control signal output section 22 for outputting an output signal y _M output from the control model 16, the control target controller 14 for controlling the controlled object 20.

Control target controller 14, the output signal y _M outputted from the feedback control simulation apparatus 12, and the output signal y _r for feedback control output from the control object 20, based on, output from the controlled object 20 output signal y _R is to generate a control signal u _R as follows the output signal y _M output from the control model 16 and outputs to the controlled object 20. The output signal y _r for feedback control, not only the output signal y _R, is a signal including observables other than the output signal y _R.

In other words, the control target controller 14 performs feedback control on the control target 20 using the output signal y _M output from the control model 16.

  The control model controller 18 can be configured by a computer including, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and a nonvolatile memory. In this case, a feedback control simulation program that causes the computer to execute a feedback control simulation process described later is stored in, for example, the ROM, and is read and executed by the CPU. The feedback control simulation program may be provided by a recording medium such as a CD-ROM or DVD-ROM.

  Next, feedback control simulation processing executed in the control model controller 18 will be described with reference to the flowchart shown in FIG.

First, in step S100, and inputs a target value r of the output signal _{y M} output from the control model 16.

At step S102, and inputs an output signal y _m for feedback control output from the control model 16 in a state in which the input virtual disturbance dv the control model 16.

In step S104, the target value r entered in step S100, the output signal _{y m} input at step S102, based on, it generates a control signal _{u M} as the output signal _{y M} to follow the target value r Output to the control model 16.

Thus, the control model 16 outputs an output signal y _M corresponding to the control signal u _M to the control target controller 14. Then, the control target controller 14, the output signal y _M output from the control model 16, and the output signal y _r for feedback control output from the control object 20, on the basis, which is output from the control object 20 output signal y _R is to generate a control signal u _R as follows the output signal y _M output from the control model 16 and outputs to the controlled object 20.

Thus, the controlled object 20 is feedback-controlled based on the output signal y _M generated by the feedback control by simulation.

In the control model 16 is only to accurately correspond to the controlled object 20, if the virtual disturbance dv is not, can be accurately follow the actual target value r output signal y _R of the controlled object 20. Further, by giving an appropriate virtual disturbance dv, an output signal y _M that is a trajectory for avoiding the moving object from the obstacle is automatically calculated by the feedback control simulation, and the trajectory of the control object 20 that is the moving object is calculated. by certain output signal y _R to follow this, the moving body can be avoided accurately obstacle. Since the simulation environment is an ideal environment for feedback control, the feedback control simulation is generally stable even if a virtual disturbance dv is given.

  (Second Embodiment)

  Next, a second embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the same part as 1st Embodiment, and the detailed description is abbreviate | omitted.

FIG. 3 shows a block diagram of the feedback control simulation apparatus 12A according to the present embodiment. The feedback control simulation device 12A in FIG. 3 is different from the feedback control simulation device 12 in FIG. 1 in that an output signal output unit 24 for outputting the control signal u _M output from the control model controller 18 to the control target 20 is provided. that it includes, in that the control signal output section 22 for outputting an output signal y _M output from the control model 16 to the control target controller 14 is omitted. Since other configurations are the same as those of the feedback control simulation apparatus 12 of FIG. 1, detailed description thereof is omitted.

Further, since the control signal output section 22 for outputting an output signal y _M to the control target controller 14 are omitted, in the configuration of FIG. 3 have been omitted also controlled object controller 14.

Thus, in the present embodiment, the control signal u _M generated by the control model controller 18 by the feedback control by simulation is output to the controlled object 20. That is, the control target 20 is subjected to feedforward control based on the control signal u _M generated by the control model controller 18.

If the controlled object 20 is an automobile that can be automatically operated for example, the control signal u _M may be, for example, a steering wheel operation amount of the steering wheel to operate the vehicle orientation. In this case, it is not necessary to calculate a trajectory for avoiding an obstacle as in the past, and then to perform a complicated calculation such as determining a handle operation amount for causing the control target to follow the calculated trajectory. The amount of steering wheel operation that can avoid the collision of an automobile with an obstacle by feedback control through simulation can be directly obtained with a small amount of calculation.

  (Third embodiment)

  Next, a third embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the same part as 1st Embodiment and 2nd Embodiment, and the detailed description is abbreviate | omitted.

FIG. 4 shows a block diagram of the feedback control simulation apparatus 12B according to the present embodiment. The feedback control simulation device 12B of FIG. 4 is a device that combines the feedback control simulation device 12 of FIG. 1 and the feedback control simulation device 12A of FIG. That is, feedback control simulation apparatus 12B includes a control signal output section 22 for outputting an output signal y _M output from the control model 16 to the control target controller 14, control signals u which is output from the control model controller 18 _And an output signal output unit 24 for outputting _M to the controlled object 20.

Then, the control signal u _M output from the control model controller 18 and the control signal u _R output from the control target controller 14 are added and output to the control target 20.

Thus, in the present embodiment, the output signal y _M generated by the feedback control by simulation is input to the controlled object controller 14, a control signal u _R which is outputted from the control target controller 14, the controller for the control object since the control signal u _M generated by the 14 control object 20 is controlled by a control signal added, it is possible to accurately track the output signal y _M output signals y _R of the controlled object 20.

  (Fourth embodiment)

  Next, a fourth embodiment of the present invention will be described. In the fourth embodiment, a case where the present invention is applied to an automatic driving control system that considers obstacle avoidance will be described.

  FIG. 5 shows a block diagram of the feedback control simulation apparatus 30 according to the present embodiment. The configuration in FIG. 5 corresponds to the configuration in FIG. 4 described in the third embodiment.

In the following, referred to as "PD" control model controller PD, referred to as control model _{P M} to _{"P M",} referred to as a control target controller PID to "PID", referred to as the control object _{P R} to _{"P R"} .

As shown in FIG. 5, from the target value r outputted from target value decision unit 32, output signal q outputted from the control model P _M ^- (hereinafter, for convenience thus described the superscript bar q) A signal obtained by subtracting is output to the control model PD as a deviation. The control model controller PD and the part for subtracting the output signal q ⁻ from the target value r correspond to the control model controller 18 in FIG.

The control signal output from the control model controller PD is added to the virtual disturbance dv output from the virtual disturbance generation unit 34 and output to the coordinate conversion unit 36. Coordinate converted control signal u by a coordinate transformation unit 36 ^- (hereinafter, for convenience thus described the superscript bar u) is output to the control model P _M.

Control target P _R is controlled by the control signal u. The output signal outputted from the control target P _R, disturbance d is outputted as an output signal q is added. Incidentally, the control target _{P R} corresponds to the controlled object 20 in FIG.

The control target controller PID generates a control signal based on a signal obtained by subtracting the output signal q from the output signal q ⁻ and outputs the control signal to the coordinate conversion unit 38. Coordinate converted control signal at the coordinate converter 38, a control signal outputted from the coordinate converter 36 of the feedback control simulation apparatus 30 u ^- is output is added to the controlled object P _R as a control signal u. Note that the control target controller PID and the part of subtracting the output signal q from the output signal q ⁻ correspond to the control target controller 14 of FIG.

Here, the control object P _R is the real vehicle, control model P _M is a simulation model corresponding to the actual vehicle. Therefore, the target value r is a target value representing the position and speed of the actual vehicle, the control signal u is a control signal for controlling the driving braking force and the steering angle of the actual vehicle, and the output signal q is the position and speed of the actual vehicle. Is an output signal. In the following, simply referred to as a vehicle control model P _M.

  The virtual disturbance generator 34 outputs a virtual disturbance dv for applying a repulsive force from an obstacle to the vehicle traveling on the simulation. The virtual disturbance dv, that is, the repulsive force from the obstacle, can be calculated from a preset potential function. If there are a plurality of obstacles, the sum of the repulsive forces from the respective obstacles is taken. If the potential function is U and the position is q ^, the force F ^ can be expressed as the following equation (1). It should be noted that the bold display symbol in the mathematical expression represents a vector, and in the text, the symbol is represented by “^”.

... (1)

As the control model controller PD, one that is affected by the virtual disturbance dv while following the path is used. The control object P _R is the real vehicle, the control signals calculated in the feedback control simulation apparatus 30 u ^- for control signals based on the u is input, the control target controller PID, disturbance d and control model P _M The one intended to suppress the error is used. FIG. 6 shows a determination method in determining the force applied to the target value and the control model P _M. In order to prevent the vehicle 44 from being delayed from the target position 46 due to the repulsive force from the obstacle 40, that is, the repulsive force 42 (Force of potential) due to the potential function, the target is the distance traveled by the vehicle 44 in the direction along the target route 48. The position is advanced on the target path 48. Then, the target speed set at the target position 46 is set as the target speed of the vehicle 44. Next, a force 50 (course keeping force) for maintaining the target path 48 is calculated by PD control (proportional derivative control) from the target value, that is, the deviation between the target position and target speed and the position and speed of the vehicle 44. Next, a repulsive force 42 from the potential function is applied. The sum of these forces, and converts the combined force added to the coordinates of the control model P _M, and forward the driving force, those lateral forces divided by the cornering power and steering angle, the control of these models and it outputs the P _M and the control object _{P R.}

  Next, an example of overtaking a parked vehicle or a pedestrian will be described.

  When a car is parked at the end of the road, driving over a safe distance is required. In addition, it is necessary to decelerate in consideration of popping out of the shadow of the parked vehicle. Similarly, when there is a pedestrian, it is necessary to perform deceleration in consideration of overtaking and jumping out at a safe distance. Therefore, consider a potential function that applies force in the direction away from the obstacle. The following equation (2) shows a potential function that increases as the distance from the obstacle increases. Here, K is the distance to the obstacle, and A and α are constants.

... (2)

  In the potential function of the above equation (2), a large repulsive force is applied only when the vehicle is present nearby, so a safe distance can be maintained, but when approaching from a distant place, an avoidance action is initiated from an early stage. Is difficult. Therefore, as shown in FIG. 7, the obstacle 40 is formed into a rectangle that is enlarged by the width of the vehicle, and an edge 52 protruding in the direction opposite to the right direction in FIG. 7, that is, the left direction, is attached. FIG. 7 is a view of the road as viewed from above, and the vehicle is moving in the right direction in FIG. 7, and it is assumed that there is an obstacle 40 at a position contacting the vehicle. In this case, the potential field 54 can be drawn with contour lines as shown in FIG. The potential field 54 increases logarithmically as it goes inward. By using such a potential function, a greater repulsive force is applied to the vehicle as the obstacle 40 is approached, but sudden braking is applied as it is. Therefore, as the relative speed with the obstacle 40 increases, the potential function is calculated using the position approaching the obstacle 40 as a temporary position. As a result, the potential field 54 is as shown in FIG. 7, and obstacles can be avoided from an early stage.

Specifically, when there is a vehicle on the protrusion side of the edge 52, a pseudo distance that becomes smaller according to the relative speed with the obstacle 40 is used. The pseudorange is a distance from the obstacle 40 when the position q ^ = [XY] is replaced with the following equation (3). At this time, the traveling direction of the vehicle is the direction of the X axis in FIG. 7, V is the speed of the vehicle, V _ob is the speed of the obstacle, X _ob is the position of the obstacle 40 in the traveling direction of the vehicle, and β is a constant. .

... (3)

  By using the pseudo distance, the value of the potential function in the case where the obstacle 40 is closer to the obstacle 40 is applied as it approaches the obstacle 40 earlier.

In Figure 7, the logarithm above (2), the potential field 54 indicating the potential function of the obstacle 40 when the equation (3), the vehicle is traveling at 40 km / h, from ¹⁰⁵ as the distance from the obstacle 40 This is represented by contour lines that were reduced.

  Next, an example in the case of passing through a no-signal intersection will be described.

Consider the priority side of a no-signal intersection. It is necessary to decelerate before entering the intersection so that it is possible to stop at a no-signal intersection when a pedestrian suddenly jumps out. Therefore, consider adding the traveling direction and opposite force on the control model P _M before entering the intersection by the potential function. The formula of the potential function at the no-signal intersection is shown in the following formula (4). FIG. 8 is a graph showing the following expression (4) when the speed limit is 25 km / h. Here, the proportional gain of speed is P _V , the target speed is Vr, the road width is w, the center of the intersection is Xu, the minimum speed obtained when entering the intersection is Vm, and the speed when passing through Xu is V _{X = Xu} . . In FIG. 8, the intersection center Xu = 0. A is a constant.

... (4)

  This potential function is before the intersection entrance by a distance where the maximum force is proportional to the target speed. In this way, the minimum speed can be set near the intersection entrance. Also in this potential, pseudorange is used so that the vehicle can be decelerated early when the vehicle speed is high.

  Next, a simulation for overtaking a parked vehicle will be described.

  FIG. 9 shows the result of verifying only the simulation portion executed in the feedback control simulation apparatus 30 of FIG. 5 when there is a parked vehicle. As shown in FIG. 9, it can be seen that the trajectory of the vehicle is an avoidance trajectory 58 (Simulation) that avoids a parked vehicle as an obstacle 40 existing on the target trajectory 56 (Reference course). From this, it was confirmed that the overtaking operation is possible by using the feedback control simulation apparatus 30 according to the present embodiment and the potential function as the virtual disturbance.

  Next, a simulation when passing through a no-signal intersection will be described.

  FIG. 10 shows the result of verifying only the simulation portion executed in the feedback control simulation apparatus 30 of FIG. 5 when passing through a no-signal intersection.

  In addition to the simulation result (simulation) when the intersection center Xu = 100, FIG. 10 shows the result of the road speed limit as the target speed (Regulation speed) as a comparison target. The driving data was shown.

  As the driving data, an average value of five data obtained by removing 31 data from data with deceleration considering oncoming vehicles and data with low deceleration is used.

  As shown in FIG. 10, it can be seen that the simulation result (Simulation) reproduces the same deceleration as the actual driving data (Driving data). There is a difference at the time of acceleration. At this time, the force from the potential function is 0, so it is considered that the PD control for maintaining the speed cannot reproduce the acceleration operation at the time of driving. It is done.

  Thus, it has been found that the present invention can be applied to an automatic driving control system that applies a force to a vehicle on a simulation by a potential function. Moreover, the potential function when passing an obstacle or passing through a no-signal intersection was set, and its effectiveness was confirmed by simulation.

  In the present embodiment, a specific example in which the form described in the third embodiment is applied to automatic driving control in consideration of obstacle avoidance has been described. However, in FIG. 5, the control target controller PID and the coordinate conversion unit The control unit 39 including 38 may be omitted. In this case, the configuration described in the second embodiment is applied to automatic operation control.

  (Fifth embodiment)

  Next, a fifth embodiment of the present invention will be described. In the fifth embodiment, a specific example when the present invention is applied to collision avoidance control of a modularized multi-body system will be described.

  First, the concept of the module according to the present embodiment will be described.

  An overview of the concept is shown in FIG. For a hardware module, a dynamic model and a controller for operating the hardware module are modularized on the software and given to the hardware, and the hardware and the software are handled together as one module. This module is defined as Model-Contained Module (MCM). Hereinafter, this module is referred to as MCM.

In FIG. 11, the module of the rotation joint (Rotational joint) is MCM1. MCM1 is for hardware _{H R} of the rotary joint, model _{P R} models the hardware _{H R,} the controller _{C R} which controls the model _{P R} and modularized.

Similarly, MCM2 is for hardware _{H L} of prismatic joint (Linear joint), model _{P L} modeling the hardware _{H L,} the controller _{C L} for controlling the model _{P L} and modularized. By combining MCM1 and MCM2, hardware _H R, together with the hardware is constructed by combining the _{H L,} model _P R, software that combines _{P L} constituted.

  When hardware is configured with MCM, a dynamic model or controller corresponding to the hardware can be obtained by combining modules on software. This simplifies the process of model creation and controller design. In this embodiment, hardware for a rotary joint and a direct acting joint is created for MCM verification, and a dedicated dynamic model is given to handle it as an MCM.

  Next, a model given to the MCM will be described.

  Conventionally, a method of performing control by giving an inverse dynamic model to the MCM is known, but in this embodiment, in order to verify the expandability of the concept of the MCM, a controller using a potential method is given to the MCM. Control. The potential method is a method of designing a path in consideration of collision avoidance by giving a potential function to an obstacle and applying a repulsive force according to the gradient of the potential function in a virtual space. In the present embodiment, two arms are configured by the MCM, and are treated as obstacles. The potential function U given to the MCM is shown in the following equation (5).

... (5)

  In the above equation (5), r is the distance between the MCMs, and a is a constant that adjusts the gradient of the potential field. The gradient of the potential function increases as the distance between the MCMs decreases. The repulsive force f ^ is expressed by the following equation (6), which is the gradient of the above equation (5).

... (6)

Further, by defining A in the above equation (5) as in the following equation (7), the gradient of the potential function is changed according to the relative speed v _r between the MCMs. Here, B is a constant for adjusting the gradient of the potential field, and c is a constant for adjusting the influence of the relative velocity.

... (7)

  As the MCM approaches at high speed, the potential field increases and deceleration begins earlier. In addition, when the two are stationary, the potential field is reduced and the influence is reduced.

  Next, the configuration of the feedback control simulation apparatus according to the present embodiment will be described.

  FIG. 12 shows a block diagram of a feedback control simulation apparatus 70 according to the present embodiment. The configuration in FIG. 12 corresponds to the feedback control simulation apparatus 12 in FIG. 1 described in the first embodiment. Hereinafter, “PD” is referred to as a control model controller PD, and “P” is referred to as a control model P.

A general potential method automatically designs a route by giving an attractive potential to a destination. In many cases, a robot arm is intended to freely follow a trajectory desired by a user. Therefore, it is not compatible with the attractive potential that automatically determines the route. Therefore, in this embodiment, the user generates the trajectory by giving the target joint angle θ _in of the arm himself, and obtains the torque τ necessary for driving the arm with the inverse dynamic model P ⁻¹ and is a control model that is a plant model. Output to P. An output joint angle θ _out of the arm is output from the control model P. Controller PD control model controls the control model P by performing the PD control based on the deviation between the output joint angle theta _out and the target joint angle theta _in. In this way, control is performed in a two-degree-of-freedom system. Further, by adding a virtual disturbance dv obtained by converting the external force f generated from the potential field to the torque τ output from the control model controller PD and applying it to the plant model P, an operation in consideration of collision avoidance is performed. The correction is performed. Here, the control model controller PD and the inverse dynamic model P- ¹ correspond to the control model controller 18 of FIG. The control model P corresponds to the control model 16 of FIG.

The output joint angle θ _out is input to the controller that controls the actual plant, and the form is the same as that of the first embodiment, and the torque τ is input to the actual plant and the form is the same as that of the second embodiment. By combining these forms, a form similar to that of the third embodiment is obtained.

Next, a deviation suppression input for suppressing a deviation between the output joint angle θ _out and the target joint angle θ _in will be described.

For example, when the target trajectory at the tip of the arm is a target trajectory 72 from the start point Start to the end point Goal as shown in FIG. 13, the output joint angle θ _out and the target by the external force F from the potential field 54 of the obstacle 40. In some cases, the deviation from the joint angle θ _in is too large. When this deviation is corrected by the PD control of the control model controller PD, since the deviation is large, an excessive torque τ is input to the control model P, and overshoot may occur.

Therefore, as a method of preventing the deviation from expanding, the advance speed of the target joint angle θ _in is delayed according to the deviation. A polynomial used in the fifth-order polynomial interpolation used for generating the target joint angle θ _in is set as shown in the following equation (8). Here, θ ₀ is an initial angle, θ _f is a final angle, t ₀ is a start time, t _f is an end time, Δt is a sampling time, and d is a constant.

... (8)

Above (8) of the variable t of the target value error e, namely in accordance with the magnitude of the deviation e, change the traveling speed of the target joint angle theta _in the input. As a result, when the deviation e is large, the progress speed of the target joint angle θ _in is slow, and when there is no deviation e, the progress speed of the target joint angle θ _in is the original speed. Thereby, overshoot is suppressed.

  Next, simulation results will be described.

  Two simulations were performed to verify whether or not the overshoot is reduced by performing the collision avoidance operation while decelerating by combining the potential field that changes with the relative speed and the deviation suppression input.

  In the first simulation, the simulation was performed using a potential field that does not change with the coefficient A in the above equation (5) being a constant. In the second simulation, the simulation was performed by combining the potential field that changes the coefficient A in the above equation (7) with the relative speed and the deviation suppression input.

  As shown in FIG. 14, a collision avoidance operation in the case where two arms 76A and 76B in which three rotary joints 74 are combined is crossed is simulated.

  For the rotary joint 74, parameters of an actual machine manufactured for MCM verification were used. The potential field for one arm giving the potential function is as shown in FIG.

  The results of the first and second simulations are shown in FIGS. The arms 76A and 76B operate in order from the smallest number in the figure. However, the description of numbers greater than 7 is omitted. The maximum speed at the tip of the left arm 76B and the shortest distance between the arms 76A and 76B are shown in Table 1 below.

  As shown in FIGS. 16 and 17, both can move from the initial posture to the target posture without colliding. Further, as shown in Table 1 above, the maximum speed during the collision avoidance operation is slower in the second simulation. This is because the potential field is increased by the relative velocity, and the arm receives an external force at an early stage. The shortest distance between the arms in Table 1 is larger in the second simulation. This is due to moderate PD control by the deceleration by the potential field that changes according to the relative speed and the deviation suppression input. Furthermore, since PD control is operating moderately, when attention is paid to the target posture, FIG. 17 shows that overshoot is suppressed compared to FIG. From the above, it was found that the collision avoidance operation can be performed while maintaining a wide interval between the arms and decelerating.

  Thus, it was confirmed that a good collision avoidance operation was possible on the simulation by combining the potential field that changes depending on the relative speed and the deviation suppression input that changes the traveling speed due to the deviation.

  (Sixth embodiment)

  Next, a sixth embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the same part as 1st Embodiment, and the detailed description is abbreviate | omitted.

  FIG. 18 shows a block diagram of a control device 10A according to the present embodiment. The control device 10A in FIG. 18 is different from the control device 10 in FIG. 1 in that the control target controller that controls the control target 20 is configured by an inverse model 14A of the control target. Since other configurations are the same as those of the feedback control simulation apparatus 12 of FIG. 1, detailed description thereof is omitted.

As described above, in the present embodiment, the controlled object inverse model 14 </ b> A controls the controlled object 20. Here, the “inverse model of the control target” is input to the control target so that when the output signal from the control target (for example, the output signal y _R from the control target 20) is input, the output signal is obtained. This is a model that outputs a control signal (for example, a control signal u _R input to the controlled object 20). If there is an accurate inverse model of the control target, in an environment where there is no disturbance or the like, the control amount can be matched with the target value with high accuracy by using it in a feedforward manner. That is, the control device 10A in FIG. 18 is obtained by replacing the portion of the control device 10 in FIG. 1 that actually performs feedback control of the control target 20 with feed-forward control using an inverse model of the control target. Note that, as the inverse model 14A to be controlled, for example, the inverse model of the control model 16 can be used, but is not limited thereto.

  (Seventh embodiment)

  Next, a seventh embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the same part as 2nd Embodiment, and the detailed description is abbreviate | omitted.

  FIG. 19 shows a block diagram of a feedback control simulation apparatus 12C according to the present embodiment.

  As illustrated in FIG. 19, the feedback control simulation apparatus 12 </ b> C according to the present embodiment includes an equivalent virtual disturbance generation unit 60.

The virtual disturbance dv is input to the equivalent virtual disturbance generation unit 60. Then, the equivalent virtual disturbance generation unit 60 generates an equivalent virtual disturbance dv1 based on the input virtual disturbance dv. Here, the equivalent virtual disturbance dv1, in order to obtain a state equivalent to a state of being entered virtual disturbance dv the control model 16, a signal to be added to the control signal u _M output from the control model for the controller 18.

Control signal output unit 24 outputs a control signal u _m obtained by adding an equivalent virtual disturbance dv1 to the control signal u _M output from the control model for the controller 18 to the control model 16 and the controlled object 20.

As described above, in this embodiment, the control signal u _m obtained by adding the equivalent virtual disturbance dv1 to the control signal u _M generated by the control model controller 18 by feedback control by simulation is output to the control target 20. That is, the control object 20 becomes to be feed-forward control based on the control signal u _m.

If the controlled object 20 is an automobile that can be automatically operated for example, the control signal u _m can be, for example, a steering wheel operation amount of the steering wheel to operate the vehicle orientation. In this case, it is not necessary to calculate a trajectory for avoiding an obstacle as in the past, and then to perform a complicated calculation such as determining a handle operation amount for causing the control target to follow the calculated trajectory. The amount of steering wheel operation that can avoid the collision of an automobile with an obstacle by feedback control through simulation can be directly obtained with a small amount of calculation.

  (Eighth embodiment)

  Next, an eighth embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the same part as 3rd Embodiment and 7th Embodiment, and the detailed description is abbreviate | omitted.

FIG. 20 shows a block diagram of a control device 10B according to the present embodiment. A control device 10B of FIG. 20 has a configuration in which a feedback control simulation device 12C of FIG. 19 is provided instead of the feedback control simulation device 12B of the control device 10 shown in FIG. The output signal y _M output from the control model 16 is output from the control signal output section 22 to the control target controller 14.

Control target controller 14, the output signal y _M output from the control model 16, and the output signal y _r as a third output signal output from the controlled object 20, on the basis, generating a control signal u _R And output.

The control target 20 includes a control signal u _R output from the control target controller 14 and a control signal u _m obtained by adding the equivalent virtual disturbance dv1 to the control signal u _M output from the control model controller 18. The controlled signal is input.

As described above, in the present embodiment, the control signal u _m obtained by adding the equivalent virtual disturbance dv1 to the control signal u _M generated by the control model controller 18 by feedback control by simulation is output from the control target controller 14. is output to the controlled object 20 the control signal is added to u _R. This makes it possible to accurately track the output signal y _M output signals y _R of the controlled object 20.

  (Ninth embodiment)

  Next, a ninth embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the same part as 8th Embodiment, and the detailed description is abbreviate | omitted.

  FIG. 21 shows a block diagram of a control device 10C according to the present embodiment. As illustrated in FIG. 21, the control device 10 </ b> C includes an equivalent virtual disturbance generation unit 60 and a control target controller 14.

Thus, in this embodiment, the control signal equivalent virtual disturbance dv1 output from the equivalent virtual disturbance generator 60 to the control signal u _R which is outputted from the control target controller 14 is added is outputted to the controlled object 20 The Thereby, the controlled object 20 can be controlled with high accuracy.

  As mentioned above, although embodiment of this invention was explained in full detail with reference to drawings, the specific structure is not restricted to this embodiment, The design change etc. of the range which does not deviate from the summary of this invention are included.

DESCRIPTION OF SYMBOLS 10 Control apparatus 12, 12A, 12B, 30, 70 Feedback control simulation apparatus 14 Control object controller 16 Control model 18 Control model controller 20 Control object 22 Control signal output part 24 Output signal output part 70 Feedback control simulation apparatus

Claims (13)

A control model corresponding to the controlled object;
Based on the target value of the first output signal output from the control model and the second output signal output from the control model in a state where a virtual disturbance is input to the control model, the control model A controller for a control model that generates a control signal such that the first output signal output from the target value becomes the target value and outputs the control signal to the control model;
A feedback control simulation apparatus comprising: The feedback control simulation apparatus according to claim 1, further comprising: a control signal output unit configured to output the control signal output from the control model controller to the control target. An equivalent virtual disturbance generation unit that generates an equivalent virtual disturbance that is a signal to be added to the control signal in order to obtain a state equivalent to a state in which the virtual disturbance is input to the control model based on the virtual disturbance;
The feedback control simulation apparatus according to claim 2, wherein the control signal output unit outputs a control signal obtained by adding the control signal and the equivalent virtual disturbance to the control target. The output signal output part for outputting the said 1st output signal output from the said control model to the control object controller which controls the said control object is given in any one of Claims 1-3 Feedback control simulation device. The control object is a moving object, the target value is a target trajectory of the moving object, and the virtual disturbance is a disturbance that virtually imagines an obstacle that becomes an obstacle when the moving object moves. The feedback control simulation apparatus according to claim 1, wherein the output signal is an obstacle avoidance trajectory that avoids the obstacle. The feedback control simulation apparatus according to claim 5, wherein the virtual disturbance is a disturbance obtained from a potential function. The feedback control simulation apparatus according to claim 6, wherein the moving body is a vehicle, and the potential function is a function that applies a repulsive force to the vehicle. The feedback control simulation apparatus according to claim 7, wherein the potential function is a function set so that the repulsive force is maximized at an entrance of an intersection where the vehicle enters. The feedback control simulation apparatus according to claim 6, wherein the moving body includes two arms, and the potential function is set by regarding each arm as the obstacle. A feedback control simulation apparatus according to claim 4,
A control target controller for controlling the control target based on the first output signal output from the control model of the feedback control simulation device;
A control device comprising: The control device according to claim 10, wherein the control target controller is an inverse model of the control target. Computer
Based on the target value of the first output signal output from the control model corresponding to the control target, and the second output signal output from the control model in a state where a virtual disturbance is input to the control model. A feedback control simulation method comprising: generating a control signal such that the first output signal output from the control model becomes the target value and outputting the control signal to the control model. Computer
The feedback control simulation program for functioning as each means which comprises the feedback control simulation apparatus of any one of Claims 1-9.