Control device, electric power steering device and control method

Abstract

Control device (100) which controls as a control target (560) at least a section with a motor (543) in a steering mechanism (530) which has an input shaft (524a) to which a steering wheel (521) is connected which is steered by a steering person, an output shaft (524b) which is connected to the input shaft (524a) via a torsion bar (546) and the motor (543) which is connected to the output shaft (524b), wherein the control device (100) has the following features: a reaction force control unit (210) which provides an input torque (T r ), which is entered into the control target (560), based on a torsion bar torque (T h ) generated in the torsion bar (546) and controls a reaction force that is transmitted from the steering wheel (521) to the steering person; a support control unit (230) which provides a correction torque (T f ) to correct the input torque (Tr ) generated based on an output of the tax target (560) and a nominal model; and a state feedback unit (280) which provides a state compensation value (V) s ) based on the output of the control target (560) to the input torque (T r ) traces back to, characterized in that the support control unit (230) is designed such that a transfer function P(s) of the control target (560) is limited by a transfer function P n (s) of the nominal model in a frequency band in which a gain for a gain property of a complementary sensitivity function with respect to a modeling error (Δ(s)) between the control target (560) and the nominal model is approximately 1, wherein the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit (230), wherein the support control unit (230) is designed such that the control target (560) is controlled such that the transfer function P(s) of the control target (560) appears as a transfer function P n (s) of the nominal model appears to be when considering the input / output relationship, where the transfer function P n (s) of the predetermined nominal model is given by the following formula (3) P n ( s ) = 1 JSTG ns + BSTG ns 2 + 2 ζ 1 n ω 1 ns + ω 1 n 2 s 2 + 2 ζ 2 n ω 2 ns + ω 2 n 2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function Pn (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is, and the state feedback unit (280) the state compensation value (V s ) so based on the output of the control target (560) such that an apparent transfer function P(s) of the control target (560) corresponds to a transfer function P n (s) approximates the nominal model, where the apparent transfer function P(s) of the control target (560) is a transfer function of an entire section from a subtractor (SU1) to an output of the control target (560) and is a transfer function of a section combining the state feedback unit (280) and the control target (560).

Description

[0001] The present invention relates to a control device according to the preamble of claim 1 or 4, an electric power steering device with such a control device and a control method according to the preamble of claim 15 or 16.

[0002] An electric power steering system mounted on a vehicle is known. For example, an electric power steering system described in JP 2018-183046A includes a motor control device with a disturbance monitor that estimates a disturbance torque.

[0003] With the electric power steering system described above, it is necessary to improve the steering feel experienced by the driver. However, in the control system of the electric power steering system, all elements related to stability, noise suppression, and responsiveness are in a compromise relationship with each other. Therefore, it is difficult to adjust each element individually, and thus it is challenging to improve the steering feel experienced by the driver.

[0004] DE 603 ​​20 115 T2 discloses an electric power steering device in which a steering assistance force, generated by a motor, is transmitted to a control system of a car or vehicle. The electric power steering device disclosed in DE 603 ​​20 115 T2 for controlling a motor is operated on the basis of a current command value calculated from a steering assistance command value, which in turn is calculated based on a steering torque generated in a steering shaft and a current sensing value of the motor that provides a steering assistance force to a steering mechanism. To simultaneously achieve ideal road information sensitivity and ideal steering feel in the frequency domain, the electric power steering system includes a torque filter for processing a torque signal, a SAT estimation function, and a SAT filter for processing SAT information from the SAT estimation function.Furthermore, the electric power steering includes a control system with two degrees of freedom that can independently control the frequency characteristics of the steering feel and the road information sensitivity.

[0005] JP H08 - 310 417 A discloses a control device for an electric power steering system.

[0006] DE 695 33 576 T2 discloses an electric power steering system in which a steering support torque is generated by a motor.

[0007] In view of the above circumstances, an object of the present invention is to provide a control device capable of improving the steering feel experienced by a steering person, an electric power steering device with such a control device, and a control method capable of improving the steering feel experienced by the steering person.

[0008] This problem is solved by a control device according to claim 1 or 4, an electric power steering device according to claim 14 or a control method according to claim 15 or 16.

[0009] A control device according to the invention is a control device which, as a control target, controls at least a section with a motor in a steering mechanism, which has an input shaft to which a steering wheel is connected, which is steered by a steering person, an output shaft which is connected to the input shaft via a torsion bar, and the motor which is connected to the output shaft, wherein the control device comprises: a reaction force control unit which generates an input torque that is input into the control target based on a torsion bar torque that is generated in the torsion bar, and controls a reaction force that is transmitted from the steering wheel to the steering person; an assistance control unit which generates a correction torque to correct the input torque based on an output of the control target and a nominal model;and a state feedback unit that feeds a state compensation value based on the output of the control target back to the input torque. A support control unit is configured such that the transfer function of the control target is limited by the transfer function of the nominal model in the frequency band where the gain of the gain property of the complementary sensitivity function with respect to the modeling error between the control target and the nominal model is approximately 1, where the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit. The support control unit is configured such that the control target is controlled such that the transfer function, P(s), of the control target appears as a transfer function, P; n(s) of the nominal model appears to be when considering the input / output relationship, where the transfer function P n (s) of the predetermined nominal model is given by the following formula Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n(s) is. The state feedback unit feeds back the state compensation value based on the output of the control target such that an apparent transfer function of the control target approximates the transfer function of the nominal model, where the apparent transfer function of the control target is a transfer function of an entire section from a subtractor to an output of the control target and is a transfer function of a section that combines the state feedback unit and the control target.

[0010] A control device according to the invention is a control device which, as a control target, controls at least a section with a motor in a steering mechanism, which has an input shaft to which a steering wheel is connected, which is steered by a steering person, an output shaft which is connected to the input shaft via a torsion bar, and the motor which is connected to the output shaft, wherein the control device comprises: a reaction force control unit which generates an input torque that is applied to the control target based on a torsion bar torque that is generated in the torsion bar, and controls a reaction force that is transmitted from the steering wheel to the steering person; an assistance control unit which generates a correction torque to correct the input torque based on an output of the control target and a nominal model; and a state feedback unit.which feeds a state compensation value back to the input torque, wherein the state compensation value compensates at least part of an inertial force generated in the control target, a viscosity force generated in the control target, and a friction force generated in the control target. A support control unit is configured such that the transfer function of the control target is limited by the transfer function of the nominal model in the frequency band in which the gain of the gain property of the complementary sensitivity function with respect to the modeling error between the control target and the nominal model is approximately 1, wherein the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit. The support control unit is configured such that the control target is controlled in such a way thatthat the transfer function, P(s), of the control target appears as a transfer function, P, n (s) of the nominal model appears to be when considering the input / output relationship, where the apparent transfer function of the control target is a transfer function of an entire section from a subtractor to an output of the control target and is a transfer function of a section combining the state feedback unit and the control target, where the transfer function P n (s) of the predetermined nominal model is given by the following formula Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1na frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is. The state feedback unit feeds the state compensation value back to the input torque after the state compensation value has been corrected by the correction torque and before the state compensation value is entered into the control target.

[0011] An electric steering device according to the invention comprises the control device and the steering mechanism.

[0012] A control method according to the invention is a control method for steering, as a control target, at least one section with a motor in a steering mechanism, which has an input shaft to which a steering wheel is connected, which is steered by a steering person, an output shaft which is connected to the input shaft via a torsion bar, and the motor which is connected to the output shaft, wherein the control method comprises the following steps: generating an input torque which is fed into the control target, based on a torsion bar torque which is generated in the torsion bar, and controlling a reaction force which is transmitted from the steering wheel to the steering person; generating a correction torque to correct the input torque based on an output of the control target and a nominal model; and feeding back a state compensation value to the input torque based on the output of the control target.Generating the correction torque involves constraining a transfer function of the control target by a transfer function of the nominal model in a frequency band where the gain of a complementary sensitivity function with respect to a modeling error between the control target and the nominal model is approximately 1, where the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit. The support control unit is configured such that the control target is controlled such that the transfer function, P(s), of the control target appears as a transfer function, P. n (s) of the nominal model appears to be when considering the input / output relationship, where the transfer function P n (s) of the predetermined nominal model is given by the following formula Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n(s) is. Feeding the state compensation value back to the input torque involves feeding back the state compensation value based on an output of the control target in such a way that an apparent transfer function of the control target approximates a transfer function of the nominal model, where the apparent transfer function of the control target is a transfer function of an entire section from a subtractor to an output of the control target and is a transfer function of a section combining the state feedback unit and the control target.

[0013] A control method according to the invention is a control method for steering, as a control target, at least one section with a motor in a steering mechanism, which has an input shaft to which a steering wheel is connected, which is steered by a steering person, an output shaft which is connected to the input shaft via a torsion bar, and the motor which is connected to the output shaft, wherein the control method comprises the following steps: generating an input torque which is fed into the control target, based on a torsion bar torque which is generated in the torsion bar, and controlling a reaction force which is transmitted from the steering wheel to the steering person; generating a correction torque to correct the input torque based on an output of the control target and a nominal model; and feeding back to the input torque.a state compensation value to compensate for at least part of an inertial force, a viscosity force, and a friction force generated in the control target. Generating the correction torque involves constraining a transfer function of the control target by a transfer function of the nominal model in a frequency band where the gain of a complementary sensitivity function with respect to a modeling error between the control target and the nominal model is approximately 1, where the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit. The support control unit is configured such that the control target is controlled such that the transfer function, P(s), of the control target appears as a transfer functionP, n (s) of the nominal model appears to be when considering the input / output relationship, where the apparent transfer function of the control target is a transfer function of an entire section from a subtractor to an output of the control target and is a transfer function of a section combining the state feedback unit and the control target, where the transfer function P n (s) of the predetermined nominal model is given by the following formula Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function Pn (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is.. Feeding the state compensation value back to the input torque includes feeding the state compensation value back to the input torque after the input torque has been corrected by the correction torque and before the input torque is fed into the control target.

[0014] According to one aspect of the present invention, it is possible to improve the steering feel experienced by a person steering the steering wheel of a vehicle equipped with an electric power steering device.

[0015] Preferred embodiments of the present invention are explained in more detail below with reference to the accompanying drawings. These show: Fig. 1 a diagram schematically representing an electric power steering device according to an exemplary embodiment; Fig. 2 a block diagram illustrating a design of a control device according to an exemplary embodiment; Fig. 3 a functional block diagram illustrating the functions of a processor in the control device according to an exemplary embodiment; Fig. 4 a graph representing a gain property of a complementary sensitivity function and a gain property of an inverse of a modeling error between a transfer function of a control target and a transfer function of a nominal model; Fig. 5 a graph that represents an example of measurement results of a steering angle and a torsional torque in a case where no model-following control is applied; Fig. 6 a graph that represents an example of measurement results of a steering angle and a torsional torque in a case where a model-following control is applied; Fig. 7 a graph showing an example of measurement results of a steering angle and a torsional torque in a case where no model-following control is applied; Fig. 8 a graph that provides another example of measurement results of a steering angle and a torsional torque in a case where a model-following control is applied; Fig. 9 a graph which provides yet another example of measurement results of a steering angle and a torsional torque in a case where no model-following control is applied; Fig. 10 a graph which provides yet another example of measurement results of a steering angle and a torsional torque in a case where a model-following control is applied; Fig. 11 a graph which provides yet another example of measurement results of a steering angle and a torsional torque in a case where no model-following control is applied; and Fig. 12 a graph which presents yet another example of measurement results of a steering angle and a torsional torque in a case in which a model-following control is applied.

[0016] The following describes exemplary embodiments of a control device, an electric power steering device, and a control method according to the present disclosure with reference to the accompanying drawings. However, unnecessarily detailed descriptions may be omitted. Detailed descriptions of known circumstances and duplicate descriptions of essentially identical designs, for example, may be omitted. This is intended to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art.

[0017] The following embodiments are for illustrative purposes only, and the control device and control method for an electric power steering device according to the present disclosure are not limited to the following embodiments. For example, numerical values, steps, the sequence of steps, and the like shown in the following embodiments are merely illustrative, and various modifications can be made, unless this results in technical consistency. The embodiments or examples described below are merely exemplary, and various combinations are possible as long as no technical contradiction arises.

[0018] An electric power steering device 1000 of the present embodiment, which is in Fig. Item 1, shown, is mounted on a vehicle. As shown in Fig. As shown in Figure 1, the electric power steering device 1000 comprises a steering mechanism 530 and a control device 100. The steering mechanism 530 comprises a steering mechanism unit 520 and an auxiliary mechanism unit 540. The electric power steering device 1000 controls the auxiliary mechanism unit 540 via the control device 100 to generate an auxiliary torque that reduces the steering torque T. h The auxiliary torque generated in the steering mechanism unit 520 is provided when the driver, operating the vehicle, steers the steering wheel 521. This auxiliary torque reduces the load on the driver's operation when the driver operates the steering wheel 521. The driver of the vehicle is a person steering the vehicle's steering wheel 521.

[0019] The steering mechanism unit 520 comprises a steering wheel 521, a steering shaft 522, universal joints 523A and 523B, an input shaft 524a, an output shaft 524b, a rack and pinion mechanism 525, a rack and pinion shaft 526, a left and a right ball joint 552A and 552B, tie rods 527A and 527B, hinge joints 528A and 528B and a right and a left steered wheel 529A and 529B. This means that the steering mechanism 530 includes the steering wheel 521, the steering shaft 522, the universal joints 523A and 523B, the input shaft 524a, the output shaft 524b, the rack and pinion mechanism 525, the rack and pinion shaft 526, the right and left ball joints 552A and 552B, the tie rods 527A and 527B, the hinge joints 528A and 528B, and the right and left steered wheels 529A and 529B.

[0020] The steering shaft 522 is a shaft extending from the steering wheel 521, which is steered by a person. An end section of the input shaft 524a is connected to an end section of the steering shaft 522 on a side opposite the side connected to the steering wheel 521 via the universal joints 523A and 523B. Consequently, the steering wheel 521 is connected to the input shaft 524a via the universal joints 523A and 523B and the steering shaft 522. The output shaft 524b is connected to the input shaft 524a via a torsion bar 546, which will be described later. In particular, one end section of the output shaft 524b is connected to the other end section of the input shaft 524a via the torsion bar 546. The other end section of the output shaft 524b is connected to the rack shaft 526 via the rack and pinion mechanism 525.

[0021] The input shaft 524a and the output shaft 524b are arranged coaxially. The input shaft 524a and the output shaft 524b are rotatable about the same central axis. The input shaft 524a and the output shaft 524b are rotatable relative to each other within a range in which the torsion bar 546, which will be described later, can be rotated.

[0022] The auxiliary mechanism unit 540 comprises a steering torque sensor 541, a steering angle sensor 542, a motor 543, a deceleration mechanism 544, an inverter 545, and a torsion bar 546. This means that the steering mechanism 530 comprises the steering torque sensor 541, the steering angle sensor 542, the motor 543, the deceleration mechanism 544, the inverter 545, and the torsion bar 546. The torsion bar 546 comprises the input shaft 524a and the output shaft 524b. The torsion bar 546 is arranged coaxially with the input shaft 524a and the output shaft 524b. In the following description, a virtual axis passing through a common central axis of the input shaft 524a, the output shaft 524b, and the torsion bar 546 is referred to as the axis of rotation R. The torsion bar 546 can be rotated around the axis of rotation R.

[0023] The steering torque sensor 541 detects a steering torque T hin the steering mechanism unit 520 by detecting a quantity of torsion about the rotation axis R of the rotating rod 546. The steering torque T h The torque is a torsional bar torque generated in the torsional bar 546 and is a torsional moment about the axis of rotation R. The steering angle sensor 542 can detect a rotation angle θa about the axis of rotation R of the input shaft 524a. The rotation angle θa of the input shaft 524a is equal to the steering angle of the steering wheel 521. This means that the steering angle sensor 542 can detect the steering angle of the steering wheel 521 by detecting the rotation angle θa of the input shaft 524a. A rotation angle θb of the output shaft 524b can be detected based on the steering torque sensor 541 and the steering angle sensor 542.

[0024] For example, the inverter 545 converts direct current (DC) power into three-phase alternating current (AC) power with U-phase, V-phase, and W-phase pseudo-sine waves according to the motor driver signal input from the control device 100 and supplies the power to the motor 543. The motor 543 is connected to the output shaft 524b via the deceleration mechanism 544. Three-phase AC power is supplied to the motor 543 by the inverter 545. The motor 543 is, for example, an internal permanent magnet synchronous motor (IPMSM), a surface permanent magnet synchronous motor (SPMSM), or a switched reluctance motor (SRM). When three-phase AC power is supplied by the inverter 545, the motor 543 generates an auxiliary torque according to the steering torque T. h The motor 543 transmits the generated auxiliary torque to the output shaft 524b via the deceleration mechanism 544.

[0025] The control device 100 controls a control target 560 with at least the input shaft 524a, the output shaft 524b, and the motor 543 in the steering mechanism 530. In the present embodiment, the control target 560 comprises the steering wheel 521, the universal joints 523A and 523B, the input shaft 524a, the output shaft 524b, the torsion bar 546, the motor 543, and the deceleration mechanism 544. Since the control target 560 comprises the input shaft 524a and the output shaft 524b, which can rotate relative to each other above the torsion bar 546, the movement of the control target 560 cannot be described solely by a simple equation of motion of the inertial system. The steering target 560 changes between the inertia-one system and the inertia-two system depending on the force with which the driver grips the steering wheel 521. The more forcefully the driver grips the steering wheel 521, the closer the steering target 560 is to the inertia-one system.The weaker the steering person grips the steering wheel 521, the closer the steering target 560 is to the inertial-two system.

[0026] The control device 100 is electrically connected to the inverter 545. The control device 100 generates a motor driver signal based on the sensing signals acquired by the steering torque sensor 541, the steering angle sensor 542, a vehicle speed sensor 300 mounted on the vehicle, and the like, in order to output the motor driver signal to the inverter 545. The control device 100 controls the control target 560 by controlling the rotation of the motor 543 via the inverter 545. In particular, the control device 100 controls the switching operation of the plurality of switching elements included in the inverter 545. Specifically, the control device 100 generates a control signal to control the switching operation of each switching element and outputs the control signal to the inverter 545. Each switching element is, for example, a MOSFET.In the following description, a control signal for controlling the switching operation of each switching element is called a "gate control signal".

[0027] The control device 100 generates a torque command value based on the steering torque T. hand the like, and controls the torque and rotational speed of the motor 543 by means of, for example, vector control. Vector control is a method in which the current flowing through the motor 543 is separated into a current component that contributes to generating torque and a current component that contributes to generating magnetic flux, with the current components, which are orthogonal to each other, being controlled independently. The control device 100 can perform not only vector control but also other closed-loop control. The rotational speed of the motor 543 is represented, for example, by a rotational speed [rpm] at which the rotor rotates in one minute, a rotational speed [rpm] at which the rotor rotates in one second, or the like.

[0028] The value of the steering torque T hThe steering torque sensor 541 can be directly input into the control device 100, or the control device 100 can receive the steering torque value T. h The steering angle value of the steering wheel 521 can be calculated from the output value of the steering torque sensor 541. The steering angle value of the steering wheel 521 can be directly input from the steering angle sensor 542 into the control device 100, or the control device 100 can calculate the steering angle value from the output value of the steering angle sensor 542.

[0029] The control unit 100 and the motor 543 are modularized and manufactured and sold as a single motor module. This motor module comprises the motor 543 and the control unit 100 and is suitable for use with the electric power steering system 1000. The control unit 100 can also be manufactured and sold independently of the motor 543 as a separate control unit for the electric power steering system 1000.

[0030] Fig. Figure 2 represents a typical example of the design of the control device 100 according to the present embodiment. The control device 100 comprises, for example, a power source circuit 111, an angle sensor 112, an input circuit 113, a communication interface or I / F 114, a driver circuit 115, a ROM 116, and a processor 200. The control device 100 can be implemented as a printed circuit board (PCB) on which these electronic components are implemented.

[0031] In the Processor 200, the vehicle speed sensor 300, the steering torque sensor 541, and the steering angle sensor 542, which are mounted on the vehicle, are connected to the Processor 200 via communication. The vehicle speed is transmitted from the vehicle speed sensor 300 to the Processor 200. The steering torque T hThe steering torque sensor 541 transmits the steering angle to the processor 200. The steering angle sensor 542 transmits the steering angle to the processor 200.

[0032] The Processor 200 is an integrated semiconductor circuit, also known as a central processing unit (CPU) or microprocessor. The Processor 200 sequentially executes computer programs stored in ROM 116, which describe instructions for controlling motor drives and implement the desired processing. In addition to or instead of the Processor 200, the Control Device 100 can include a Free-Programmable Gate Array (FPGA) equipped with a CPU, a Graphics Processing Unit (GPU), an Application-Specific Integrated Circuit (ASIC), an Application-Specific Standard Product (ASSP), or a combination of two or more circuits selected from these.The processor 200 sets a current command value according to the actual current value, the rotation angle of the rotor and the like of the motor 543, generates a pulse width modulation (PWM) signal and outputs the PWM signal to the driver circuit 115.

[0033] The power source circuit 111 is connected to an external power source (not shown). The power source circuit 111 generates a DC voltage required for each unit of the control device 100. The DC voltage generated in the power source circuit 111 is, for example, 3V or 5V.

[0034] The angle sensor 112 detects the rotation angle of the rotor in the motor 543 and outputs this angle to the processor 200. The angle sensor 112 can be a coordinate converter or resolver, a Hall element (such as a Hall IC), or an MR sensor with a magnetoresistive element. The processor 200 can calculate the angular velocity ω [RAD / s] of the motor 543 based on the electrical angle θm of the motor 543, which is obtained from the angle sensor 112. The control device 100 can include a velocity sensor capable of detecting the rotational angular velocity of the motor 543, and an acceleration sensor capable of detecting the rotational angular acceleration of the motor 543, instead of the angle sensor 112.

[0035] A motor current value, detected by a current sensor (not shown), is input to the input circuit 113. In the following description, a motor current value detected by a current sensor (not shown) is referred to as the "actual current value". The input circuit 113 converts the level of the input actual current value, as needed, into an input level for the processor 200 and outputs the actual current value to the processor 200. A typical example of the input circuit 113 is an analog-to-digital conversion circuit.

[0036] The communication I / F 114 is an input / output interface for sending / receiving data, for example in accordance with a vehicle's internal control unit network (CAN).

[0037] The driver circuit 115 is typically a gate driver or a pre-driver. The driver circuit 115 generates a gate control signal according to the PWM signal and outputs this gate control signal to the gates of the majority of switching elements included in the inverter 545. For example, if the motor 543 to be driven is a low-voltage motor, the driver circuit 115 as a gate driver is not strictly necessary in some cases. In this case, the gate driver function of the driver circuit 115 can be implemented in the processor 200.

[0038] ROM 116 is electrically connected to processor 200. ROM 116 can be, for example, writable memory, rewritable memory, or read-only memory. Examples of writable memory include programmable read-only memory (PROM). Examples of rewritable memory include flash memory and electrically erasable programmable read-only memory (EEPROM). ROM 116 stores a control program containing instructions for causing processor 200 to control the motor drive. For example, the control program stored in ROM 116 is initially developed once in RAM (not shown) at the time of initial loading.

[0039] Fig. Figure 3 represents an example of functional blocks of the processor 200 according to the present embodiment. The processor 200, which is a computer, sequentially performs processing or tasks necessary for controlling the motor 543 using each functional block. Each functional block of the processor 200, which is in Fig. The function shown in Figure 3 can be implemented in the processor 200 as software, such as firmware, as hardware, or as a combination of both. The processing of each functional block in the processor 200 is typically described in a computer program in units of software modules and stored in ROM 116. However, if an FPGA or similar device is used, all or some of the functional blocks can be implemented as hardware accelerators. The control method of the control device 100 according to the present embodiment is implemented in a computer and can be implemented by causing the computer to perform a desired operation.

[0040] The processor 200 comprises a reaction force control unit 210, a support control unit 230, a state feedback unit 280, a subtractor SU1, an adder AD1, and an adder AD2. This means that the control device 100 is equipped with the reaction force control unit 210, the support control unit 230, the state feedback unit 280, the subtractor SU1, the adder AD1, and the adder AD2. In other words, functions corresponding to the reaction force control unit 210, the support control unit 230, the state feedback unit 280, the subtractor SU1, the adder AD1, and the adder AD2 are implemented in the processor 200 of the control device 100.

[0041] The steering torque T h The torque detected by the steering torque sensor 541 is input into the reaction force control unit 210. The reaction force control unit 210 generates the input torque T. r, which is entered into the control target 560, based on the steering torque T h , that is, the torsion bar torque generated in the torsion bar 546. The input torque T r is a target torque of the motor 543 and is a torque command value. The reaction force control unit 210 generates the input torque T. r and controls the torque of the motor 543 to control the reaction force transmitted from the steering wheel 521 to the steering person. The reaction force control unit 210 generates the input torque T r by applying phase compensation to the steering torque T h, if the steering frequency or steering speed is within a predetermined range. The steering frequency is a frequency of the steering angle that changes based on the driver's operation of the steering wheel 521. The steering speed is a speed of the steering wheel that changes based on the driver's operation of the steering wheel 521. The reaction force control unit 210, which is in Fig. Figure 3 shows a basic support calculation unit 211 and a phase compensator 212.

[0042] The basic support calculation unit 211 records the steering torque T h and the vehicle speed. The basic support calculation unit 211 generates a basic support torque based on the steering torque T. hand the vehicle speed. The basic support calculation unit 211, for example, includes a lookup table (LUT) in which a relationship between the steering torque T is defined. h , which is defined by the vehicle speed and the basic support torque. The basic support calculation unit 211 can calculate the basic support torque using a correspondence relationship based on the steering torque T. h and the vehicle speed with reference to the reference table. The basic support calculation unit 211 can determine a basic support gain based on a gradient defined by a ratio of a change quantity of the basic support torque to a fluctuation quantity of the steering torque T. h .

[0043] In the present embodiment, the phase compensator 212 adjusts the assistance gain within a range of the steering frequency when the driver operates the steering wheel 521 and compensates for the rigidity of the torsion bar 546. The range that the steering frequency can assume is, for example, 5 Hz or less. The phase compensator 212 can, for example, provide first-order phase compensation to the steering torque T. h Apply the torsion bar torque when the steering frequency is 5 Hz or less. First-order phase compensation is expressed, for example, by a transfer function of formula (1). C(s)=12πf1s+112πf2s+1

[0044] In formula (1), s is a Laplace transformer, f1 is a frequency (Hz) used to determine the zero point of the transfer function, and f2 is a frequency (Hz) used to determine the pole of the transfer function. A graph where the gain (or loop gain) is set as a vertical axis and the logarithm of the frequency is set as a horizontal axis is called a gain diagram. In the gain diagram, the zero point is the intersection of the gain curve and the horizontal axis at 0 dB, and the pole is the maximum point of the gain curve. For example, by setting the pole frequency to a point higher than the zero point frequency, phase advance compensation can be applied. The greater the difference between the pole frequency and the zero point frequency, the greater the amount of phase advance.

[0045] The phase compensator 212 generates the input torque T rbased on the basic support torque and the basic support gain output from the basic support calculation unit 211. The phase compensator 212 can, for example, be a stabilization compensator and apply a stability phase compensation to the basic support torque. The phase compensator 212 can have a second-order or higher transfer function whose frequency property is variable according to the basic support gain. The second-order or higher transfer function is expressed using a response parameter and an attenuation parameter. The second-order or higher transfer function can be expressed, for example, by formula (2). By setting the order of the transfer function to two, attenuation can be imparted to the transfer function property. A phase property can be adjusted by changing the attenuation. C(s)=s2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2(ω22ω12)

[0046] In formula (2), s is a Laplace transformer, ω1 is a zero-point frequency of the transfer function, ω2 is a pole frequency of the transfer function, ζ1 is a zero-point attenuation ratio, and ζ2 is a pole attenuation ratio. The pole frequency ω2 is lower than the zero-point frequency ω1.

[0047] The support control unit 230 generates a correction torque T f to correct the input torque T r based on the output of the control target 560 and the nominal model. In the present embodiment, the correction torque T f a feedback torque that is applied to the input torque T r is returned.

[0048] The nominal model is an internal model used to constrain the control target 560 when the control target 560 is being controlled. The nominal model will be described in detail later. In the present embodiment, the support control unit 230 is a model-following controller configured to perform model-following control. A specific configuration of the support control unit 230 will be described in detail later.

[0049] The subtractor SU1 subtracts the correction torque T. f , which is output from the support control unit 230, from the input torque T rThe output from subtractor SU1 is fed into adder AD1 and support control unit 230. Adder AD1 outputs a value obtained by adding the output from state feedback unit 230 to the output from subtractor SU1, which is then fed to adder AD2. Adder AD2 outputs a value obtained by adding a disturbance torque T. d to the output from the adder AD1, to the control target 560.

[0050] The disturbance torque T d is a difference between the actual output torque of motor 543 and the ideal output torque of motor 543. The disturbance torque T d This includes a disturbance torque that is applied externally to the control target 560. The disturbance torque T dThis includes, for example, an additional torque generated by friction and chatter due to mechanical elements, such as the motor 543 and the deceleration mechanism 544; torque ripple generated in the motor 543; a restoring torque; and a disturbance torque that could be generated when driving on an unpaved, uneven path, a gravel road, or the like. Here, the restoring torque means a torque acting in the direction in which the steering wheel 521 returns due to the elasticity of the tire, which is twisted when the steering wheel 521 is turned.

[0051] In the present embodiment, the support control unit 230 generates the correction torque T f based on an angular velocity w θ , which is calculated from the rotation angle θa of the input shaft 524a, and introduces the correction torque T f to the input torque T rback. The angular velocity ω θ is a value corresponding to the angular velocity of the motor 543, which is theoretically calculated from the rotation angle θa of the input shaft 524a. For example, immediately after the person steering begins to turn the steering wheel 521, the input shaft 524a may rotate with the rotation of the steering wheel 521, but the motor 543 may not yet be driven and the output shaft 524b may not yet be rotating. In this case, the actual angular velocity ω of the motor 543 is 0; however, theoretically, if the input shaft 524a rotates, the motor 543 will also rotate, and the output shaft 524b will also rotate. The angular velocity ω θ is a value that corresponds to the angular velocity of motor 543 in a case where motor 543 is theoretically rotated as described above. Therefore, the angular velocity ω can θdiffers from the actual angular velocity ω of motor 543. The rotation angle θa, which is used to calculate the angular velocity ω θ The value used could be a value detected by the steering angle sensor 542, or a value calculated from the rotation angle θb of the output shaft 524b.

[0052] The support control unit 230 comprises an inverse nominal model 231, a low-pass filter 232, a high-pass filter 233, a support matching unit 270, a subtractor SU2, and an adder AD3. The high-pass filter 233 has a first cutoff frequency Cf1. The first cutoff frequency Cf1 ranges, for example, from 2 Hz inclusive to 10 Hz, preferably from 5 Hz inclusive to 7 Hz.

[0053] The low-pass filter 232 has a second cutoff frequency Cf2 that is higher than the first cutoff frequency Cf1. For example, the second cutoff frequency Cf2 ranges from 3 Hz to 50 Hz inclusive. However, the upper limit of the second cutoff frequency Cf2 can be in a range of approximately 140 Hz or more and 200 Hz or less. The order of the low-pass filter 232 is 3 or higher. The low-pass filter 232 can, for example, comprise multiple low-pass filters. The low-pass filter 232 and the high-pass filter 233 are connected in series.

[0054] If a transfer function of the low-pass filter 232 is Q(s) and a transfer function of the high-pass filter 233 is HPF(s), the support control unit 230 is configured such that the transfer function P(s) of the control target 560 is mapped to a transfer function P n(s) of a predetermined nominal model is restricted in a frequency band in which a gain of Q(s)·HPF(s) is 1. Q(s)·HPF(s) is a complementary sensitivity function T(s) of the inner loop formed by the support control unit 230. As in Fig. As shown in Figure 4, Q(s)·HPF(s), that is, the complementary sensitivity function T(s), exhibits a gain of 0 dB, that is, a gain of 1 in the transfer function in a frequency band where the frequency f is the first cutoff frequency Cf1 or higher and the second cutoff frequency Cf2 or lower. Fig. Figure 4 represents the absolute value of the complementary sensitivity function T(s). In the present description, "the transfer function of the control target is restricted to the transfer function of the nominal model" means, for example, that the control target is controlled in such a way that the transfer function of the control target appears to be a transfer function of the nominal model when considering the input / output relationship.

[0055] The inverse nominal model 231 is an inverse model of a predetermined nominal model (plant model) used to constrain the control target 560. In the present embodiment, the transfer function P n (s) of the predetermined nominal model is expressed by the following formula (3). The transfer function P n -1 (s) of an inverse nominal model 231 is expressed by the following formula (4). Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 Pn−1(s)=(JSTGns+BSTGn)(s2+2ζ2nω2ns+ω2n2s2+2ζ1nω1ns+ω1n2)

[0056] In formulas (3) and (4) s is a Laplace transformer, J STGn is a parameter that represents a moment of inertia of the nominal model, B STGn ω is a parameter that represents a viscosity friction coefficient of the nominal model. 1n is a frequency at a zero point of the transfer function P n (s), ω 2n is a frequency of one pole of the transfer function P n (s), ζ 1n is a damping ratio at the zero point of the transfer function P n (s) and ζ 2n is a damping ratio at the pole of the transfer function P n (s).

[0057] In the present embodiment, the nominal model is a model with frequency characteristics between an inertia-one system and an inertia-two system. Formula (3), which defines the transfer function P n (s) of the nominal model is a formula obtained by adding a damping term to a formula representing an inertia-two system. In the above formula (3), the damping terms are 2ζ 1n ω 1n s and 2ζ 2n ω 2n An equation obtained by removing these damping terms from formula (3) is an equation representing an inertia-two system. In the present embodiment, the degree of the transfer function P is n (s) of the nominal model 3.

[0058] In the present embodiment, the nominal model is a model that takes into account mechanical properties when the steering wheel 521 is steered by the person steering. As described above, the steering target 560 approaches the inertia-one system when the person steering grips the steering wheel 521 firmly and approaches the inertia-two system when the person steering grips the steering wheel 521 loosely. Therefore, the transfer function P(s) of the steering target 560 changes between the inertia-one system and the inertia-two system depending on how a force is exerted on the steering wheel 521 by the person steering. In the present embodiment, by setting the nominal model to a model with frequency characteristics between an inertia-one system and an inertia-two system, the modeling error Δ(s) between the transfer function P can be prevented. n(s) of the nominal model and the transfer function P(s) of the control target 560 becomes too large, regardless of whether the state of the control system 560 is an inertia-one system or an inertia-two system. Therefore, the control target 560 can be appropriately controlled using the nominal model, irrespective of how the driver steers the steering wheel 521. As described above, in the present embodiment, the nominal model is a model that takes into account mechanical properties of the control target 530 based on how the driver grips the steering wheel 521. The control device 100 can appropriately control the control target 560 by having such a nominal model as an internal model.

[0059] In the present description, "the nominal model is a model taking into account mechanical properties when the driver steers the steering wheel" means, for example, that the nominal model can be a model that can compensate for at least part of the influence that the mechanical property gives to the steering objective when the driver steers the steering wheel. The nominal model can, for example, be a model that directly incorporates mechanical properties of the driver's arm movement.

[0060] As in Fig. As shown in Figure 3, the output of the control target 560 is input into the inverse nominal model 231. In particular, the angular velocity ω is θ The value calculated from the rotation angle θa of the input shaft 524a is entered into the inverse nominal model 231. The inverse nominal model 231 yields a torque T. pbased on the above formula (4) and the input angular velocity ω θ This means that the support control unit 230 controls the torque T. p calculated using the nominal model based on the output of the control target 560. The torque T p is equal to the value of the torque input into the nominal model if the output value of the nominal model is the same value as the output value of the control target 560.

[0061] The subtractor SU2 subtracts the output of the subtractor SU1 from the output of the inverse nominal model 231 to obtain a differential torque T. a to generate. This means that the subtractor SU2 generates the differential torque T. a by subtracting from the torque T p , of the input torque T r , before a state compensation value V s , which will be described later, is returned after the correction torque T fThe differential torque T is generated and fed back. d For example, an estimated value of the disturbance torque T d The differential torque T a The output from subtractor SU2 undergoes filtering by low-pass filter 232 and high-pass filter 233, connected in series in that order, and is fed into adder AD3. The differential torque T a The signal, which is filtered by the low-pass filter 232 and the high-pass filter 233, is in a state where a frequency component lower than the first cutoff frequency Cf1 and a frequency component higher than the second cutoff frequency Cf2 are removed. This means that the differential torque T a , which is filtered by the low-pass filter 232 and the high-pass filter 233, a frequency component T aMis which is greater than or equal to the first cutoff frequency Cf1 and less than or equal to the second cutoff frequency Cf2.

[0062] The support adjustment unit 270 generates a compensation value for friction and disturbance and adjusts the differential torque T. d In one embodiment, the support adjustment unit 270 adjusts the frequency component T. aM of the differential torque T d The support adjustment unit 270 is connected in parallel to the high-pass filter 233. The support adjustment unit 270 comprises a friction compensation value calculation unit 250, a disturbance compensation value calculation unit 260, and a subtractor SU2.

[0063] The subtractor SU2 subtracts the output value from the high-pass filter 233 from the output value from the low-pass filter 232. Here, the output value from the low-pass filter 232 is a value obtained by removing a frequency component higher than the second cutoff frequency Cf2 from the differential torque T. d The output value from the high-pass filter 233 is a value obtained by removing a frequency component higher than the second cutoff frequency Cf2 and a frequency component lower than the first cutoff frequency Cf1 from the differential torque T. d Therefore, the value output from subtractor SU3 is the frequency component T. aL , which is lower than the first cutoff frequency Cf1, at which the differential torque T aThe output of subtractor SU3 is entered into friction compensation value calculation unit 250 and disturbance compensation value calculation unit 260. The frequency component T aL includes a frictional force, a restoring torque, a disturbance torque caused by play in the control target 560, a torque ripple generated in the control target 560, and the like.

[0064] The friction compensation value calculation unit 250 calculates the friction compensation value V f , which at least partially compensates for the frictional force generated in the control target 560, based on the differential torque T a As described above, the value from the subtractor SU3, which is entered into the friction compensation value calculation unit 250, is the frequency component T. aL , which is lower than the first cutoff frequency Cf1, at which the differential torque T aTherefore, in the present embodiment, the friction compensation value calculation unit 250 calculates the friction compensation value V. f based on a component with a frequency lower than the first cutoff frequency Cf1, in which the differential torque T a .

[0065] The friction compensation value calculation unit 250 comprises a limiter 252 and a gain adjuster 253. The limiter 252 limits the output value from the subtractor SU3. The limiter 252 clips the input value to the upper or lower threshold if the input value exceeds the upper or lower threshold. The gain adjuster 253 applies a gain K1 to the output value from the limiter 252. The friction compensation value calculation unit 250 calculates the friction compensation value V. fby applying the limit by the limiter 252 and the gain K1 to the component of the frequency that is lower than the first cutoff frequency Cf1, in which the differential torque T a . The threshold of the limiter 252 and the value of the gain K1 are determined beforehand based on, for example, the frictional force that is actually generated in the control target 560.

[0066] To compensate for friction, the correction torque T fTo apply the control used for the model-following control in the support control unit 230, it is necessary to pay attention to the stability condition of the model-following control. This condition consists of the fact that the gain in the gain property of the transfer function of the friction compensation value calculation unit 250, which is limited to the property that considers stability, does not exceed 1 according to the low gain theorem, which is described later. This is derived from the design condition of the low-pass filter 232. In the present embodiment, the subtractor SU3 is provided in the preceding stage of the limiter 252, so that the value of the gain K1 in the gain adjuster 253 is set to a maximum of 1 and the gain in the gain property becomes 1 under this condition, and the subtraction processing is applied.In other words, the friction compensation value calculation unit 250 behaves like a low-pass filter with a transfer function of 1 - HPF(s).

[0067] The friction compensation value V f , which is output from the friction compensation value calculation unit 250, is a value that compensates at least part of the friction force component that is in the frequency component T aL of the differential torque T a is included. Generally calculated, since suitable friction is required for the control target 560, the friction compensation value calculation unit 250 calculates a value that is smaller than the frictional force actually generated in the control target 560, as the friction compensation value V. fThis makes it possible to achieve very precise friction compensation while maintaining a suitable frictional force on the control target 560. The friction compensation target using the friction compensation value V f Examples include the friction of the motor 543, the friction of the deceleration mechanism 544, the right / left difference in the friction of the deceleration mechanism 544, and the like.

[0068] Here, the frequency component T includes aL of the differential torque T a In addition to the frictional force component, the restoring torque generated in the control target 560, the disturbance torque caused by the backlash generated in the control target 560, and the torque ripple generated in the control target 560 are also included. Therefore, the friction compensation value V comprises f , which is obtained by processing the frequency component T aLthrough the limiter 252 and the gain adjuster 253, also a compensation value to compensate for at least part of the restoring torque generated in the control target 560, the disturbance torque caused by the backlash generated in the control target 560, and the torque ripple generated in the control target 560.

[0069] The vehicle, equipped with the electric power steering device 1000, can operate in both automatic and manual driving modes. In this case, the gain K1 of the gain adjuster 253 can be switched according to the driving mode. The greater the gain K1 of the gain adjuster 253, the greater the degree of friction reduction. The gain K1 in automatic driving mode is preferably greater than the gain K1 set in manual driving mode. Consequently, it is possible to apply optimal friction compensation to an automatic driving mode where a greater reduction in friction is required.

[0070] The interference compensation value calculation unit 260 calculates an interference compensation value V dto compensate for at least part of the restoring torque generated in the control target 560. In the present embodiment, the disturbance compensation value V comprises d A compensation value to compensate for at least part of the frictional force generated in the control target 560, the disturbance torque caused by the backlash generated in the control target 560, and the torque ripple generated in the control target 560. The disturbance compensation value calculation unit 260 calculates a disturbance compensation value V d based on a differential torque T a , which is a difference between the torque T p , which is output from the inverse nominal model 231, and the input torque T r This means that the interference compensation value calculation unit 260 has the interference compensation value V. d based on the differential torque T dcalculated that the difference between the torque T p , which is calculated using the nominal model based on the output of the control target 560, and the input torque T r minus the correction torque Tf. As described above, the value from subtractor SU3, which is entered into the disturbance compensation value calculation unit 260, is a frequency component that is lower than the first cutoff frequency Cf1, in which the differential torque T a Therefore, in the present embodiment, the interference compensation value calculation unit 260 calculates the interference compensation value V. d based on a component with a frequency lower than the first cutoff frequency Cf1, in which the differential torque T d .

[0071] The disturbance compensation value calculation unit 260 comprises a limiter 262 and a gain adjuster 263. The limiter 262 limits the output value from the subtractor SU3. The limiter 262 truncates the input value to the upper or lower threshold if the input value exceeds the upper or lower threshold. The threshold of the limiter 262 differs, for example, from the threshold of the limiter 252. The gain adjuster 263 applies a gain K2 to the output value from the limiter 262. The maximum value of the gain K2 of the gain adjuster 263 is determined under the condition that the transfer function P(s) of the control target 560 is given by the transfer function P n(s) of the nominal model is limited. The value of the gain K2 differs, for example, from the value of the gain K1. The value of the gain K2 is, for example, approximately 0.3 or more and 0.8 or less. The gain K2 of the gain adjuster 263 can be switched according to the movement mode of the vehicle.

[0072] The interference compensation value V d is a value that compensates for at least part of the restoring torque component that is in the frequency component T aL of the differential torque T d This is included. For example, the disturbance compensation value calculation unit 260 calculates a value that corresponds to approximately half of the restoring torque actually generated in the control target 560, as the disturbance compensation value V. dThe restoring torque actually generated in the control target 560 is obtained experimentally beforehand for each frequency. The threshold of the limiter 262 of the noise compensation value calculation unit 260 and the value of the gain K2 are adjusted to values ​​at which the noise compensation value V d as a value calculated that is approximately half the magnitude of the restoring torque previously obtained. The disturbance compensation value V d , which is calculated by the disturbance compensation value calculation unit 260, is a value that differs from the friction compensation value V f This differs, and is calculated using the friction compensation value calculation unit 250.

[0073] Here, the frequency component T includes aL of the differential torque T aIn addition to the restoring torque, a frictional force generated in the control target 560, a disturbance torque caused by backlash generated in the control target 560, and a torque ripple generated in the control target 560 are also included. Therefore, the disturbance compensation value V comprises d , which is obtained by processing the frequency component T aL through the limiter 262 and the gain adjuster 263, also a compensation value to compensate for at least part of the frictional force generated in the control target 560, the disturbance torque caused by backlash generated in the control target 560, and the torque ripple generated in the control target 560.

[0074] The adder AD3 adds the output value from the support adjustment unit 270 to the output value from the high-pass filter 233. This means that the adder AD3 calculates the friction compensation value V. fand the interference compensation value V d to the frequency component T aM The adder AD3 provides the correction torque T. f from which is calculated by adding the frequency component T aM , of the friction compensation value V f and the disturbance compensation value V d The correction torque T f The value output from adder AD3 is fed back to the input of control target 560, i.e., the input torque T. r As described above, in the present embodiment the support control unit 230 generates the correction torque T f by adding the friction compensation value V f and the disturbance compensation value V d to the differential torque T a , from which the frequency component lower than the first cutoff frequency Cf1 was removed by the high-pass filter 233, that is, the frequency component T aM .

[0075] The state feedback unit 280 carries out the state compensation value V s based on the output of the control target 560 to the input torque T r back, so that the apparent transfer function of the control target 560 corresponds to the transfer function P n(s) of the nominal model. The apparent transfer function of the control target 560 is, for example, a transfer function of a section in a case where a section located inside the feedback loop generated by the support control unit 230 is considered to be that section. In particular, in the present embodiment, the apparent transfer function of the control target 560 is a transfer function of the entire section from the subtractor SU1 to the output of the control target 560 and is a transfer function of a section that combines the state feedback unit 280 and the control target 560. In the present embodiment, the state feedback unit 280 introduces the state compensation value V. s back to the input torque T r , after the state compensation value is determined by the correction torque T fwas corrected and before the state compensation value is entered into control target 560.

[0076] The state compensation value V s comprises a compensation value that compensates for at least part of the inertial force, viscosity force, and frictional force generated in the control target 560. In particular, the state compensation value V comprises s a compensation value that compensates for at least part of the inertial force, the viscosity force, and the frictional force generated in the motor 543. In the present embodiment, the state compensation value V sa compensation value that includes the inertial force generated in the motor 543, the viscosity force generated in the motor 543, and the friction force generated in the motor 543.

[0077] The state feedback unit 280 comprises an inertial compensator 281, a viscosity compensator 282, and a friction compensator 283. The inertial compensator 281 calculates a compensation value to compensate for at least part of the inertial force generated in the motor 543, based on the angular velocity ω of the motor 543. The viscosity compensator 282 calculates a compensation value to compensate for at least part of the viscosity force generated in the motor 543, based on the angular velocity ω of the motor 543. The friction compensator 283 calculates a compensation value to compensate for at least part of the friction force generated in the motor 543, based on the angular velocity ω of the motor 543. In the present embodiment, the State compensation value V sA compensation value calculated by the inertial compensator 281, a compensation value calculated by the viscosity compensator 282, and a compensation value calculated by the friction compensator 283. The compensation value calculated by the inertial compensator 281, the compensation value calculated by the viscosity compensator 282, and the compensation value calculated by the friction compensator 283 are output to the adder AD1 and added to the input torque T. r added by the correction torque T f has been corrected.

[0078] Next, the control by the support control unit 230 is described in more detail. The support control unit 230 controls the control target 560 using the inverse model of the nominal model as an internal model, that is, the inverse nominal model 231. In the present embodiment, the torque ripple or the like, depending on the angular velocity ω of the motor 543, can be compensated by the feedback loop formed by the support control unit 230. The angular velocity ω signal used for control can be corrected for each type of motor 543, and the accuracy of the angular velocity ω signal can be improved compared to the current signal and the like. Consequently, torque ripple compensation can be applied to torque control with high accuracy.

[0079] The support control unit 230 is structurally similar to a conventional disturbance estimator (disturbance observer), but differs from the conventional disturbance estimator in its operation and effect. A conventional disturbance estimator estimates a disturbance torque by using an inverse plant model as an internal model, approximating a model close to the control target 560, and reduces the influence of disturbances by pre-adjusting the disturbance torque.

[0080] The control by the support control unit 230 according to the present embodiment utilizes the effect that the transfer function P(s) of the control target 560 is determined by the transfer function P n(s) of the nominal model is restricted as an internal model by the feedback loop. For example, if the nominal model is defined such that there is no torque ripple, the transfer function P(s) of the control target 560 is restricted to the property without torque ripple by the model-following control, and consequently, the torque ripple can be reduced by applying torque ripple compensation. By setting the nominal model as a low-inertia model and restricting the control target 560 with the nominal model, the control target 560 can be treated as a low-inertia model. The control target 560 can be treated as a low-viscosity model by setting the nominal model as a low-viscosity model and restricting the control target 560 with the nominal model.In addition to compensating for the torque ripple of the motor 543, for example, compensation for lost torque or compensation for motor inertia is performed by executing model-following control by the support control unit 230. In the above formulas (3) and (4), by appropriately adjusting J. STGn and B STGn a desired frequency property is assigned to the transfer function P(s) of the control target 560.

[0081] If there is a modeling error between the transfer function P(s) of the control target 560 and the transfer function P n (s) of the nominal model Δ(s), the transfer function P(s) of the control target 560 is expressed by the following formula (5). Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2(1+Δ(s))

[0082] The gain property of the transfer function P(s) of the control target 560, for example, exhibits peaks at two frequency values. The modeling error Δ(s) appears, for example, near the higher frequency peak of the two peaks in the gain property of the control target 560. Therefore, the reciprocal 1 / Δ(s) of the modeling error Δ(s), as shown in Fig. Figure 4 shows a valley floor in a relatively high-frequency area. Fig. In section 4, the modeling error Δ(s) is shown as an absolute value. As the modeling error Δ(s) increases, the deviation between the transfer function P(s) of control target 560 and the transfer function P increases. n(s) of the nominal model, and the control of the control target 560 using the nominal model by the support control unit 230 becomes unstable. Therefore, in the region where the modeling error Δ(s) is relatively small, the control target 560 can be controlled stably and appropriately by restricting the control target 560 to the nominal model, with the gain of the complementary sensitivity function T(s), i.e., Q(s)·HPF(s), set to 1. The frequency property of the modeling error Δ(s) can be adjusted by adjusting J STGn and B STGn in the transfer function P n (s) of the nominal model. The frequency band at which the gain of Q(s)·HPF(s) becomes 1 can be adjusted by adjusting the first cutoff frequency Cf1 and the second cutoff frequency Cf2. Consequently, the gain of Q(s)·HPF(s) can be adjusted to 1 in the frequency band where the modeling error Δ(s) is small.

[0083] In Fig. 4. 1 / Δ(s) is relatively high in a frequency band of the second cutoff frequency Cf2 or less and decreases rapidly in a frequency band higher than the second cutoff frequency Cf2. The model-following control to restrict the control target 560 to the nominal model can be performed stably, for example, in a range where 1 / Δ(s) is greater than 1, that is, in a range where 1 / Δ(s) is greater than 0 dB. Therefore, as in Fig. As shown in Figure 4, if the gain of Q(s)·HPF(s) 1 is increased by adjusting 1 / Δ(s) to a higher value than 1 in the frequency band in which the gain of Q(s)·HPF(s) 1 is increased, it is possible to control the control target 560 stably and appropriately by restricting it to the nominal model.

[0084] For example, to extend a frequency band in which the control target can be stably and appropriately controlled by restricting the control target 560 to the nominal model, the second cutoff frequency Cf2 can be increased within a range where 1 / Δ(s) is not 1 or less, that is, within a frequency band lower than a frequency at which a curve containing 1 / Δ(s) in Fig. 4 indicates that it intersects the horizontal axis.

[0085] However, if the second cutoff frequency Cf2 is set too high, the gain of Q(s)·HPF(s) remains relatively high, even though 1 / Δ(s) becomes low in a frequency band higher than the second cutoff frequency Cf2, and the control could become unstable. On the other hand, in the present embodiment, since the order of the low-pass filter 232 is set to 3 or more, the gain of Q(s)·HPF(s) can be steeply reduced in a region where the frequency is higher than the second cutoff frequency Cf2. Consequently, even if the second cutoff frequency Cf2 is set relatively high, the gain of Q(s)·HPF(s) can be immediately reduced in the frequency band higher than the second cutoff frequency Cf2, thus preventing the control of the target 560 from becoming unstable.

[0086] The robust stability of the support control unit 230 is guaranteed when the low gain theorem, represented in the following formula (6), is established between the complementary sensitivity function T(s) and the modeling error Δ(s). |T(jω)|<1|Δ(jω)| or |T(jω)Δ(jω)|<1, ∀s=jω

[0087] As described above, to perform model-following control using the nominal model in the support control unit 230, it is sufficient if T(s) = 1. However, considering robust stability, it is necessary to satisfy the formula (6) above. As can be seen from this, T(s) = 1 and formula (6) cannot be compatible in all frequency bands, and the suppression of interference and the like by the support control unit 230 and robust stability are incompatible.

[0088] As in Fig. As shown in Figure 4, even in a region where the frequency is lower than the first cutoff frequency Cf1, the gain of Q(s)·HPF(s), that is, the gain of the complementary sensitivity function T(s), is less than 1. In a region where the gain of Q(s)·HPF(s) is less than 1, the reaction force control unit 210 controls the input torque T. r , to control the target frequency 560. As described above, in the region where the frequency is higher than the second cutoff frequency Cf2, the gain of Q(s)·HPF(s) is greatly reduced and the correction torque T is increased. f The signal from the support control unit 230 is barely fed back to the input of the control target 560. On the other hand, in a region where the frequency is lower than the first cutoff frequency Cf1, the correction torque T fThe input of the control target 560 is fed back, with the gain of Q(s)·HPF(s) set to a specific value. In a region where the frequency is lower than the first cutoff frequency Cf1, the compensation value generated in the support adjustment unit 270 described above is fed back to the input of the control target 560 according to the gain of Q(s)·HPF(s).

[0089] The control device 100 performs torque control in the reaction force control unit 210 with respect to a low-frequency torque signal, lower than the first cutoff frequency Cf1, and performs control such that the angular velocity ω ≈ 0 with respect to a high-frequency disturbance, higher than the second cutoff frequency Cf2, thereby stabilizing the steering so that the steering wheel 521 is not turned. To achieve this, the control device 100 reduces the high-frequency gain of the torque control using the reaction force control unit 210 and limits the transfer function P(s) of the control target 560 to the property at which the high-frequency gain decreases, using the support control unit 230.The latter processing is carried out in such a way that the control target 560 does not react to a disturbance when the disturbance is entered into the control target 560.

[0090] The effective range of the model-following control by the support control unit 230 is a region from the first cutoff frequency Cf1 to the second cutoff frequency Cf2, inclusive. This means that the lower limiting frequency of the effective range of the model-following control depends on the first cutoff frequency Cf1. Therefore, the lower limiting frequency of the effective range of the model-following control is determined by adjusting the first cutoff frequency Cf1 of the high-pass filter 233 so as not to impede the control of the reaction force control unit 210 in the low-frequency range.

[0091] According to the present embodiment, the support control unit 230, which controls the correction torque T fto correct the input torque T r generated based on the output of the control target 560 and the nominal model, designed such that the transfer function P(s) of the control target 560 is bounded by the transfer function P n (s) of the nominal model in the frequency band where the gain of the gain property of the complementary sensitivity function T(s) with respect to the modeling error Δ(s) between the control target 560 and the nominal model is 1. In particular, in a region where the modeling error Δ(s) between the nominal model and the control target 560 is small, the gain of the gain property of Q(s)·HPF(s) is set to 1, and the difference in output between the control target 560 and the nominal model is defined as the correction torque T. f to the input torque T rtraced back so that the transfer function P(s) of the control target 560 appears close to the transfer function P n (s) of the nominal model can be brought into play. Therefore, for example, by setting the nominal model as a model in which no torque ripple occurs, the torque ripple can be removed or reduced from the output of control target 560 in the frequency band in which the gain for the gain property of Q(s)·HPF(s) is 1. By setting the nominal model as a low-inertia, low-viscosity model, control target 560 can be restricted to a low-inertia, low-viscosity model, and control target 560 can be controlled without further ado.

[0092] For example, a conventional disturbance estimator incorporates a model that approximates the control target 560 as its internal model and compensates for disturbances generated in the control target 560. However, it is difficult to have a model that is exactly the same as the control target 560 as the internal model, and a modeling error Δ(s) inevitably occurs. Therefore, in order to suppress control instability, the gain of the gain property Q(s)·HPF(s) in all frequency bands is set to a value less than 1 in the conventional disturbance estimator. Because the internal model is only an approximation of the actual model of the control target 560, the torque ripple or similar phenomena generated in the control target 560 itself cannot be removed by the conventional disturbance estimator, although the disturbance applied externally to the control target 560 can be estimated.

[0093] On the other hand, in the present embodiment, the nominal model, which is included as an internal model in the control device 100, is set as the ideal model for the control target 560 instead of the model that reproduces the actual model of the control target 560, and the gain of Q(s)·HPF(s) is set to 1 in the region where the modeling error Δ(s) is small. Consequently, by appropriately adjusting the nominal model, not only the disturbance applied externally to the actual control target 560 but also the torque ripple generated internally in the control target 560 can be eliminated. Therefore, according to the present embodiment, the control device 100 can appropriately control the control target 560, and the steering feel experienced by the driver can be improved.

[0094] For example, the steering mechanism 530 has a structure in which the input shaft 524a and the output shaft 524b are connected to the torsion bar 546, which is arranged between them, and is not a simple inertia-one system. Therefore, if the object to be controlled by the control device 100 is, for example, an inertia-one system comprising only the motor 543, it may be difficult to adequately guarantee torque ripple, disturbance, and the like. On the other hand, as in the present embodiment, it is conceivable to consider a section comprising sections on both sides that sandwich around the torsion bar 546 as the control target 560; however, the control target 560 is not considered a simple inertia-two system.As described above, the control target 560 changes between the inertia-one system and the inertia-two system depending on how the steering wheel 521 is steered by the driver or similar factors. Therefore, even if the control target 560 is simply a model of an inertia-two system, it may be difficult to adequately guarantee torque ripple, disturbance, and the like.

[0095] On the other hand, according to the present embodiment, the nominal model is a model in which mechanical properties are taken into account when the steering wheel 521 is steered by the driver. Therefore, the nominal model can be suitably adapted to the property of the control target 560, which changes according to the way the steering wheel 521 is steered by the driver. Consequently, torque ripple, disturbance, and the like can be more suitablely compensated for by limiting the control target 560 to the nominal model by the model-following control described above. Therefore, the steering feel experienced by the driver can be further improved.

[0096] According to the present embodiment, the degree of the transfer function P is n(s) of the nominal model 3 or more. Here, the present inventors have found that the order of the transfer function of the steering mechanism 530 is, for example, 6. Therefore, by adjusting the transfer function P n (s) of the nominal model to a higher-order transfer function closer to the order of the transfer function of the steering mechanism 530, the steering mechanism 530 can be controlled more appropriately by the control device 100. Therefore, the steering feel experienced by the person steering can be further improved.

[0097] In the present embodiment, the control target is captured in the area of ​​the third-order transfer function, and the degree of the transfer function P is n (s) of the nominal model is also set to 3. However, the control target could, for example, be captured in a range of a transfer function of four or more degrees, and the degree of the transfer function P n(s) of the nominal model could be four or more. If the degree of the transfer function in the range considered to be the control target, and the degree of the transfer function P n By bringing the order (s) of the nominal model closer to the sixth degree, which is the degree of the transfer function of the steering mechanism 530, the control can be performed even more favorably. If, for example, the steering torque sensor 541 is a coordinate converter or resolver or the like, it is conceivable that the transfer function of the steering torque sensor 541 is quadratic. Therefore, if, for example, the order of the transfer function P n (s) of the nominal model is set to the fifth order, whereby the control target is obtained by adding the steering torque sensor 541 to the control target 560 of the preferred embodiment, a more suitable control is carried out.

[0098] According to the present embodiment, the nominal model is a model with frequency characteristics between an inertial-one system and an inertial-two system. As described above, the characteristics of the control target 560 change between the inertial-one system and the inertial-two system depending on how the steering wheel 521 is steered by the driver. Therefore, by setting the nominal model to a model with frequency characteristics between the inertial-one system and the inertial-two system, the control target 560 can be controlled more effectively using the nominal model. Consequently, the steering feel experienced by the driver can be further improved.

[0099] According to the present embodiment, the formula that defines the transfer function P is n(s) of the nominal model represents a formula obtained by adding a damping term to a formula representing the inertia-two system. Therefore, the transfer function P n (s) of the nominal model are suitable and can readily be made into a model exhibiting frequency characteristics between the inertia-one system and the inertia-two system.

[0100] According to the present embodiment, the transfer function P n (s) of the nominal model can be expressed by the formula (3) above. Therefore, the transfer function P n (s) of the nominal model can be made more suitable and easier to a model with frequency properties between the inertia-one system and the inertia-two system.

[0101] As described above, the control target 560 is wider than the inertia-one system, and the nominal model according to the control target 560 is set such that the disturbance, which includes torque ripple in a wider frequency band than conventional, can be compensated for by feedback control using the nominal model. Consequently, the frequency band in which the disturbance can be suppressed can be wider than the conventional frequency band. In particular, for example, the torque ripple generated in the control target 560 includes torque ripple caused by a worm gear used for the deceleration mechanism 544. The torque ripple caused by the worm gear could be a disturbance of, for example, about 50 Hz.In the design of a conventional disturbance estimator or similar device, the torque ripple cannot be suppressed in the first place, and the frequency band that can be suppressed as disturbance is also lower than 50 Hz, so that the torque ripple caused by the worm gear described above cannot be suppressed. However, according to the design and method of the present embodiment, by appropriately adjusting the nominal model, the support control unit 230 can also compensate for the torque ripple of a relatively high frequency caused by the worm gear described above, and the torque ripple with a relatively high frequency can be suppressed.

[0102] According to the present embodiment, the control device 100 comprises the state feedback unit 280, which provides the state compensation value V sbased on the output of the control target 560 to the input torque T r returns, so that the apparent transfer function of the control target 560 corresponds to the transfer function P n (s) of the nominal model. Therefore, the control target 560, which is to be controlled by the feedback through the support control unit 230, can be brought close to the nominal model in appearance as the internal model. Consequently, when the support control unit 230 performs model-following control, the control target 560 can be considered a model close to the nominal model, and the modeling error Δ(s) between the control target 560 and the nominal model can be reduced. Therefore, it is possible to broaden the frequency band at which the transfer function P(s) of the control target 560 approximates the transfer function P. n(s) of the nominal model may be limited, with the gain set to 1 for the gain property of Q(s)·HPF(s). Therefore, the control following a model can be carried out by the support control unit 230 in a wider frequency band, and the steering feel experienced by the driver can be further improved.

[0103] According to the present embodiment, the state feedback unit 280 carries out the state compensation value V. s to the input torque T r back after this is due to the correction torque T fwas corrected and before it is input into the control target 560. Therefore, feedback from the state feedback unit 280 can be fed into the feedback loop of the support control unit 230. Consequently, when viewed from the perspective of the support control unit 230, the state feedback unit 280 and the control target 560 can be considered collectively as a single control target. Therefore, by considering the collective transfer function of a control target as the apparent transfer function P(s) of the control target 560, control by the support control unit 230 using the nominal model can be performed more appropriately.

[0104] According to the present embodiment, the state compensation value V comprises sa compensation value to compensate for at least part of the inertial force, the viscosity force, and the friction force generated in the control target 560. Therefore, the apparent transfer function of the control target 560 can more closely approximate the transfer function P. n (s) of the nominal model. In the present embodiment, the state compensation value V is s , which includes both the inertial force and the viscosity force and the friction force, is fed back through the inertial compensator 281, the viscosity compensator 282 and the friction compensator 283, so that the apparent transfer function of the control target 560 is more suitablely close to the transfer function P n(s) of the nominal model. The gains of the inertial compensator 281, the viscosity compensator 282 and the friction compensator 283 are appropriately adjusted to values ​​that bring the control target 560 closer to the nominal model.

[0105] According to the present embodiment, the state compensation value V comprises s a compensation value to compensate for at least part of the inertial force, the viscosity force, and the friction force generated in the motor 543. Therefore, the inertial force or similar force generated in the motor 543 can be compensated so that it approximates the nominal model. Consequently, the apparent transfer function of the control target 560 can be brought more appropriately close to the transfer function P. n (s) of the nominal model are brought.

[0106] According to the present embodiment, the support control unit 230 generates the correction torque T f based on a difference between the torque T p , which is calculated using the nominal model based on the output of the control target 560, and the input torque T r , before the state compensation value V s is returned after being subjected to the corrective torque T f was corrected. This means that the input torque T r , before the state compensation value V sThe value added can be input to the subtractor SU2 of the support control unit 230. Therefore, when viewed from the perspective of the support control unit 230, the state feedback unit 280 and the control target 560 can be more appropriately considered as a single control target. Consequently, the apparent transfer function P(s) of the control target 560, viewed from the perspective of the support control unit 230, can be brought closer to the nominal model.

[0107] For example, in the inertia-two system, it is difficult to estimate the motion of the entire inertia-two system from the motion of the inertia system at the output side. This means that even in the present embodiment, it is difficult to estimate the motion of the nominal model between the inertia-one system and the inertia-two system solely from the information about the rotation angle θb of the output shaft 524b, and it may be difficult to determine the torque T. p, which is output from the inverse nominal model 231, is suitable for calculation. On the other hand, the support control unit 230 generates the correction torque T according to the present embodiment. f based on the rotation angle θa of the input shaft 524a. Therefore, by using the rotation angle θa of the input shaft 524a at the input side, the motion of the nominal model between the inertia-one system and the inertia-two system can be suitably estimated. Consequently, the torque T can be p , which is output from the inverse nominal model 231, can be appropriately calculated and the correction torque T can be determined. f be produced appropriately.

[0108] The restoring torque is transmitted to the driver as a response when the driver turns the steering wheel 521. Therefore, for example, in a low-frequency region encompassing the restoring torque, it is conceivable to significantly reduce the gain of the support control unit 230 so that the restoring torque is not compensated. In this case, since no compensation is performed by the support control unit 230, the reaction force control unit 210 performs compensation and controls the steering target 560. However, in this case, the amount of compensation in the reaction force control unit 210 increases, and the gain of the reaction force control unit 210 can become too high. Therefore, control by the control device 100 can become unstable.

[0109] On the other hand, according to the present embodiment, the support control unit 230 comprises the disturbance compensation value calculation unit 260, which calculates the disturbance compensation value V d Calculated to compensate for at least part of the restoring torque generated in control target 560. The correction torque T f includes a disturbance compensation value V dTherefore, the support control unit 230 can perform compensation, even in the low-frequency region containing the restoring torque, by at least the restoring torque itself. Consequently, the amount of compensation required in the reaction force control unit 210 can be reduced, and the gain in the reaction force control unit 210 can be reduced. Therefore, unstable steering can be suppressed by the control device 100. Therefore, the steering feel experienced by the driver can be further improved. Since at least part of the restoring torque is compensated by the support control unit 230, it is possible to reduce the reaction force exerted on the driver when the driver turns the steering wheel 521. Therefore, the steering feel experienced by the driver can be further improved.In particular, in the present embodiment, since only part of the restoring torque is compensated by the support control unit 230 when the steering person steers the steering wheel 521, the steering person can continue to steer the steering wheel 521 without further ado, while the steering person receives a suitable response.

[0110] According to the present embodiment, the interference compensation value V comprises d a compensation value to compensate for at least part of the frictional force generated in the control target 560, the disturbance torque caused by the backlash generated in the control target 560, and the torque ripple generated in the control target 560. Therefore, not only the restoring torque, but also at least part of the frictional force, the disturbance torque caused by the backlash, and the torque ripple can be compensated by the disturbance compensation value V. dThis is compensated for by the disturbance compensation value calculation unit 260 of the support control unit 230. Consequently, at least a portion of the frictional force, the disturbance torque caused by the play, and the torque ripple itself can be compensated in a low-frequency region lower than the first cutoff frequency Cf1, at which model-following control cannot be performed. Therefore, the steering feel experienced by the driver can be further improved in the low-frequency region lower than the first cutoff frequency Cf1.

[0111] According to the present embodiment, the interference compensation value calculation unit 260 calculates the interference compensation value V. d based on the differential torque T a , which is the difference between the torque T p, which is calculated using the nominal model based on the output of the control target 560, and the input torque T r Therefore, a value such as the restoring torque can be suitablely derived from the differential torque T. d can be estimated and the disturbance compensation value V can be determined. d appropriately calculated.

[0112] According to the present embodiment, the interference compensation value calculation unit 260 calculates the interference compensation value V. d based on a component of a frequency that is lower than the first cutoff frequency Cf1, in which the differential torque T a , that is, the frequency component T aL Since the restoring torque is in the relatively low-frequency component T aL Included, the restoring torque can be appropriately compensated by calculating the disturbance compensation value V. d based on the frequency component T aLThe frequency component T aL It also includes a frictional force generated in the control target 560, a disturbance torque caused by backlash generated in the control target 560, and a torque ripple generated in the control target 560. Therefore, it is possible to calculate the disturbance compensation value V. d by performing processing on the frequency component T through the limiter 262 and the gain adjuster 263 aL possible to adjust the interference compensation value V d to calculate which is able to compensate for the frictional force generated in the control target 560, the disturbance torque caused by the backlash generated in the control target 560, and the torque ripple generated in the control target 560, in addition to the restoring torque.

[0113] According to the present embodiment, the support control unit 230 comprises the friction compensation value calculation unit 250, which calculates the friction compensation value V f to compensate for at least part of the frictional force generated in the control target 560, based on the differential torque T d calculated. The support control unit 230 adds the friction compensation value V. f and the interference compensation value V d to the differential torque T a , from which the frequency component T aL , which is lower than the first cutoff frequency Cf1, which was removed by the high-pass filter 230, that is, the frequency component T aM to generate the correction torque T f Therefore, the frictional force generated in the control target 560 can be more appropriately compensated by the support control unit 230.

[0114] According to the present embodiment, the friction compensation value calculation unit 250 calculates the friction compensation value V f based on a component with a frequency lower than the first cutoff frequency Cf1, in which the differential torque T a , that is, the frequency component T aL Therefore, the frictional force can be adequately compensated even in a low-frequency region lower than the first cutoff frequency Cf1, at which model-based control cannot be performed. Consequently, in the low-frequency region lower than the first cutoff frequency Cf1, the gain of the reaction force control unit 210 can be further reduced, and the instability of the control target 560 can also be suppressed.

[0115] In conventional friction compensation control, when the angular velocity ω of the motor 543 is close to zero, any change in the friction compensation value relative to the angular velocity ω of the motor 543 must be gradual to avoid fluttering. Consequently, highly accurate friction compensation control may not be possible under certain circumstances. According to the inventor's study, to solve this problem, it is desirable to estimate and compensate for the friction in a predictive manner. According to the friction compensation provided by the friction compensation value calculation unit 250 of the present embodiment, since friction is estimated in a predictive manner and the friction compensation value V is determined... f The problem can be solved by calculating the solution.

[0116] For example, an auxiliary device has been developed that detects a lane, such as a white or yellow line, when driving on a highway and assists in the automatic driving of a vehicle following the lane. It is known that in a vehicle equipped with such an auxiliary device and an electric power steering system, if a right / left differential exists due to friction in the deceleration mechanism 544, the control of the auxiliary device, which causes the vehicle to move precisely along the center of the lane, may be impaired. According to the friction compensation provided by the friction compensation value calculation unit 250 of the present embodiment, even in a case where a right / left differential exists due to friction in the deceleration mechanism 544, the friction can be estimated, thus solving the above problem.The angular velocity ω of the motor 543, which is the output of the control target 560, includes information about a right / left difference in the case of friction of the deceleration mechanism 544.

[0117] The present inventors have confirmed the effect obtained by applying model-following control through the control device 100 of the embodiment described above by performing an actual vehicle measurement. In the actual vehicle measurement, a case in which model-following control is not applied and a case in which model-following control is applied are compared for torque ripple, viscosity sensation, friction, and inertia.

[0118] The Fig. 5, Fig. 7, Fig. 9 and Fig. Figure 11 presents measurement results of the steering angle [degrees] and the torsional torque [Nm] in a case where the model-based control is not applied. Fig. 6, Fig. 8, Fig. 10 and Fig. Figure 12 presents measurement results of the steering angle [degrees] and the torsional torque [Nm] in a case where model-based control is applied. The graphs of Fig. From 5 to 12, the horizontal axis represents the steering angle [degrees] and the vertical axis represents the torsional torque [Nm]. The torsional torque is a torsion bar torque and is a steering torque T. h .

[0119] The Fig. 5 and Fig. Figure 6 shows signal waveforms when the steering wheel 521 is turned at 90 degrees / second. Referring to the enlarged views of the Fig. 5 and Fig. 6 confirms that the signal pattern of Fig. 6, to which the model-following control is applied, exhibits a smaller torque ripple than the signal waveform of Fig. 5, to which the model-following control is not applied. Consequently, it was confirmed that the torque ripple at the time of steering of the steering wheel 521 can be reduced by applying the model-following control.

[0120] The Fig. 7 and Fig. Figure 8 represents signal waveforms when the steering wheel 521 is steered at a steering frequency of 2 Hz. From the Fig. 7 and Fig. 8 confirms that a fluctuation quantity D1em of the torsional torque in a signal waveform EM1 from Fig. 8, to which the control following a model is applied, is smaller than a fluctuation quantity D1ce of the torsional torque in a signal waveform CE1 from Fig. 7, to which the model-following control is not applied. Consequently, it was confirmed that the viscosity sensation at the time of steering the steering wheel 521 can be reduced by applying the model-following control.

[0121] The Fig. 9 and Fig. Figure 10 represents signal waveforms when the steering wheel 521 is turned by ± 10 degrees, with the steering frequency set to 0.5 Hz. From the Fig. 9 and Fig. 10 confirms that a width D2em, which is indicated by the arrow in a signal waveform EM2 in Fig. 10 is displayed, to which the control following a model is applied, is smaller than a width D2ce, which is defined by the arrow in a signal waveform CE2 in Fig. 9 is displayed, to which the model-following control is not applied. Both widths D2ce and D2em correspond to the amount of friction. Therefore, it was confirmed that the friction at the time of steering wheel 521 can be reduced by applying the model-following control.

[0122] The Fig. 11 and Fig. Figure 12 represents signal waveforms when the steering wheel 521 is turned back at a steering frequency of 2 Hz. Fig. Figure 11 shows an ellipse in a dashed line, representing a signal waveform section when the steering wheel 521 is turned back. Fig. 12 shows an ellipse in a dashed line Eem a signal waveform segment when the steering wheel 521 is turned back. From the Fig. 11 and Fig. 12 confirms that the detent caused by inertia at the time of reversal in the signal path EM3 from Fig. 12 is caused, to which the control following a model is applied, is smaller than the detent caused by the inertia at the time of a reversal in the signal path CE3 from Fig. 11, on which the model-following control is not applied. Consequently, it was confirmed that the feeling of inertia at the time of steering the steering wheel 521 can be reduced by applying the model-following control.

[0123] The present invention is not limited to the embodiments described above, and other embodiments and methods may be used within the scope of protection of the technical idea of ​​the present invention. In the embodiment described above, the support control unit is configured such that the transfer function of the control target is limited by the transfer function of the nominal model in the frequency band in which the gain of the complementary sensitivity function with respect to the modeling error between the control target and the nominal model is 1; however, the present invention is not limited thereto.The support control unit could be designed such that the transfer function of the control target is limited by the transfer function of the nominal model in the frequency band where the gain of the complementary sensitivity function with respect to the modeling error between the control target and the nominal model is approximately 1. "The gain is essentially 1" includes, for example, a case where the gain is 0.8 or greater and 1.2 or less, in addition to a case where the gain is 1. The numerical range is, for example, a range where the gain of the essential noise suppression property can be adjusted to 1, taking into account the positive efficiency and the inverse efficiency of the worm gear when the deceleration mechanism associated with the motor includes the worm gear.Since the efficiency of the worm gear is approximately 0.8, it is necessary to adjust the gain by ±0.2 with respect to the target value of 1.

[0124] In the above embodiment, if the transfer function of the low-pass filter is Q(s) and the transfer function of the high-pass filter is HPF(s), the complementary sensitivity function is a function represented by Q(s)·HPF(s). However, the present invention is not limited to this. For example, in a case where the transfer function of the control target is equal to the transfer function of the nominal model, the complementary sensitivity function can be represented by Q(s).

[0125] The nominal model can be a model with any transfer function. For example, the transfer function P can be... n (s) of the nominal model can be expressed by the following formula (7), where the control target comprises the engine as an inertia-one system. Pn(s)=1Jmns+Bmn

[0126] In formula (7), s is a Laplace transformer, J mn is a parameter that represents the moment of inertia of the nominal model, and B mn is a parameter that represents the viscosity friction coefficient of the nominal model.

[0127] The corrective torque generated by the support control unit could be any torque that corrects the input torque, and the input torque could be corrected in any way. The support control unit could correct the input torque using the corrective torque to be controlled, which is not feedback control, such as feedforward control.

[0128] The control target of the control device and control method in the present disclosure could be any part of the steering mechanism, as long as the control target includes at least one motor. The control target could be a control target of an inertial-one system, a control target of an inertial-two system, or more. Description of the reference symbols 100 Control device 210 Reaction force control unit 230 Support control unit 232 Low-pass filters 233 High-pass filters 250 Friction compensation value calculation unit 260 interference compensation value calculation unit 280 State feedback unit 521 Steering wheel 524a Input shaft 524b Output wave 530 Steering mechanism 543 Engine 546 Torsion bar 560 Tax target 1000 electric power steering device Cf1 first cutoff frequency Cf2 second cutoff frequency P(s) Transfer function of the control target Pn(s) Transfer function of the nominal model T(s) complementary sensitivity function T d Differential torque T f Correction torque T h Steering torque (torsion bar torque) T p torque T r Input torque V d interference compensation value V f Friction compensation value V s State compensation value θa Rotation angle

Claims

[1] Control device (100) which controls as a control target (560) at least a section with a motor (543) in a steering mechanism (530) which has an input shaft (524a) to which a steering wheel (521) is connected which is steered by a steering person, an output shaft (524b) which is connected to the input shaft (524a) via a torsion bar (546) and the motor (543) which is connected to the output shaft (524b), wherein the control device (100) has the following features: a reaction force control unit (210) which provides an input torque (T r ), which is entered into the control target (560), based on a torsion bar torque (T h ) generated in the torsion bar (546) and controls a reaction force that is transmitted from the steering wheel (521) to the steering person; a support control unit (230) which provides a correction torque (T f) to correct the input torque (T r ) generated based on an output of the tax target (560) and a nominal model; and a state feedback unit (280) which provides a state compensation value (V) s ) based on the output of the control target (560) to the input torque (T r ) traces back to, characterized by , that the support control unit (230) is designed such that a transfer function P(s) of the control target (560) is bounded by a transfer function P n(s) of the nominal model in a frequency band in which a gain for a gain property of a complementary sensitivity function with respect to a modeling error (Δ(s)) between the control target (560) and the nominal model is approximately 1, wherein the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit (230), wherein the support control unit (230) is designed such that the control target (560) is controlled such that the transfer function P(s) of the control target (560) appears as a transfer function P n (s) of the nominal model appears to be when considering the input / output relationship, where the transfer function P n (s) of the predetermined nominal model is given by the following formula (3) Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is, and the state feedback unit (280) the state compensation value (V s ) so based on the output of the control target (560) such that an apparent transfer function P(s) of the control target (560) corresponds to a transfer function P n(s) approximates the nominal model, where the apparent transfer function P(s) of the control target (560) is a transfer function of an entire section from a subtractor (SU1) to an output of the control target (560) and is a transfer function of a section combining the state feedback unit (280) and the control target (560). [2] Control device (100) according to claim 1, wherein the state feedback unit (280) determines the state compensation value (V s ) to the input torque (T r ) returns after the input torque (T r ) by the correction torque (T f ) was corrected and before the input torque (T r ) is entered into the tax target (560). [3] Control device (100) according to claim 1 or 2, wherein the state compensation value (V) s) has a compensation value that compensates at least part of an inertial force generated in the control target (560), a viscosity force generated in the control target (560), and a friction force generated in the control target (560). [4] Control device (100) which controls as a control target (560) at least a section with a motor (543) in the steering mechanism (530) which has an input shaft (524a) to which a steering wheel (521) is connected which is steered by a steering person, an output shaft (524b) which is connected to the input shaft (524a) via a torsion bar (546) and the motor (543) which is connected to the output shaft (524b), wherein the control device (100) has the following features: a reaction force control unit (210) which provides an input torque (T r ), which is entered into the control target (560), based on a torsion bar torque (T h) generated in the torsion bar (546) and controls a reaction force that is transmitted from the steering wheel (521) to the steering person; a support control unit (230) which provides a correction torque (T f ) to correct the input torque (T r ) generated based on an output of the tax target (560) and a nominal model; and a state feedback unit (280) which provides a state compensation value (V) s ) to the input torque (T r ) returns, where the state compensation value (V s ) at least part of an inertial force generated in the control target (560), a viscosity force generated in the control target (560), and a friction force generated in the control target (560), characterized by, that the support control unit (230) is designed such that a transfer function P(s) of the control target (560) is bounded by a transfer function P n (s) of the nominal model in a frequency band in which a gain for a gain property of a complementary sensitivity function with respect to a modeling error (Δ(s)) between the control target (560) and the nominal model is approximately 1, wherein the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit (230), wherein the support control unit (230) is designed such that the control target (560) is controlled such that the transfer function P(s) of the control target (560) appears as a transfer function P n (s) of the nominal model appears to be when considering the input / output relationship, where the apparent transfer function P(s) of the control target (560) is a transfer function of an entire section from a subtractor (SU1) to an output of the control target (560) and is a transfer function of a section that combines the state feedback unit (280) and the control target (560), where the transfer function P n (s) of the predetermined nominal model is given by the following formula (3) Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1na damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is, and the state feedback unit (280) the state compensation value (V s ) to the input torque (T r ) returns after the input torque (T r ) by the correction torque (T f ) was corrected and before the input torque (T r ) is entered into the tax target (560). [5] Control device (100) according to one of claims 1 to 4, wherein the support control unit (230) has the following features: a high-pass filter (233) with a first cutoff frequency; and a low-pass filter (232) with a second cutoff frequency that is higher than the first cutoff frequency, and If the transfer function of the low-pass filter is Q(s) and the transfer function of the high-pass filter is HPF(s), the complementary sensitivity function is expressed by Q(s)·HPF(s). [6] Control device (100) according to claim 5, wherein one order of the low-pass filter (232) is 3 or more. [7] Control device (100) according to one of claims 1 to 6, wherein the state compensation value (V) s ) has a compensation value to compensate for at least part of an inertial force generated in the motor, a viscosity force generated in the motor, and a friction force generated in the motor. [8] Control device (100) according to one of claims 1 to 7, wherein the support control unit (230) controls the correction torque (T f) based on a difference between a torque calculated using the nominal model based on an output of the control target (560) and the input torque (T r ) generated before the state compensation value (V s ) is returned after being subjected to the correction torque (T f ) was corrected. [9] Control device (100) according to one of claims 1 to 8, wherein the support control unit (230) controls the correction torque (T f ) based on a rotation angle of the input wave (524a). [10] Control device (100) according to any one of claims 1 to 9, wherein the support control unit (230) has a disturbance compensation value calculation unit (260) which calculates a disturbance compensation value to compensate for at least part of a restoring torque generated in the control target (560), and the correction torque (T f) exhibits the interference compensation value. [11] Control device (100) according to claim 10, wherein the disturbance compensation value comprises a compensation value for compensating at least part of a frictional force generated in the control target (560), a disturbance torque caused by backlash generated in the control target (560), and a torque ripple generated in the control target (560). [12] Control device (100) according to claim 10 or 11, wherein the support control unit (230) is configured to: Includes a friction compensation value calculation unit that calculates a friction compensation value to compensate for at least part of a frictional force generated in the control target (560) based on a differential torque that represents a difference between the input torque (T r) and a torque calculated using the nominal model based on an output of the control target (560); and Generating the correction torque (T f ) by adding the friction compensation value and the disturbance compensation value to the differential torque from which a frequency component lower than a first cutoff frequency has been removed by a high-pass filter with the first cutoff frequency. [13] Control device (100) according to claim 12, wherein the friction compensation value calculation unit (250) calculates the friction compensation value based on a component with a frequency lower than the first cutoff frequency in which the differential torque is calculated, and The disturbance compensation value calculation unit (260) calculates the disturbance compensation value based on a component of the differential torque with a frequency that is lower than the first cutoff frequency. [14] Electric power steering device (1000) comprising the following features: the control device (100) according to any one of claims 1 to 13; and the steering mechanism (530). [15] Control method for steering, as a control target (560), at least one section with a motor (543) in a steering mechanism (530) comprising an input shaft (524a) to which a steering wheel (521) is connected which is steered by a steering person, an output shaft (524b) which is connected to the input shaft (524a) via a torsion bar (546) and the motor (543) which is connected to the output shaft (524b), wherein the control method comprises the following steps: Generating an input torque (T r ), which is entered into the control target (560), based on a torsion bar torque (T h), which is generated in the torsion bar (546), and controlling a reaction force transmitted from the steering wheel (521) to the steering person; Generating a corrective torque (T) f ) to correct the input torque (T r ) based on an output of the tax target (560) and a nominal model; and Reducing a state compensation value (V) s ) to the input torque (T r ) based on the output of the tax target (560), characterized by , that the generation of the correction torque T f a restriction of a transfer function P(s) of the control target (560) to a transfer function P n (s) of the nominal model in a frequency band in which a gain for a gain property of a complementary sensitivity function T(s) with respect to a modeling error (Δ(s)) between the control target (560) and the nominal model is approximately 1, wherein the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit (230), wherein the support control unit (230) is designed such that the control target (560) is controlled such that the transfer function P(s) of the control target (560) appears as a transfer function P n (s) of the nominal model appears to be when considering the input / output relationship, where the transfer function P n (s) of the predetermined nominal model is given by the following formula (3) Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where s is a Laplace transformer, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1na frequency at a zero point of the transfer function P n ( S ) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1n a damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is, and the return of the state compensation value (V s ) to the input torque (T r ) a return of the state compensation value (V s) based on an output of the control target (560) in such a way that an apparent transfer function P(s) of the control target (560) approximates a transfer function of the nominal model, wherein the apparent transfer function P(s) of the control target (560) is a transfer function of an entire section from a subtractor (SU1) to an output of the control target (560) and is a transfer function of a section combining the state feedback unit (280) and the control target (560). [16] Control method for steering, as a control target (560), at least one section with a motor (543) in a steering mechanism (530) which is provided with an input shaft (524a) to which a steering wheel (521) is connected which is steered by a steering person, an output shaft (524b) which is connected to the input shaft (524a) via a torsion bar (546) and the motor (543) which is connected to the output shaft (524b), wherein the control method comprises the following steps: Generating an input torque (T r ), which is entered into the control target (560), based on a torsion bar torque (T h ), which is generated in the torsion bar (546), and controlling a reaction force transmitted from the steering wheel (521) to the steering person; Generating a corrective torque (T) f ) to correct the input torque (T r) based on an output of the tax target (560) and a nominal model; and Return to the input torque (T) r ), a state compensation value (V s ) to compensate for at least part of an inertial force generated in the control target (560), a viscosity force generated in the control target (560), and a friction force generated in the control target (560), characterized by , that the generation of the corrective torque (T f ) a restriction of a transfer function P(s) of the control target (560) by a transfer function P n(s) of the nominal model in a frequency band in which a gain of a gain property of a complementary sensitivity function T(s) with respect to a modeling error (Δ(s)) between the control target (560) and the nominal model is approximately 1, wherein the gain property is a complementary sensitivity function T(s) of an inner loop formed by the support control unit (230), wherein the support control unit (230) is designed such that the control target (560) is controlled such that the transfer function P(s) of the control target (560) appears as a transfer function P n (s) of the nominal model appears to be when considering the input / output relationship, where the apparent transfer function P(s) of the control target (560) is a transfer function of an entire section from a subtractor (SU1) to an output of the control target (560) and is a transfer function of a section that combines the state feedback unit (280) and the control target (560), where the transfer function P n (s) of the predetermined nominal model is given by the following formula (3) Pn(s)=1JSTGns+BSTGns2+2ζ1nω1ns+ω1n2s2+2ζ2nω2ns+ω2n2 where its Laplace transformer is, J STGn a parameter that represents a moment of inertia of the nominal model, B STGn a parameter that represents a viscosity friction coefficient of the nominal model, ω 1n a frequency at a zero point of the transfer function P n (s) is, ω 2n a frequency of one pole of the transfer function P n (s) is, ζ 1na damping ratio at the zero point of the transfer function P n (s) and ζ 2n a damping ratio at the pole of the transfer function P n (s) is, and the return of the state compensation value (V s ) to the input torque (T r ) a return of the state compensation value (V s ) to the input torque (T r ) exhibits after the input torque (T) r ) by the correction torque (T f ) was corrected and before the input torque (T r ) is entered into the tax target (560).

Images