JP5272949B2 - Vehicle vibration suppression control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the structure of a vibration damping control device in order to cope with a malfunction concerning a phenomenon of a waveform peak of a value calculated by a model computing element in the vibration damping control device by wheel torque control or drive/brake force control. <P>SOLUTION: This vibration damping control device for a vehicle for suppressing the vibration of a vehicle body by controlling wheel torque includes a compensation component determination section calculating a compensation component which is superposed on a wheel torque required value in order to reduce the vibration amplitude of the vehicle body based on the vehicle body vibration model of the vehicle and compensates the wheel torque. The compensation component determination section includes the model computing element for calculating the compensation component. When the magnitude of the output value of the model computing element reaches a predetermined value, the compensation component calculated based on the output value of the model computing element reaching the predetermined value is not superposed on the wheel torque required value. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a vibration damping control device for a vehicle such as an automobile, and more specifically, controls a torque (hereinafter referred to as “wheel torque”) that acts between a vehicle wheel and a grounding road surface. The present invention relates to a vibration suppression control device for suppressing vehicle body vibration such as pitch / bounce vibration of a vehicle body.

  Vehicle vibrations such as pitch and bounce vibrations while the vehicle is running are generated by braking / driving force (or inertial force) acting on the vehicle body during acceleration / deceleration of the vehicle or other external force acting on the vehicle body. It is reflected in the wheel torque. Therefore, in the field of vehicle vibration suppression control, it has been proposed to adjust the wheel torque through vehicle braking / driving force control to suppress vibration of the vehicle body while the vehicle is running (for example, Patent Documents). 1 and 2). In vibration suppression control of vehicle body vibration by such wheel torque or braking / driving force control, when fluctuations in wheel torque due to vehicle acceleration / deceleration requests or turning requests are expected, or external force (unevenness or foreign matter on the road surface, A so-called vehicle body when the wheel torque changes due to the action of a change in road reaction force due to a change in the state of the vehicle on the road surface such as a change in gradient or friction, or a mechanical disturbance such as a force such as wind. Predicts pitch and bounce vibrations generated in the vehicle body and suppresses the predicted vibrations using a motion model of vehicle body vibration that is built on the assumption of a dynamic model of sprung vibration or sprung / unsprung vibration Thus, the torque or braking / driving force of the wheel is adjusted. In the case of this type of vibration suppression control, the generation of vibration energy is suppressed by adjusting the source of the force that generates vibration, rather than by absorbing vibration energy generated as in the case of vibration suppression control by the suspension. Therefore, the vibration damping action is relatively quick and the energy efficiency is good. Further, in the vibration damping control as described above, since the object to be controlled is concentrated on the wheel torque or the braking / driving force of the wheel, the control adjustment is relatively easy.

  In the vibration damping control by the wheel torque control or the braking / driving force control as described above, typically, first, the wheel corresponding to the predicted value of the wheel torque corresponding to the acceleration / deceleration request or the turning request of the vehicle or the external force acting on the vehicle body. The actual value or estimated value of the torque is input to a computing unit (model computing unit) configured based on the vehicle body vibration model, where the displacement in the pitch direction around the center of gravity of the vehicle and / or the bounce direction of the center of gravity and A predicted value of the rate of time change (hereinafter, these values are referred to as “pitch / bounce vibration state values”) is calculated. Then, an actuator that executes wheel torque control or braking / driving force control so that each of the calculated pitch / bounce vibration state values converges to 0, that is, a vehicle driving device such as an engine or a motor, a braking device, A compensation component (a correction amount of the wheel torque) for the wheel torque or braking / driving force is given to the steering device as a control command. The predicted value of the pitch bounce vibration state value is calculated by solving the differential equation of the pitch bounce vibration state value derived from the balance of force and moment (including wheel torque) acting on the vehicle body in the pitch bounce direction. The model computing unit incorporates an integrator for solving such a differential equation numerically sequentially (see the column of the embodiment).

JP 2004-168148 A JP 2006-69472 A

  Regarding a model computing unit that solves the above differential equation and calculates a predicted value of the pitch bounce vibration state value, if an excessive amplitude is output from the model computing unit, an excessive amplitude compensation component is correspondingly generated. There may be a case where the wheel torque or the braking / driving force is given to an actuator. Therefore, the model computing unit is generally provided with a configuration so that the amplitude of the output value is limited within a predetermined range, that is, the output value is “guarded”. Is. However, when the output value of the model computing unit is guarded, the waveform of the output value reaches its peak, and the accuracy of the compensation component calculated from the output value is deteriorated. In addition, since the calculation in the integrator included in the model calculator is executed in such a manner that the sequentially calculated values are accumulated, the integration calculation is continued while the guard is applied to the output value. Then, an error remains or accumulates in the predicted value of the pitch bounce state value given by the output value of the model calculator or integrator.

Moreover, the phenomenon of the waveform peak of the output value of the model calculator as described above also occurs when the value calculated by the calculation exceeds the capacity digit of the model calculator, that is, when the calculated value overflows. Since the model computing unit and the integrator inside the model computing unit used in the vibration suppression control device are typically configured by a digital computer, the numerical values that can be processed in the model computing unit have the number of digits in advance. limited. For example, if the 16-bit digital computer, the number of representable numbers, for each of the positive and negative ^directions, a ^{2 15} (32768). In other words, if the resolution of the value calculated by the model calculator is set to 1/8192, the numerical value range that can be output is ± 4. If the calculated value deviates from this range, an overflow occurs and the calculation is disabled. It becomes possible. Therefore, if the resolution is set finely in an attempt to calculate the calculated value with high accuracy, when the absolute value of the calculated value increases, the waveform peaks, and this is the same as when the above output value is guarded. In addition, the accuracy of the compensation component may be deteriorated, and errors such as remaining or accumulation of errors in the output value of the model computing unit (particularly including an integrator) may occur.

  Thus, the main problem to be solved by the present invention is a problem related to the phenomenon of the peak of the waveform of the output value or the calculated value of the model computing unit in the vibration damping control device by wheel torque control or braking / driving force control. In order to cope with this, it is to improve the configuration of the vibration suppression control device.

  According to the present invention, there is provided a vibration suppression control device for a vehicle that suppresses vehicle body vibration such as pitch bounce vibration of the vehicle by controlling wheel torque generated at a ground contact point between the vehicle wheel and a road surface. When the output value of the model computing unit used to calculate the compensation component for compensating the wheel torque so as to reduce the vibration state value or the vibration amplitude of the vehicle body has or is likely to occur In addition, there is provided a vibration damping control device configured such that such “waveform peak” is not reflected in a compensation component for vibration damping control applied to an actuator that controls wheel torque or braking / driving force.

  First, the vibration suppression control device of the present invention calculates a compensation component for compensating for the wheel torque superimposed on the required value of the wheel torque so as to reduce the vibration amplitude of the vehicle body based on the vehicle body vibration model of the vehicle. A compensation component determination unit is included. That is, the vibration suppression control device of the present invention basically has a vehicle body pitch / frequency that can be generated by a driver or a braking / driving request (or turning request) by automatic driving control or a disturbance acting on the vehicle body while the vehicle is running. It is a vibration damping control device that compensates for wheel torque so as to reduce or cancel body vibration such as bounce. Typically, the compensation component calculated by the compensation component determining unit is a vehicle such as an engine or a motor. Superimposed on the required value of the drive torque given to the drive device (or the required value for the braking device or steering device of the vehicle), thereby reducing or eliminating the component that causes the vehicle body vibration included in the required value of the drive torque Alternatively, the drive torque is controlled in a direction that cancels the action of the component (vibration force) that causes the vehicle body vibration in the disturbance acting on the vehicle body, and the wheel torque in the wheel is compensated.The compensation component determination unit includes a model computing unit for calculating the compensation component. Typically, the model computing unit uses the integrator to calculate the vibration displacement of the vehicle body based on the vehicle body vibration model of the vehicle. The differential value is predicted, and the compensation component may be determined based on the predicted vibration displacement of the vehicle body and the differential value.

  However, as already described, when the absolute value of the output value of the model computing unit increases and the output value is guarded, or when the output value deviates from the operable range of the model computing unit (overflow). Therefore, the waveform of the compensation component calculated from the output value of the model computing unit (typically, the predicted value of the vibration displacement of the vehicle body and its differential value) will occur. Will also be deformed. Then, the frequency and phase of the wheel torque actually generated in response to the compensation component is different from the frequency and phase of the wheel torque planned by the vibration suppression control, or can be amplified rather than suppressing the vibration of the vehicle body. There may also be cases where frequency components occur in the wheel torque.

  Therefore, in the vibration suppression control apparatus of the present invention, the magnitude of the output value of the model calculator is predetermined in order to avoid components other than the frequency and phase planned by the vibration suppression control being input to the wheel torque control. When the value is reached, the compensation component calculated based on the output value of the model computing unit that has reached the predetermined value is not superimposed on the required value of the wheel torque. Here, the “predetermined value” is the absolute value of the upper limit value and the lower limit value of the guard or model calculation for avoiding that a compensation component having an excessive absolute value is superimposed on the required value of the wheel torque. It may be an absolute value of a limit value of the numerical value range that can be calculated or a value that is set slightly smaller than the numerical value range that can be calculated by the model arithmetic unit so that the output value of the device does not exceed the numerical value range that can be calculated.

  As one specific embodiment when the magnitude of the output value of the model computing unit reaches a predetermined value, the control of the wheel torque for suppressing the vehicle body vibration of the vehicle may be stopped. Thereby, it becomes possible to avoid the malfunction of control with a simple configuration. [For example, the output value of the model calculator is one of the predicted values of the pitch bounce vibration state value, and the “predetermined value” is set to the maximum value of the pitch bounce state value that can be assumed in the normal vibration state. If the output value exceeds the predetermined value, the model computing unit does not accurately predict the pitch bounce vibration state value. Therefore, in that case, the failure due to inaccurate vibration suppression control can be avoided by stopping the control of the wheel torque that suppresses the vehicle body vibration of the vehicle. ]

  In particular, when the model computing unit is a digital computer capable of computing a numerical value represented by a predetermined number of bits, another specific example when the output value of the model computing unit reaches a predetermined value As an embodiment, when the magnitude of the output value of the model calculator reaches a predetermined value determined by the predetermined number of bits, the numerical value width per bit of the model calculator is increased, that is, the model calculator The resolution of the calculation may be coarsened.

  As will be understood by those skilled in the art, when a numerical operation is executed by a model arithmetic unit constituted by a digital computer, a numerical value width assigned per bit, that is, a numerical value value that can be resolved by the model arithmetic unit. Is set in accordance with the accuracy required for the output value and the numerical value expected to be output. In order to calculate the output value with higher accuracy, the numerical value width per bit may be set to be small so that the resolution of the calculation value in the model calculator is fine. However, the number of bits of the digital computer used for the model calculator is usually fixed, and there are restrictions on the upper and lower limit values that can be expressed by the computer. If the absolute value of the value is exceeded, the model computing unit overflows and cannot be computed. Therefore, in order to avoid overflow, it is necessary to set a larger numerical value width per bit. However, when this happens, the resolution of the output value becomes coarse, and a highly accurate output value cannot be obtained. In fact, in the vibration suppression control subject to the present invention, an output value deviating from the normally assumed value range may be output from the model computing unit due to various factors, and such an unusual absolute value is large. If the resolution of the model calculator is set in preparation for the value, the accuracy of the control will be reduced. In addition, the model computing unit may include an integrator. In that case, since the computation of the integrator is executed in a mode of accumulating values sequentially, even temporarily, When a value deviating from the normally assumed value range occurs and overflow occurs, an error at the time of overflow remains in values calculated thereafter.

  Therefore, the apparatus according to the above aspect of the present invention may cause the output value of the model calculator to overflow when the magnitude of the output value of the model calculator reaches a predetermined value determined by a predetermined number of bits. At some point, it is configured to increase the numerical width per bit of the model operator, thereby avoiding overflow. In this case, the “predetermined value” is determined so as to be a maximum value of a numerical value that can be calculated by the model calculator or a value having a smaller absolute value. By increasing the numerical value width per bit of the model calculator, the output value of the model calculator falls within the range that can be calculated by the model calculator. The compensation component calculated based on the output value is not superimposed on the required value of the wheel torque. In addition, since overflow is avoided, there is no longer any error in the overflow in the output value, so even after the output value deviating from the normally assumed value range is output from the model computing unit ( Although the accuracy is reduced), the vibration suppression control may be continued.

  Further, in the above configuration, the numerical value width per bit of the model computing unit may be increased as the output value of the model computing unit increases. In this case, since the resolution is changed depending on the magnitude of the output value, the model calculation can be executed accurately and without overflowing in a wide amplitude range.

  By the way, in the above-described series of embodiments, the model computing unit includes an integrator, and the magnitude of the output value of the model computing unit reaches a predetermined value. Is calculated based on the output value of the model computing unit that has reached a predetermined value by stopping the control of the wheel torque that suppresses or by increasing the numerical value range per bit of the model computing unit When the vehicle wheel speed differential value becomes substantially zero over a predetermined period after the compensation component is not superimposed on the required value of the wheel torque, the output value of the integrator included in the model calculator May be set to 0, and the superimposition of the compensation component calculated on the basis of the output value of the model calculator to the required value of the wheel torque may be resumed. When the wheel speed differential value becomes substantially zero over a predetermined period (for example, when the vehicle stops), the vehicle body vibration is substantially extinguished and the output value of the integrator is forced to zero. It is considered that the state may be reset automatically. Accordingly, by setting the output value of the integrator to 0, an error that may remain in the integrator is removed, and the superimposition of the compensation component on the required value of the wheel torque is resumed to execute the vibration damping control. It may be.

  As described above, when the model computing unit for calculating the compensation component of the compensation component determining unit includes an integrator, the computation of the integrator is executed in a manner in which values are sequentially accumulated. Therefore, once an error enters the integral value, the error continues to remain and deteriorates the accuracy of the compensation component. Therefore, in the vibration suppression control device of the present invention, as another aspect, when the vehicle stops or not regardless of whether the output value of the model computing unit or the integrator reaches a predetermined value or When the wheel speed differential value of the vehicle becomes substantially zero over a predetermined period, the output value of the integrator may be set to 0, and the error that has entered the integral value may be removed.

  In general, according to the present invention, in the vibration damping control device based on wheel torque control or braking / driving force control, the “topped” waveform in the output value of the model calculator used for calculating the compensation component and As a result, it is avoided that the compensation component including the error is input to the actuator that controls the wheel torque or the braking / driving force, and thus, the magnitude of the output value of the model calculator is excessively controlled. Problems due to the inability to generate the frequency and phase of the wheel torque planned by the control are avoided. In particular, according to the aspect in which the resolution of the model computing unit is made variable in accordance with the output value of the model computing unit, the amplitude range of vibration that can be suppressed by the damping control is expanded compared to the conventional case. It is advantageous.

  Other objects and advantages of the present invention will become apparent from the following description of preferred embodiments of the present invention.

FIG. 1 shows a schematic diagram of an automobile in which a preferred embodiment of a vibration damping control device according to the present invention is realized. FIG. 2 shows the internal configuration of the embodiment of the electronic control device of FIG. 1 in the form of a control block diagram. Various parameters other than those shown in the figure, such as engine temperature, may be input to the drive torque request value determination unit 51 and the control command determination unit 53. FIG. 3A is a diagram illustrating a state variable of vehicle body vibration that is suppressed in the operation of vibration suppression control according to a preferred embodiment of the present invention. FIG. 3B is a view for explaining a “sprung vibration model” which is one of the mechanical motion models of the vehicle body vibration assumed in the vibration damping control according to the preferred embodiment of the present invention, and FIG. It is a figure explaining an upper and unsprung vibration model. FIG. 3D is a diagram showing the configuration of the compensation component determination unit in the preferred embodiment of the present invention in the form of a control block diagram. 4A is a diagram showing the configuration of the model calculator in the compensation component determination unit in FIG. 3D in the form of a more detailed control block diagram, and FIG. 4B is a control block diagram showing the configuration of the integrator in the model calculator. FIG. FIG. 5 is a diagram showing, in the form of a flowchart, processing for changing the control mode in accordance with the output value of the model computing unit (integrator) in the compensation component determination unit of FIG.

  The present invention will now be described in detail with reference to a few preferred embodiments with reference to the accompanying drawings. In the figure, the same reference numerals indicate the same parts.

Diagram 1 of the apparatus is a vehicle such as an automobile in which the preferred embodiment of the vibration damping control device is mounted of the present invention is schematically shown. In the figure, the vehicle 10 having the left and right front wheels 12FL and 12FR and the left and right rear wheels 12RL and 12RR is driven in the normal manner according to the depression of the accelerator pedal 14 by the driver. A drive device 20 that generates torque is mounted. In the illustrated example, as illustrated in FIG. 1, the driving device 20 has a driving torque or a rotational driving force from the engine 22 via the torque converter 24, the automatic transmission 26, the differential gear device 28, and the like. It is configured to be transmitted to the rear wheels 12RL and 12RR. However, the drive device may be an electric drive in which an electric motor is used instead of the engine 22 or a hybrid drive device having both an engine and an electric motor. The vehicle may be a four-wheel drive vehicle or a front wheel drive vehicle. Although not shown for simplicity, the vehicle 10 includes a braking system device that generates a braking force on each wheel in response to the depression of the brake pedal 16 and the steering angle of the front wheels or the front and rear wheels, as in a normal vehicle. A steering device for controlling the motor is provided.

  The operation control of the driving device 20 and the operation control of other parts of the vehicle are controlled by the electronic control device 50. The electronic control unit 50 may include a microcomputer having a CPU, a ROM, a RAM, and an input / output port device, which are connected to each other by a bidirectional common bus, and a driving circuit. The electronic control unit 50 includes a signal Vwi (i = FL, FR, RL, RR) representing a wheel speed from a wheel speed sensor 30i (i = FL, FR, RL, RR) mounted on each wheel, a vehicle Signals such as the engine rotation speed ne, the transmission rotation speed no, the accelerator pedal depression amount θa, and the brake pedal depression amount θb are input from sensors provided in these parts (see FIG. 1). In addition to the above, it should be understood that various detection signals for obtaining various parameters necessary for various controls to be executed in the vehicle of the present embodiment may be input.

  The vibration damping control device of the present invention is realized in the electronic control device 50 described above. FIG. 2 shows the internal configuration of an embodiment of such an electronic control device 50 in the form of a control block. Referring to the figure, an electronic control unit 50 includes a drive control unit 50a for controlling the operation of the engine, a braking control unit 50b for controlling the operation of a braking unit (not shown), and a known electronic control unit for a vehicle. May include various control devices (not shown). It is understood that the configuration and operation of various control devices such as a drive control device including a vibration suppression control device are realized by processing operations of the CPU and the like in the electronic control device 50 during operation of the vehicle. Should.

  As shown in the figure, the braking control device 50b includes a pulse-type electric power that is sequentially generated every time the wheel rotates by a predetermined amount from the wheel speed sensor 30i (i = FR, FL, RR, RL) of each wheel. The wheel rotational speed ω is calculated by measuring the time interval between arrival of such sequentially input pulse signals, and the wheel radius value r · ω is multiplied by the wheel rotational speed r. Is calculated. Then, the wheel speed value r · ω is transmitted to the drive control device 50a (wheel torque estimator 52c) and used for calculation of the wheel torque estimated value in order to execute vibration suppression control which will be described in detail later. . The calculation from the wheel rotation speed to the wheel speed may be performed by the drive control device 50a. In this case, the wheel rotation speed is given from the braking control device 50b to the drive control device 50a.

  As a basic configuration, the drive control device 50a has a drive torque request value determining unit 51 that determines the engine drive torque request value requested by the driver based on the accelerator pedal depression amount or the accelerator opening θa from the accelerator pedal sensor. A compensation component determining unit 52 that calculates a compensation component for executing vehicle body pitch / bounce vibration damping control by wheel torque (drive torque) control to compensate (correct) the drive torque request value, and the compensation component Based on the drive torque request value compensated by the compensation component calculated by the determination unit, a control command for each part of the engine is generated in an arbitrary known format and a corresponding controller (not shown) is achieved. Control command determination unit 53 to be transmitted to

  In such a basic configuration, the drive torque request value determination unit 51 may determine and output a drive torque request value (before compensation) corresponding to the accelerator opening θa by any known method. . Note that the “accelerator opening” means the amount of depression or operation of the accelerator pedal by the driver of the vehicle, or in the case of a vehicle equipped with an automatic travel control device (not shown). It is an amount that represents the required amount of drive torque or drive output, and represents the required amount of acceleration / deceleration force or braking / driving torque for the vehicle. The unit of the drive torque request value may be basically the drive torque in the engine, but in the case of a gasoline engine, the intake air amount or throttle opening, in the case of a diesel engine, the fuel injection amount, the motor If so, it may be a current value (hereinafter, unless otherwise specified, the unit of the drive torque request value is the drive torque in the engine).

  As shown in the figure, the compensation component determination unit 52 converts the drive torque request value (before compensation) determined by the drive torque request value determination unit 51 into a wheel torque (wheel torque request value), and wheel torque estimation. 52c receives the estimated value of the wheel torque actually acting on the wheel estimated from the wheel speed r · ω, and the vehicle vibration model (required wheel torque request value and estimated value) according to a mode described in detail later. Using a model calculator that outputs the vibration displacement in the pitch / bounce direction of the vehicle body as an input (torque input), the vibration component that can cause pitch / bounce vibration in the vehicle body at the wheel torque required value and estimated value A compensation component (K · X) to be reduced or canceled is calculated. When the wheel torque Tw is input from the wheel torque estimator 52c, the wheel torque Tw is used to adjust the balance of contribution between the driver-requested wheel torque Tw0 and the estimated wheel torque Tw in the vehicle body vibration model. The feedback control gain (input gain) λin may be multiplied and then input to the compensation component determination unit (multiplier 52d). Still further, the compensation component determination unit may calculate a compensation component for damping pitch / bounce vibration caused by a change in wheel torque generated in the wheel by a driver's brake operation or steering operation. In that case, as indicated by a dotted line in the figure, a wheel torque estimated value estimated based on the brake operation amount or the steering operation amount by the wheel torque estimator 52x is input to the compensation component determination unit, and the wheel The compensation component is calculated in the same manner as the torque request value and the like. The estimation of the change amount of the wheel torque based on the brake operation amount or the steering operation amount may be performed by an arbitrary method.

  Thus, the compensation component (K · X) calculated by the compensation component determination unit 52 is converted into a unit of the drive torque request value (compensation component U), transmitted to the adder a1, and is sent to the adder a1. Thus, the drive torque request value is compensated by superimposing the compensation component on the drive torque request value (before compensation) (in the illustrated example, the compensation component U is subtracted from the drive torque request value). ). When the compensation component (K · X) is output from the compensation component determination unit 52, a multiplier 52f that multiplies the compensation component K · X by the control gain λout to adjust the contribution of the compensation component for an arbitrary purpose. (That is, the compensation component is sent to the adder a1 in the state of λout · K · X). Then, the control command determination unit 53 is experimentally or theoretically determined in advance based on the compensated drive torque request value with reference to the engine speed and / or the engine temperature at that time. Using a map, in a known manner, control commands to drive units (not shown) of each part of the engine are determined and control commands are transmitted to the drive units so as to achieve drive torque. .

  In the configuration of the compensation component determination unit 52 described above, as will be described in more detail later, the compensation component is the vibration state value of the vehicle body (vibration displacement and its time differential value) calculated in the vehicle body vibration model. Determined from the predicted value. In such a configuration, when the amplitude of the predicted value of the vibration state value of the vehicle body increases, and as a result, the amplitude of the compensation component increases, the engine drive output vibrates correspondingly. Therefore, in order to avoid such excessive vibration of the drive output, a guard is usually provided to prevent a value exceeding a predetermined limit value from being output from the compensation component determination unit 52 or the vehicle body vibration model. That is, when the vibration state value of the vehicle body or the value of the compensation component reaches a predetermined limit value, the change in the output value reaches a peak at the predetermined limit value, and a control command that gives an excessive drive output is sent to each part of the engine. Is transmitted to the other driver. However, when the vibration state value of the vehicle body or the peak of the compensation component occurs, the waveform of the compensation component is deformed, so that good vibration suppression control is not achieved, or due to a change in the waveform, the compensation component There is a possibility that the frequency characteristic or the phase characteristic is changed, and an unexpected frequency / phase component is generated in the drive output.

  Further, as already mentioned, the model computing unit constituting the vehicle body vibration model in the compensation component determining unit 52 is constituted by a digital computer having a predetermined number of bits, and the number of digits of the processable number (capacity digit) is previously set. However, when the absolute value of the calculated value (vibration state value) of the vehicle body vibration model becomes large, the computer becomes an overflow state, and the calculated value reaches a peak. In this case as well, the waveform of the compensation component is deformed in the same manner as in the above case, so that good vibration suppression control is not achieved.

  Therefore, in the control device of the present invention, in addition to the above basic configuration, the vibration state value or the compensation component value of the vehicle body that has reached its peak at the predetermined limit value is not reflected in the drive torque request value. Alternatively, a configuration is provided so as not to superimpose, thereby attempting to avoid excessive frequency and phase components from being generated in the drive output while avoiding excessive vibration due to vibration suppression control. Some of its specific aspects are described in more detail later.

Outline of Device Operation Hereinafter, the configuration and operation of vibration suppression control executed by the device illustrated in FIG. 2 will be described. It should be understood that the vibration suppression control device is realized with the compensation component determination unit 52 as a main component in the drive control device in the illustrated example. The pitch / bounce vibration suppression control using the compensation component calculated by the compensation component determination unit 52 may be performed in the following manner.

(Principle of vibration suppression control)
In the vehicle, when the driving device is activated based on the driving request of the driver and the wheel torque fluctuates, the vertical position of the center of gravity Cg of the vehicle body in the vehicle body 10 as illustrated in FIG. The bounce vibration in the direction (z direction) and the pitch vibration in the pitch direction (θ direction) around the center of gravity of the vehicle body can occur. In addition, if a force or torque (disturbance) acts on wheels due to changes in road surface conditions or wind while the vehicle is running, the disturbance is transmitted to the vehicle, and vibrations in the bounce direction and pitch direction are also generated in the vehicle body. obtain. Therefore, in the pitch / bounce vibration damping control exemplified here, a motion model (body vibration model) of the body pitch / bounce vibration is constructed, and the required drive torque value is converted into wheel torque in that model. Vehicle body displacement z and θ and their rate of change dz / dt and dθ / dt when the calculated value (wheel torque request value) and / or the current wheel torque estimation value are input, that is, the state variable (vibration state) Value) and the driving torque of the driving device (engine) is adjusted so that the state variable obtained from the model converges to 0, that is, the pitch / bounce vibration is suppressed (the driving torque request value is Compensated.)

ここに於いて、Ｌｆ、Ｌｒは、それぞれ、重心から前輪軸及び後輪軸までの水平方向距離であり、ｒは、車輪半径であり、ｈは、重心の路面からの高さ（即ち、重心の車高）である。なお、式（１ａ）に於いて、第１、２項は、前輪軸から、第３、４項は、後輪軸からの力の成分であり、式（１ｂ）に於いて、第１項は、前輪軸から、第２項は、後輪軸からの力のモーメント成分である。式（１ｂ）に於ける第３項は、駆動輪に於いて発生する車輪トルクＴが車体の重心周りに与える力のモーメント成分である。 Thus, first, as a dynamic motion model in the bounce direction and the pitch direction of the vehicle body in the vibration suppression control, for example, as shown in FIG. 3B, the vehicle body is a rigid body S having a mass M and an inertia moment I. It is assumed that the rigid body S is supported by a front wheel suspension having an elastic modulus kf and a damping rate cf, and a rear wheel suspension having an elastic modulus kr and a damping rate cr (vehicle body sprung vibration model). In this case, the motion equation in the bounce direction and the motion equation in the pitch direction of the center of gravity of the vehicle body are expressed as the following Equation 1.

Here, Lf and Lr are respectively horizontal distances from the center of gravity to the front wheel shaft and the rear wheel shaft, r is a wheel radius, and h is the height of the center of gravity from the road surface (that is, the center of gravity of the center of gravity). Vehicle height). In the equation (1a), the first and second terms are components of the force from the front wheel shaft, the third and fourth terms are components of the force from the rear wheel shaft, and in the equation (1b), the first term is From the front wheel shaft, the second term is the moment component of the force from the rear wheel shaft. The third term in the equation (1b) is a moment component of the force that the wheel torque T generated in the drive wheel gives to the periphery of the center of gravity of the vehicle body.

であり、行列Ａの各要素a1-a4及びb1-b4は、それぞれ、式（１ａ）、（１ｂ）のｚ、θ、ｄｚ／ｄｔ、ｄθ／ｄｔの係数をまとめることにより与えられ、
a1=-(kf+kr)/M、a2=-(cf+cr)/M、
a3=-(kf・Lf-kr・Lr)/M、a4=-(cf・Lf-cr・Lr)/M、
b1=-(Lf・kf-Lr・kr)/I、b2=-(Lf・cf-Lr・cr)/I、 …（２ｂ）
b3=-(Lf²・kf+Lr²・kr)/I、b4=-(Lf²・cf+Lr²・cr)/I
である。また、ｕ(t)は、
ｕ(t)＝Ｔ
であり、状態方程式（２ａ）にて表されるシステムの入力である。従って、式（１ｂ）より、行列Ｂの要素p1は、
p1=h/(I・r)
である。 The above formulas (1a) and (1b) are expressed as (linear) as shown in the following formula (2a) with the vehicle body displacements z and θ and their change rates dz / dt and dθ / dt as the state variable vector X (t). It can be rewritten in the form of a system state equation.
dX (t) / dt = A · X (t) + B · u (t) (2a)
Here, X (t), A, and B are respectively

And each element a1-a4 and b1-b4 of the matrix A is given by combining the coefficients of z, θ, dz / dt, dθ / dt in the equations (1a) and (1b), respectively.
a1 =-(kf + kr) / M, a2 =-(cf + cr) / M,
a3 =-(kf ・ Lf-kr ・ Lr) / M, a4 =-(cf ・ Lf-cr ・ Lr) / M,
b1 =-(Lf.kf-Lr.kr) / I, b2 =-(Lf.cf-Lr.cr) / I, (2b)
b3 =-(Lf ²・ kf + Lr ²・ kr) / I, b4 =-(Lf ²・ cf + Lr ²・ cr) / I
It is. U (t) is
u (t) = T
And is an input of the system represented by the state equation (2a). Therefore, from equation (1b), the element p1 of the matrix B is
p1 = h / (I ・ r)
It is.

In the equation of state (2a)
u (t) = − K · X (t) (2c)
Then, the equation of state (2a) is
dX (t) / dt = (A−BK) · X (t) (2d)
It becomes. Accordingly, the initial value X ₀ (t) of X (t) is set as X ₀ (t) = (0,0,0,0) (assuming that there is no vibration before torque is input). ), The gain that converges the magnitude of X (t), that is, the displacement in the bounce direction and the pitch direction and its time change rate, to 0 when the differential equation (2d) of the state variable vector X (t) is solved When K is determined, a torque value u (t) for suppressing pitch bounce vibration is determined. A value obtained by converting the torque value u (t) into an engine drive torque request value is a compensation component given to the engine by vibration suppression control.

The gain K can be determined by using a so-called optimal regulator theory. According to this theory, a quadratic evaluation function J = 1/2 · ∫ (X ^T QX + u ^T Ru) dt (3a)
(Integral range is 0 to ∞)
When the value of is the minimum, the matrix K that minimizes the evaluation function J by the stable convergence of X (t) in the state equation (2a) is
K = R− ¹ · B ^T · P (3b)
It is known to be given by Where P is the Riccati equation
^{-dP / dt = A T P +} PA + Q-PBR -1 B T P
Is the solution. The Riccati equation can be solved by any method known in the field of linear systems, which determines the gain K.

などと置いて、式（３ａ）に於いて、状態変数ベクトルの成分のうち、特定のもの、例えば、ｄｚ／ｄｔ、ｄθ／ｄｔ、のノルム（大きさ）をその他の成分、例えば、ｚ、θ、のノルムより大きく設定すると、ノルムを大きく設定された成分が相対的に、より安定的に収束されることとなる。また、Ｑの成分の値ｑ１〜ｑ４を大きくすると、過渡特性重視、即ち、状態変数ベクトルの値が速やかに安定値に収束し、Ｒの値ρを大きくすると、消費エネルギーが低減される。 Q and R in the evaluation function J and Riccati equation are a semi-positive definite symmetric matrix and a positive definite symmetric matrix, respectively, which are arbitrarily set, and are weight matrices of the evaluation function J determined by the system designer. For example, in the case of the motion model considered here, Q and R are

In Equation (3a), a specific one of the components of the state variable vector, for example, the norm (magnitude) of dz / dt, dθ / dt is set to other components, for example, z, If the value is set larger than the norm of θ, components having a large norm are converged relatively stably. Further, when the Q component values q1 to q4 are increased, the transient characteristics are emphasized, that is, the value of the state variable vector quickly converges to a stable value, and when the R value ρ is increased, energy consumption is reduced.

ここに於いて、xf、xrは、前輪、後輪のばね下変位量であり、mf、mrは、前輪、後輪のばね下の質量である。式（４ａ）−（４ｂ）は、ｚ、θ、xf、xrとその時間微分値を状態変数ベクトルとして、図３（Ｂ）の場合と同様に、式（２ａ）の如き状態方程式を構成し（ただし、行列Ａは、８行８列、行列Ｂは、８行１列となる。）、最適レギュレータの理論に従って、状態変数ベクトルの大きさを０に収束させるゲイン行列Ｋを決定することができる。 As a dynamic motion model in the bounce direction and the pitch direction of the vehicle body, for example, as shown in FIG. 3C, in addition to the configuration in FIG. 3B, the spring elasticity of the front and rear tires A model that takes into account the above (vehicle body sprung / lower vibration model) may be employed. Assuming that the front and rear tires have the respective elastic moduli ktf and ktr, the motion equation in the bounce direction and the motion equation in the pitch direction of the center of gravity of the vehicle body are as understood from FIG. The following equation 4 is expressed.

Here, xf and xr are unsprung displacement amounts of the front and rear wheels, and mf and mr are unsprung masses of the front and rear wheels. Equations (4a)-(4b) form a state equation as shown in Equation (2a) using z, θ, xf, xr and their time differential values as state variable vectors, as in FIG. 3B. (However, the matrix A has 8 rows and 8 columns and the matrix B has 8 rows and 1 column.) According to the theory of the optimal regulator, the gain matrix K that converges the magnitude of the state variable vector to 0 can be determined. it can.

(Estimation of wheel torque)
As the torque input T for the vehicle body vibration model in FIGS. 3B and 3C, among the input torques, the currently generated wheel torque value Tw input as a disturbance action is an ideal value. Specifically, a torque sensor may be provided on each wheel and actually detected, but it is difficult to provide a torque sensor on each wheel of a normal vehicle. Therefore, in the illustrated example, the wheel torque estimated value estimated by the wheel torque estimator 52c (FIG. 2) from other detectable values in the running vehicle is used as the disturbance input of the wheel torque. The wheel torque estimated value Tw is typically obtained by using a wheel rotational speed ω obtained from a wheel speed sensor of a driving wheel or a time derivative of a wheel speed value r · ω,
Tw = M · r ² · dω / dt (5)
Can be estimated. Here, M is the mass of the vehicle, and r is the wheel radius. [If the sum of the driving forces generated at the contact points of the driving wheels on the road surface is equal to the overall driving force MG (G is acceleration) of the vehicle, the wheel torque Tw is
Tw = M · G · r (5a)
Given in The acceleration G of the vehicle is obtained from the differential value of the wheel speed r · ω,
G = r · dω / dt (5b)
Therefore, the wheel torque is estimated as shown in Equation (5). Note that the estimated wheel torque value is not estimated from the wheel speed, but from the rotational speed of the rotating shaft of the drive system operatively connected to the drive wheels, such as the engine rotational speed, the transmission rotational speed, and the turbine rotational speed. It may be. When the rotational speed ne of the output shaft of the engine or motor of the driving device is used, the wheel rotational speed of the driving wheel is
ωe = ne x transmission (transmission) gear ratio x differential (differential gear) gear ratio (6)
Given by. When using the rotational speed no of the output shaft of the transmission,
ωo = no x differential gear ratio (7)
Given by. Then, the estimated value of the wheel rotational speed ω of the drive wheel in Expression (6) or (7) is substituted into Expression (5), and the estimated wheel torque value is calculated.

(Schematic configuration of compensation component determination unit)
FIG. 3D schematically shows a basic configuration of control processing inside the compensation component determination unit 52 of FIG. 2 for calculating the compensation component U for the pitch / bounce vibration suppression control. In the control configuration of FIG. 3D, first, the wheel torque request value obtained by converting the drive torque request value from the drive torque request determination unit 51 into the wheel torque to the wheel torque input end of the vehicle body vibration model. Two and estimated wheel torque Tw (currently estimated value) Tw that is actually generated in the wheel are respectively input (in addition, as indicated by the dotted line in the figure, the estimated wheel torque value corresponding to the brake operation amount or the steering operation amount is May be entered). Next, the model calculator calculates the state variable vector X (t) by solving the differential equation (2a) using the torque input value T (= Two + Tw). Then, a value K · X (= −u (t)) obtained by multiplying the state variable vector X (t) by the gain K determined to converge the state variable vector X (t) to 0 or the minimum value as described above. ) Is calculated, and K · X is converted into a compensation component U in units of engine drive torque request values. The compensation component calculated in this way is transmitted to the adder a1, and is superposed on the drive torque request value in the adder a1, so that the component corresponding to the value of K · X (t) becomes the drive torque request value. Will be deducted from. As can be understood from FIG. 3B or 3C, the pitch bounce vibration system of the vehicle body is a resonance system, and the value of the state variable vector X (t) is substantially equal to an arbitrary input. Is only a frequency component of a band (usually about 1 to 5 Hz) having a certain spectral characteristic centered around the natural frequency of the system. Accordingly, by subtracting K · X (t) from the drive torque request value as described above, the natural frequency component of the system of the drive torque request value or the currently generated wheel torque, that is, in the vehicle body. Thus, components that cause pitch bounce vibration are reduced or eliminated, and pitch bounce vibration in the vehicle body is suppressed.

Improvement of the Model Calculator As described above, the model calculator as described above is composed of a digital computer having a predetermined number of bits and has a predetermined number of capacity digits. That is, as the resolution of the model computing unit, that is, the numerical value width per bit is made finer, the accuracy of the calculated value is improved. However, the numerical value range that can be processed becomes narrow correspondingly, and overflow easily occurs. Once the overflow occurs, the amplitude of the calculated value reaches its peak, and the vibration waveform is deformed. The frequency that is not planned in the damping control is included in the compensation component sent as a control command to the engine. Component and phase components are generated, and the model computing unit includes an integrator (typically, the output value of the integrator is the element value of the state variable vector as it is. ), And subsequent calculation results are inaccurate. Therefore, the resolution of the model calculator needs to be determined in consideration of the accuracy required for the numerical value calculated in the model calculation and the range of numerical values expected to appear in the model calculator. .

  However, regarding the range of numerical values that appear in the model calculator, the value of each element (vibration state value) of the state variable vector calculated by the model calculator is the result of the estimation calculation from the wheel torque estimated value. Therefore, an unrealistically excessive amplitude may be calculated (for example, when the wheel speed changes suddenly, the tire slips while driving on a road surface with a low road surface friction coefficient such as a snowy road, and the wheel locks. Or it can happen when you spin the wheel.) Assuming that such an excessively large state variable vector can be generated, if the resolution of the model computing unit is set so as to avoid overflow even with such a value, vibration of realistic amplitude is suppressed. As the accuracy of the power compensation component deteriorates, the compensation component whose amplitude has become excessive due to the excessive amplitude of each element of the state variable vector is transmitted to the engine control command, and the engine drive output vibrates excessively. There is a risk. A state calculated in the model calculator as one of the measures for avoiding excessive vibration of the engine drive output while dealing with such a failure, that is, while maintaining the accuracy of the compensation component for normal vibration Each element or compensation component of the variable vector is guarded so that the value of each element of the state variable vector and the value of the compensation component do not deviate from a predetermined range, for example, a realistic vibration range. However, in the case of a method of providing a guard for the calculated value in the model computing unit, the fluctuation of the value is limited by the boundary value of the predetermined range, and eventually the calculated value reaches a peak state and is controlled. In the vibration control, an unplanned frequency component / phase component is generated. Moreover, if the calculated value with the guard applied is used for the calculation of the integrator, the calculation result becomes inaccurate.

  Therefore, in the apparatus of the present invention, while maintaining the accuracy of the calculated value of the model computing unit as much as possible, it is avoided that an excessive amplitude compensation component is given as an engine control command, and it is peaked by a guard or overflow. The configuration of the model calculator is improved so that the calculated value of the model calculator is not reflected in the compensation component. Hereinafter, a preferred mode of improvement of the model calculator will be described.

(Detailed configuration of model calculator)
FIG. 4A shows a case where a sprung vibration model represented by the formula (1a)-(1b) or (2a)-(2b) illustrated in FIG. 3B is adopted as the vehicle body vibration model. This is a model computing unit in which the configuration further including the improvement according to the present invention is shown in more detail in the form of a control block diagram. Referring to the figure, in the model computing unit shown in the drawing, generally speaking, each element of time derivative dX / dt of state variable vector X is integrated, and the value of each element of the corresponding state variable vector is integrated. Integrators 100a, 100b, 100c, and 100d that calculate the values, matrix calculator 100e that executes multiplication A · X of state variable vector X and matrix A, and multiplication B · T of wheel torque T and matrix B are executed. And a multiplier 100g. In this configuration, as basic processing, first, the matrix calculator 100e and the multiplier 100g, as can be understood from the figure, show the element values [z, dz / dt, θ, dθ / dt] or receiving the wheel torque T, and finally d ² z / dt ² expressed in the left side of the equations (1a) and (1b), d ² θ / dt ² is calculated here (in this case, the value from the matrix calculator 100e is used for the next calculation cycle, and thus the previous value given by the delay units 100a ′ and b ′). is there.). These calculated values are respectively applied to the terminals i2 of the integrators 100a and 100b and integrated to calculate dz / dt and dθ / dt. Further, these calculated values are respectively calculated by the integrators 100c and 100d. Given to terminal i2 and integrated, z and θ are calculated. Then, the calculated element values [z, dz / dt, θ, dθ / dt] of the state variable vector are sent to the matrix calculator 100e, and at the same time, the compensation component is a multiplication of the state variable vector X and the gain matrix K. It is also transmitted to the matrix calculator 100f that calculates K · X. Each of the integrators 100a, 100b, 100c, and 100d has a configuration as shown in FIG. 4B. In each of the integrators 100a, 100b, 100c, and 100d, each of the integrators is basically received from the input i2. A value obtained by multiplying a time differential value [d ² z / dt ² , dz / dt, d ² θ / dt ² , dθ / dt] of each element of the variable vector X by a minute time dt in the multiplier 102a; , The integrated value of the product of each time differential value accumulated from the delay unit 102d so far and the minute time dt is added in the adder 102b, whereby each element of the state variable vector X [ dz / dt, z, dθ / dt, θ] are calculated and output through the switch sw1. Usually, the model computing unit repeatedly executes the above-described processing, thereby sequentially outputting the element value of the state variable vector while solving the differential equation of Equation (2a), and the compensation component K · X is Will be output.

  However, in the apparatus of the present invention, a guard 102c is further provided at the output terminal o1 of each integrator described above, and when the absolute value of the integral value reaches a predetermined value in the guard, A guard ON signal is emitted from the output terminal o2. When the guard ON signal is generated from at least one of the integrators in the OR computing unit 100h in FIG. 4A, vibration suppression control is performed according to any of several methods described later. Or the mode of calculation processing of the model calculator. Hereinafter, an example of changing the mode when the guard ON signal is issued will be described.

Mode when the guard ON signal is issued
Example 1
In one aspect, at least one of the elements of the state variable vector output from the model computing unit (integrator thereof) has reached an absolute value and a guard ON signal has been issued. Sometimes, the vibration suppression control may be stopped. Specifically, when the output of the OR calculator 100h is turned on, the gain λout of the control gain 52e in FIG. 2 is set to 0, thereby preventing the compensation component from being superimposed on the drive torque request value. You may be supposed to be. The predetermined value set for the guard may be set to the maximum value of the displacement that can occur in an actual vehicle and the magnitude of the time change rate of the displacement, and the resolution of the model calculator is the maximum value of the model calculator. It may be set so as to be expressed by the limit of the number of digits of the capacity of the digital computer to be configured. That is,
Predetermined value ← [Numeric value width per bit] × [Number of bits corresponding to the number of capacity digits−1]
(-1 corresponds to subtracting the number of bits assigned to the positive and negative signs). When the vibration suppression control is stopped, the calculation in the model calculator may or may not be executed. According to such a configuration, the compensation component calculated from the element value (vibration state value) of the state variable vector that has reached a predetermined value is not superimposed on the drive torque request value, and an unexpected frequency component / It is avoided that the phase component control command is given to the engine. The return of the vibration suppression control may be performed according to Example 5 or 6 below.

Example 2
In another embodiment, at least one of the elements of the state variable vector output from the model computing unit (integrator thereof) reaches its predetermined value, and a guard ON signal is generated. When this is done, the resolution of the model calculator may be coarsened, that is, the numerical value width per bit may be increased to expand the numerical range that can be processed by the model calculator. Specifically, when the output of the OR calculator 100h is turned ON, first, a multiplier 102e indicated by a dotted line on the output side of the guard 102 of all the integrators (FIG. 4B). , The calculated value is multiplied by 1 / ξ to reduce the absolute value (ξ is a positive constant). Then, in the multiplier 100j (FIG. 4A) indicated by the dotted line on the output side of the matrix calculator 100f, the calculated value is multiplied by ξ, and the absolute value of the output value is restored. It is. In subsequent cycles, in the multiplier 100i (FIG. 4A) indicated by the dotted line provided on the input side of the multiplier 100g, the wheel torque input value is multiplied by 1 / ξ. Thus, the absolute value is reduced, and the multiplier 100j multiplies the calculated value by ξ to restore the absolute value of the output value (the operation of the multiplier 102e turns on the output of the OR operator 100h). It is executed only immediately after, and not after that.) The predetermined value set for the guard may be set to the maximum value of the displacement that normally occurs in an actual vehicle and the magnitude of the time change rate of the displacement. (The maximum value may be set to a value slightly smaller than the value corresponding to the limit of the number of capacity digits).

  According to such a configuration, when at least one of the output values of the integrator reaches a predetermined value, the number of used bits of the calculated value transmitted in the model computing unit is reduced. Even if the magnitude of the rate of change exceeds the maximum value that would normally occur, the possibility of overflow of the model calculator will be reduced (although the accuracy of the calculated vibration state value will be reduced), so it will reach its peak. It is expected that the compensation component calculated from the calculated value will not be superimposed on the drive torque request value. Also, before at least one of the elements of the state variable vector output from the integrator reaches a predetermined value, the model calculator is designed to achieve the desired accuracy in the displacement that occurs normally and the rate of change of the displacement over time. This is advantageous in that the resolution can be set. Note that when the output value of the multiplier 100j deviates from the range of the displacement that can occur in the actual vehicle and the time change rate range of the displacement, for example, the value of λout of the control gain 52e is adjusted by an arbitrary method. Thus, the compensation component having an excessive amplitude may not be superimposed on the drive torque request value. [The accuracy of the compensation component superimposed on the drive torque request value may be lower than the calculation accuracy of the value calculated by the model calculator. Since the calculation in the model calculator includes an integration operation and numerical values are accumulated, errors are likely to accumulate or remain, and therefore it is desirable that the calculation accuracy be as high as possible. On the other hand, the accuracy of the control command converted from the drive torque request value and transmitted to each part of the engine does not require the accuracy of the value in the model calculator. Therefore, in general, when the compensation component is output from the compensation component determination unit, the accuracy is lowered (after the resolution is coarsened) and then superimposed on the drive torque request value. That is, by making the resolution coarse downstream from the output end of the matrix calculator 100f, the calculated value of the multiplier 100j is prevented from overflowing downstream of the signal. Note that the resolution can be restored according to Example 5 or 6 below.

Example 3
As a further development of the example 2 above, the resolution of the model calculator is set to be coarser as the absolute value of the element value of the state variable vector calculated by the model calculator increases. In this case, the numerical range that can be processed may be expanded. Specifically, in the guard 102c of each integrator, when the number of bits used for the integral value reaches a predetermined number of bits (corresponding to a predetermined value), the multiplier 102e is used in the manner described in Example 2. After reducing the number of bits used for the integral value (corresponding to reducing the absolute value of the integral value), the absolute value of the wheel torque input value is reduced by one step by the multiplier 100i and reduced by the multiplier 100j. Processing for restoring the absolute value of the compensation component is performed. Thereafter, every time the number of bits used for the integral value reaches a predetermined number of bits, the number of bits used for the integral value is reduced by the multiplier 102e, and then the absolute value of the wheel torque input value is already obtained by the multiplier 100i. one step (if reaching a predetermined number of bits of the number of available bits of the integral value is n-th, multiplied by 1 / xi] ⁿ⁾ reduced, based on the absolute value of the reduced compensation component by the multiplier 100j The returning process is executed (multiply by ξ ⁿ ). According to such a configuration, the resolution of the model computing unit is changed according to the magnitude of the absolute value of the value calculated by the model computing unit, and the possibility of occurrence of overflow in the model computing unit can be reduced. It becomes possible. Further, with the variable resolution configuration, the resolution of the model computing unit is set finely at the beginning. Therefore, even if the numerical range that can be processed by the model computing unit is relatively narrow (that is, the predetermined value in the guard 102c). Even if the value is relatively small), it is possible to continue the estimation of the vibration state value without causing an overflow. However, once the resolution is coarsened, the accuracy of the calculated value in the model computing unit is lowered. Therefore, when the resolution is restored again, it is preferable to follow the following Example 5 or 6.

Example 4
Further, as another aspect, a sensor capable of detecting displacement in the pitch / bounce direction of the vehicle body, such as a vertical G sensor for an apparatus such as an electronically controlled suspension in a vehicle, or other means capable of detecting displacement in the pitch / bounce direction (others) If the displacement of the pitch bounce direction is estimated by the above method.) Is installed, at least one of the elements of the state variable vector output from the model calculator (the integrator) When the absolute value reaches a predetermined value and a guard ON signal is generated, each element value of the state variable vector is determined using the value obtained from the sensor or other means as described above, and The compensation component K · X may be calculated by using (substituting the state value). [In the case of a sensor detection value such as a vertical G sensor, a drift phenomenon generally occurs as time elapses due to a change in temperature or the like, and this reduces the accuracy of the detection value when used for a long time. To do. Therefore, the use of a configuration that uses each element value of a state variable vector obtained from a sensor or other means should be limited. According to such a configuration, the compensation component calculated from the vibration state value that has reached its peak at the predetermined value is not superimposed on the drive torque request value, and an unexpected control command for the frequency component / phase component is generated in the vibration suppression control. It will be avoided to be given to. The return of the vibration suppression control using the output value of the model calculator may be performed according to Example 5 or 6 below.

By the way, as already mentioned, the integration operation of each integrator is performed in such a manner that the calculated values are accumulated sequentially, so that once an error enters the calculated value, the error will continue thereafter. If the error continues and an error occurs continuously, the error will be accumulated (according to simulation experiments, after the wheel torque input has changed from a certain significant value to 0, the vibration state value becomes 0. It has been found that it takes a few seconds to converge.) Therefore, as shown in FIGS. 4A and 4B, each integrator is given a reset signal to the terminal i1 when several conditions to be described later are satisfied, whereby the switch sw1 is turned on. The output value may be switched to 0 and reset (the output of the delay device 102d is set to 0). According to such a configuration, when each integrator is reset, errors accumulated or remaining in the calculated values so far are eliminated, and the vibration suppression control can be resumed with high accuracy. In addition, after the absolute value reaches a predetermined value and a guard ON signal is generated in at least one of the elements of the state variable vector output from the model computing unit, the above-described examples 1 to 4 are performed. As described, when the mode of calculation processing of the model computing unit or the mode of vibration suppression control is changed, the accuracy of the calculated value in the model computing unit is reduced. It is not preferable that the state returns. Therefore, when the changes described in Examples 1 to 4 are made, a reset signal is issued, and after the integrator is reset, the state before the change may be restored. Hereinafter, examples of conditions for generating the reset signal will be described.

Example 5
Pitch bounce vibration does not occur when the vehicle is stopped. Therefore, as one aspect, when the vehicle stops, a reset signal for forcibly resetting the integrator may be issued. Regarding the determination of the stop of the vehicle, for example, when the vehicle speed becomes 0 for 0.5 to 1 second, it may be determined that the vehicle has stopped.

Example 6
As understood from the above description, when the wheel torque is substantially zero, no vehicle body vibration should have occurred. Therefore, as another aspect, when the state where the wheel speed differential value = 0 continues for a predetermined time (for example, several seconds), a reset signal for forcibly resetting the integrator may be issued.

Flow of control processing Thus, in the device of the present invention, when at least one of the elements of the state variable vector output from the model calculator reaches a predetermined value and a guard ON signal is generated. In addition, when the control mode is changed in any one of Examples 1 to 4 and any of the conditions (return conditions) in Examples 5 to 6 is satisfied, the process of returning the control mode to the state before the change is performed. Executed. FIG. 5 shows such processing in the form of a flowchart. Referring to the figure, in this process, first, whether or not the control mode has already been changed, that is, whether or not the vibration suppression control is stopped (in the case of Example 1), and whether the resolution is being changed. It is determined whether or not (in the case of Examples 2 and 3), or whether or not the value obtained by another sensor or means is used as the state value (in the case of Example 4) (Step 10). When the mode has not been changed, it is determined whether or not any of the magnitudes of the vibration state values deviates from a predetermined value (that is, whether or not a guard ON signal has been issued) (step 20). ). If the magnitude of any vibration state value is within the predetermined range, normal vibration suppression control (no change in the mode of Examples 1 to 4) is executed (step 50). On the other hand, when one of the magnitudes of the vibration state values deviates from the predetermined value, the control mode is changed (vibration control stop, resolution change or state value substitution) (step 30). Further, when the control mode change in step 30 is executed, it is determined after the determination in step 10 whether or not the return condition is satisfied (step 40). If the return condition is not satisfied, When the changed control mode is continued and the return condition is satisfied, the calculated value of the integrator so far is reset to 0 (step 45), and the normal vibration suppression control is returned (step 50).

  Although the above description has been made in relation to the embodiment of the present invention, many modifications and changes can be easily made by those skilled in the art, and the present invention is limited to the embodiment exemplified above. It will be apparent that the invention is not limited and applies to various devices without departing from the inventive concept.

  For example, although the wheel torque estimated value in the above embodiment is estimated from the wheel speed, the wheel torque estimated value may be estimated from parameters other than the wheel speed. In the above-described embodiment, the vibration suppression control is based on the assumption that a sprung vibration model or a sprung / unsprung vibration model is used as a motion model of vehicle body vibration, and the vibration control of pitch / bounce vibration using the theory of an optimal regulator. As long as the concept of the present invention uses wheel torque, the concept of the present invention is to adopt any vehicle body by using a motion model of vehicle body vibration other than the one introduced here or by a control method other than the optimal regulator. The present invention is also applied to a device that suppresses vibration, and such a case also belongs to the scope of the present invention. In particular, even when the sprung / unsprung vibration model illustrated in FIG. 3C is used, the improvement of the model calculator described in FIGS. 4A and 4B may be applied in substantially the same manner. It should be understood that such a case is also within the scope of the present invention.

DESCRIPTION OF SYMBOLS 10 ... Vehicle body 12FL, FR, RL, RR ... Wheel 14 ... Accelerator pedal 20 ... Drive device 30FL, FR, RL, RR ... Wheel speed sensor 50 ... Electronic control device 50a ... Drive control device 50b ... Braking control device

Claims (3)

A vibration suppression control device for a vehicle that suppresses vehicle body vibration of the vehicle by controlling wheel torque generated at a contact point between a wheel and a road surface, the vehicle body of the vehicle based on a vehicle body vibration model of the vehicle A compensation component determining unit that calculates a compensation component for compensating the wheel torque so as to be superimposed on a request value of the wheel torque so as to reduce a vibration amplitude of the vehicle, and for the compensation component determining unit to calculate the compensation component When the magnitude of the output value of the model calculator reaches a predetermined value, the compensation component calculated based on the output value of the model calculator that has reached the predetermined value is the wheel torque. The wheel torque control for suppressing the vehicle body vibration of the vehicle is stopped without being superimposed on the required value.   2. The apparatus according to claim 1, wherein the model computing unit includes an integrator, and after the magnitude of the output value of the integrator reaches the predetermined value, the wheel speed differential value of the vehicle is increased over a predetermined period. The output value of the integrator is set to 0, and the superimposition of the compensation component calculated based on the output value of the integrator on the required value of the wheel torque is resumed. A device characterized by that. 3. The apparatus according to claim 1 , wherein the model calculator is a model calculator that predicts a vibration displacement of the vehicle body and a differential value thereof based on a vehicle body vibration model of the vehicle using an integrator. A component is determined based on the predicted vibration displacement of the vehicle body and its differential value.

Images