JPH04238745A - Autonomous dispersed control device for vehicle - Google Patents

Abstract

PURPOSE:To provide a vehicle autonomous dispersed control device of performing vehicle motion control of wide range containing a limit region further with excellence in a trouble resistance property. CONSTITUTION:In each intelligent means 311 to 314 in a subsystem corresponding to four wheel steering mechanisms, based on ground force of each wheel from a force sensor of condition amount, steering angle from an operating amount detecting means and longitudinal lateral and yaw angular speeds of a car body from an action amount detecting means, respectively, dynamic characteristic of car body motion, followed by each subsystem, is calculated through a net work between the intelligent means in the above and the other intelligent means, and each steering mechanism is driven controlled by a drive means based on an obtained steering angle command. Thus by autonomously dispersedly controlling a steering angle of each wheel, a complicated vehicle motion can be controlled even to a limit region without linearly changing tire characteristic and also assigning roles by the other subsystem to obtain a normal car body motion even relating to a trouble of the single steering mechanism.

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device that autonomously and decentrally controls vehicle motion over a wide range including a limit region. 2. Description of the Related Art Conventionally, there is a technique described in Japanese Patent Application No. 2-108456 filed by the present applicant as a device for controlling vehicle motion by distributing it into a plurality of subsystems. In this conventional device, "a control amount detection means for detecting the amount of operation operated by the driver, a behavior amount detection means for detecting the amount of behavior of the vehicle body, and a plurality of subsystems consisting of one wheel or multiple wheels each have to follow." Total control means for calculating dynamic characteristics and outputting the dynamic characteristics as a command signal; State quantity detection means for detecting state quantities within the subsystem; and for the subsystem to obtain the dynamic characteristics given by the command signal. a sub-control means that calculates a control amount necessary for the control and outputs the control amount; and a drive means that drives the wheel mechanism based on the control amount. [0003] This is achieved by distributing vehicle motion, which is described by complex nonlinear differential equations that take into account the nonlinear characteristics of tires, etc., into multiple subsystems, so that each subsystem is a relatively simple nonlinear Focusing on differential equations, this method performs distributed and hierarchical control of vehicle motion without linear approximation. [0004] In this way, in this prior art, since the vehicle motion is not linearly approximated, it is possible to obtain desired vehicle characteristics in a wide range including the limit range. Additionally, since the controllers are distributed, the system has excellent fault tolerance. However, in this prior art,
The total control means plays the role of a coordinator that governs the behavior of the entire system, and is an essential component for the operation of the control system. Therefore, a failure of the total control means results in a failure of the entire system. [0006] Therefore, the present inventors discovered that by autonomously decentralizing control of each subsystem, it is possible to set the vehicle body motion to desired dynamic characteristics through cooperation among the subsystems without requiring command signals from the total control means. With this in mind, we have solved the problems in the prior art. SUMMARY OF THE INVENTION The present invention provides a system that performs a wide range of vehicle motion control including the limit range and has excellent fault tolerance. By the way, the above-mentioned conventional technology includes a plurality of sub-control means for controlling the motion of each wheel, and a system for sending command signals to the sub-control means to control the vehicle body motion determined by the motion of each wheel to desired characteristics. It had a hierarchical control configuration of total control means. For this reason, if the total control means, which is a higher-order system, fails, a command signal cannot be sent to the sub-control means, which is a lower-order system, and the entire vehicle movement will be disrupted. The present invention focuses on the problem that the above-mentioned conventional technology has a problem in that the control of each wheel movement is performed by a total control means, and distributes the role performed by the total control means to each sub-control means. The present invention is an autonomous decentralized control system for a vehicle that is capable of setting desired characteristics of vehicle body motion in a wide range of regions including limit regions, and has excellent fault tolerance, by each sub-control means autonomously controlling each wheel motion. For the purpose of providing. [Means for Solving the Problems] Operation amount detection means for detecting the operation amount of the driver, behavior amount detection means for detecting the amount of behavior of the vehicle body, and a plurality of sub-subs provided corresponding to the wheel mechanism. The subsystems are necessary for the vehicle body motion to obtain the desired dynamic characteristics through a network with other subsystems based on the detected operation amount and behavior amount. intelligent means that calculates the dynamic characteristics to be followed and outputs them as command signals; state quantity detection means that detects state quantities within the subsystem; and the detected state quantities and behavior quantities and the command signals from the intelligent means. a control means for calculating and outputting a control amount necessary for the subsystem to obtain dynamic characteristics as the command signal based on the above, and a drive means for driving the wheel mechanism based on the control amount from the control means. It is what it is. [Operation] The operation amount detection means detects the operation amount of the driver and outputs a signal corresponding to the detected operation amount. Note that this operation amount includes a steering amount, an accelerator operation amount, a brake operation amount, and the like. Further, the behavior amount detection means detects the amount of behavior of the vehicle body, which is the amount of motion state on the spring that is directly felt by the driver, and outputs a signal corresponding to the amount of behavior of the vehicle body. Next, the intelligent means within each subsystem determines the dynamic characteristics that each subsystem should follow in order to obtain the optimal vehicle motion based on the amount of operation.
Using vehicle characteristic equations determined from vehicle specifications and a mathematical model representing the target dynamic characteristics of vehicle motion, each calculation is performed via a network with other subsystems based on the amount of vehicle behavior, and output as a command signal to control means. do. Note that this optimal vehicle body motion is the vehicle body dynamic characteristic that is easiest for the driver to maneuver, and is set as the target dynamic characteristic of the vehicle body motion. Here, each subsystem is provided corresponding to a wheel mechanism. For example, in a normal four-wheeled vehicle, a total of four subsystems may be provided corresponding to each of the four wheel mechanisms, front, rear, left and right. Alternatively, a total of two subsystems may be provided for each of the two sets of wheel mechanisms, the front two left and right wheels and the rear two left and right wheels. Further, the state quantity detection means in each subsystem detects the state quantity in each subsystem and outputs a signal corresponding to the detected state quantity. Next, in the control means, the state quantity which is the motion state quantity of the subsystem and the sprung mass motion are determined from each subsystem so that the output of each subsystem follows the dynamic characteristic which is the command signal outputted from the intelligent means. Each control amount is calculated and output based on the behavior amount, which is the amount of influence on the system. Further, the drive means in each subsystem drives each wheel mechanism based on the control amount. [0014] As described above, the present invention independently controls each subsystem consisting of a wheel mechanism of one wheel or multiple wheels, and also coordinates with other subsystems by intelligent means, thereby controlling the amount of vehicle behavior. The present invention relates to an autonomous decentralized control device for a vehicle that sets a desired dynamic characteristic to a vehicle. Therefore, even if a failure occurs in some subsystems, the self-organizing function of the intelligent means allows other subsystems to share the role of the failed subsystem and adjust the amount of vehicle behavior to the desired dynamic characteristics. Can be set to . That is, the present invention has the effect of being extremely resistant to failure. [Embodiments] <Specific Examples of the Invention> More specific examples of the present invention will be described below with reference to FIGS. 1 to 3. In this specific example, the operation amount detection means 1, the behavior amount detection means 2, and subsystems S1 to S4 respectively corresponding to four wheel mechanisms are used.
Each subsystem consists of intelligent means 3 (31 to 34), state quantity detection means 4 (41 to 44)
, control means 5 (51-54), drive means 6
(61-64) Consists of. In the following description, the time differential of function A will be expressed as dA/dt, and the transposed matrix of matrix B will be expressed as BT. Furthermore, in each of the following examples, the configuration of each subsystem is the same, so unless otherwise specified, a common explanation will be given. The dynamic characteristics of the vehicle body are described by the following differential equation. dφ/dt=f(φ) +G(φ)・θ

… (1) However, G(φ) = [g1(φ), g2(φ), g3(φ
), g4(φ) ]
... (2) θ=[θ1, θ2, θ3, θ4]T

... (3) Here, φ: Vehicle behavior amount θi: Force applied to the vehicle body by the i-th wheel mechanism [0018
] Also, it is assumed that the desired dynamic characteristics of the vehicle body are given by the following differential equation. dφ/dt=f0(φ)

... (4) At this time, in order to obtain equation (4), each wheel mechanism needs to generate a force as shown in the following equation. θ=G(φ)+ ・(f0(φ) −f(φ)
)
... (5) However, G(φ)+: Pseudo-inverse matrix of G(φ) 0019
That is, by inputting θ calculated by equation (5) to each control means as a command signal, the amount of vehicle body behavior can be set to the desired dynamic characteristic expressed by equation (4). The intelligent means in each subsystem performs the calculation of equation (5) via a network with intelligent means in other subsystems, and is described by the following differential equation. Intelligent means 3 in the first subsystem S1
01: τ(dθ1/dt)=-g1(φ)T
g1(φ)・θ1 −g1(φ)T g2(φ)・
θ2
-g1(φ)T g3(φ)・θ3 -g1(φ)
T g4(φ)・θ4
+g1(φ)T ・(f0(φ)
-f(φ) )... (6) Intelligent means 302 in the second subsystem S2:
τ(dθ2/dt)=-g2(φ)T g1(φ
)・θ1 −g2(φ )T g2(φ)・θ2
−g2(φ
)T g3(φ)・θ3 −g2(φ )T g4(
φ)・θ4
+g2(φ)T ・(f0(φ) −f(φ)
)... (7) Intelligent means 303 in the third subsystem S3: τ(dθ
3/dt)=-g3(φ)T g1(φ)・θ1
-g3(φ)T g2(φ)・θ2
-g3(φ)T g
3(φ)・θ3 −g3(φ )T g4(φ)・θ4
+g
3(φ)T ・(f0(φ) −f(φ))
... (8) Intelligent means 304 in the fourth subsystem S4: τ(dθ4/dt
)=-g4(φ)T g1(φ)・θ1-g4(φ
)T g2(φ)・θ2
-g4(φ)T g3(φ)・
θ3 −g4(φ)T g4(φ)・θ4
+g4(φ)
T ・(f0(φ) −f(φ))
(9) [0022] However, τ is a coefficient that determines the speed of convergence of the network, and needs to be set sufficiently fast compared to the movement of the vehicle body. These differential equations, when the initial state is the origin, θ shown by equation (5)
After sufficient convergence, this θ is used as a command signal to the control means. FIG. 2 shows how the networks are connected. That is, for example, in a network connecting intelligent means 31 and intelligent means 33, g
1(φ) T g3(φ) Through the weight, θ1
and θ3 are exchanged with each other. This corresponds to the third term on the right-hand side of equation (6) and the first term on the right-hand side of equation (8), respectively, and the signals received via the network are used to construct differential equations in each intelligent means. used as an element. In this way, by finding equation (5) through a network of intelligent means, when a failure occurs in some subsystems, other subsystems can share the role of the failed subsystem, and Can self-organize to tolerate failures. For example, consider that the third subsystem fails, as shown in FIG. 3, and the output from this wheel mechanism becomes zero. At this time, the intelligent means of the other subsystems change as follows. Intelligent means 3 in the first subsystem S1
01: τ(dθ1/dt)=-g1(φ)T
g1(φ)・θ1 −g1(φ)T g2(φ)・
θ2
-g1(φ)T g4(φ)・θ4
+g1(φ)T ・
(f0(φ) −f(φ) ) …(1
0) Intelligent means 302 in the second subsystem S2: τ(dθ2/dt)=−g2(φ
)T g1(φ)・θ1 −g2(φ )T g2(φ
)・θ2
-g2(φ)T g4(φ)・θ4
+g2(φ)T
・(f0(φ) −f(φ) ) …
(11) Intelligent means 304 in the fourth subsystem S4: τ(dθ4/dt)=−g4(
φ )T g1(φ)・θ1 −g4(φ )T g2
(φ)・θ2
-g4(φ)T g4(φ)・θ4
+g4(φ
)T ・(f0(φ) −f(φ) )
...(12) [0026] θ1, θ2, and θ4 obtained through these networks calculate the amount of behavior of the vehicle body by the three subsystems excluding the third subsystem that has failed. This is a command signal for setting the characteristics. That is, the intelligent means self-organized from cooperation among four subsystems to cooperation among three subsystems due to a failure. The autonomous decentralized control device for a vehicle of the present invention exhibits excellent fault tolerance due to the intelligent means that self-organizes in this manner. <First Embodiment> The first embodiment is a vehicle in which the suspension force of the four wheels can be controlled, and each of the four wheel suspension mechanisms is made into a subsystem, the configuration of which is shown in FIG. 4. . (1) Operation amount detection means The operation amount detection means 10 includes a steering angle sensor 101 that detects the steering angle and converts it into a corresponding electric signal, and an accelerator that detects the operation amount of the accelerator and converts it into a corresponding electric signal. It consists of a sensor group 100 consisting of a brake operation amount sensor 102, a brake operation amount sensor 103 that detects the amount of brake operation and converts it into a corresponding electric signal, and an A/D converter 120 that digitizes these signals. ing. (2) Behavior amount detection means The behavior amount detection means 20 includes a longitudinal speed sensor 201 that detects the longitudinal speed of the vehicle body and converts it into a corresponding electric signal, and a longitudinal speed sensor 201 that detects the lateral speed of the vehicle body and converts it into a corresponding electric signal. a lateral speed sensor 202 that detects the vertical speed of the vehicle body and converts it into a corresponding electrical signal; a vertical position sensor 204 that detects the vertical position of the vehicle body and converts it into a corresponding electrical signal; A roll angular velocity sensor 205 detects the angular velocity and converts it into a corresponding electrical signal, a roll angle sensor 206 detects the roll angle of the vehicle body and converts it into a corresponding electrical signal, and a roll angle sensor 206 detects the pitch angular velocity of the vehicle body and converts it into a corresponding electrical signal. a pitch angular velocity sensor 207 that
A sensor group 200 consists of a pitch angle sensor 208 that detects the pitch angle of the vehicle body and converts it into a corresponding electrical signal, and a yaw angular velocity sensor 219 that detects the yaw angular velocity of the vehicle body and converts it into a corresponding electrical signal. It is composed of an A/D converter 220 for digitizing. (3) Intelligent means Intelligent means 30 (301 to 304) in each subsystem
is composed of a CPU 70 (701-704) as well as a control means 50 (501-504). C.P.
The intelligent means 301 in U701, the steering angle signal, the accelerator operation amount signal, and the brake operation amount signal output from the A/D converter 120 of the operation amount detection means 10 are input. The intelligent means 30 also receives longitudinal speed signals, lateral speed signals, vertical speed signals, vertical position signals, roll angular speed signals, roll angle signals, and pitch angular speed signals output from the A/D converter 220 of the behavior amount detection means 20. , pitch angle signal and yaw angular velocity signal are input. Based on the above input values, the intelligent means 301 calculates a target value for the suspension control force via a network with the intelligent means 302 to 304 in other subsystems, and outputs it as a command signal to the control means 501. Next, intelligent means 301 ~
An outline of the calculation in step 304 will be explained below. There is the following relationship between the posture of the vehicle body and the suspension control force. dφ/dt=f(φ) +G(φ)・θ
…(
13) However, G(φ) = [g1(φ), g2(φ), g3(φ
), g4(φ) ]
...(14) θ=[θ1, θ2, θ3, θ4]T

…(15) φ=[φ1, φ
2, φ3, φ4, φ5, φ6 ]T
…(16)
Here, θi : Suspension control force of the i-th suspension φ1 : Vertical position φ2 : Roll angle φ3 : Pitch angle φ4 : Vertical speed φ5 : Roll angular speed φ6 : Pitch angular speed 0032] Also, f(φ) = [mf1, mf2, mf3
, mf4, mf5, mf6]T
...(17) However, mf1 = -U sinφ3 +V sinφ
2 cosφ3 +φ4 cosφ2 cosφ
3 mf2 =φ5 +φ6 sinφ2
tanφ3 +R cosφ2 tanφ3 mf
3 =φ6 cosφ2 −R sinφ2 mf4
=Uφ6 −Vφ5 −g cosφ3 cosφ
2 mf5 = (Iy - Iz) Rφ6 /Ix -
Fy h/Ix mf6 = (Iz - Ix) Rφ5
/Iy +Fx h/Iy 0033] g1(φ) = [0,0,0,1,df,-af
]T...(1
8) g2(φ) = [0,0,0,1,-df,
-af]T...(
19) g3(φ) = [0,0,0,1,dr,
ar〕T
...(20) g4(φ) = [0, 0, 0, 1, -
dr, ar]T
...(21) [0034] However, af: Distance between front axle centers of gravity ar: Distance between rear axle centers of gravity df: Front suspension spacing/2 dr: Rear suspension spacing/2 Fx: Front-rear force Fy generated by driving and braking
: Lateral force generated by steering g : Gravitational acceleration h : Center of gravity height Ix : Roll moment of inertia Iy : Pitch moment of inertia Iz : Yaw moment of inertia m : Vehicle mass R : Yaw angular velocity U : Longitudinal speed V : Lateral speed [0035 Now, let us consider controlling the attitude of the vehicle body to have the dynamic characteristics expressed by the following equation. dφ/dt=f0(φ)

…(22) f0(φ) = [nf1
, nf2, nf3, nf4, nf5,
nf6]T...(23)
However, nf1 =-U sinφ3 +V sinφ
2 cosφ3 +φ4 cosφ2 cosφ
3 nf2 =φ5 +φ6 sinφ2
tanφ3 +R cosφ2 tanφ3 nf
3 =φ6 cosφ2 −R sinφ2 nf4
=-k4 φ4 -k1 φ1 nf5 =-k5
φ5 −k2 φ2 nf6 = −k6 φ6 −k3
φ3 [0036] In order to obtain the characteristics of the above equation, (1
7) It is necessary to substitute equations (21) to (23) into equations (6) to (9) and realize the converged values θ1 to θ4 as suspension control forces. Therefore, each intelligent means 301 ~
At step 304, θ1 to θ4 thus obtained are output as command signals. In addition, in this example,
Since the intelligent means is realized by a digital computer, the calculations of equations (6) to (9) are performed using discretized recurrence equations. After all, the intelligent means 301 to 304 in the i-th (i=1, 2, 3, 4) subsystem
The processing in is the flow shown below. First, the longitudinal force Fx generated at the four wheels is estimated based on the accelerator operation amount signal and brake operation amount signal inputted from the operation amount detection means using the following equation. In addition, the lateral force Fy generated at the four wheels is calculated by the following formula based on the steering angle signal input from the operation amount detection means, the longitudinal speed signal input from the behavior amount detection means, the lateral speed signal, and the yaw angular velocity signal. Estimated by. Fx=Kp・p

...(24) Fy = -Cf (
(V+af R)/U-δsw)-Cr (V-ar
R)/U (25) However, Kp: Constant Cf: Front wheel cornering power Cr: Rear wheel cornering power p: Accelerator operation amount (p>0) or brake operation amount (p<0) δsw: Steering angle (however, the actual Rudder angle equivalent value)0040
Next, 0 is assigned as the initial value of θi and output to intelligent means in other subsystems. Then,
Estimated longitudinal force Fx and lateral force Fy,
A longitudinal speed signal, a lateral speed signal, a vertical speed signal, a vertical position signal, a roll angular speed signal, a roll angle signal, a pitch angular speed signal, a pitch angle signal, and a yaw angular speed signal input from the behavior amount detection means. and θj (j≠i
) and the current θi, (17) to (21)
θi is updated according to equations, equations (23), and recurrence equations obtained by discretizing equations (6) to (9), and output to intelligent means in other subsystems. Next, the recurrence formula is repeated to converge θi. Then,
Output θi as a command signal to the control means and return to the first step. (4) State quantity detection means State quantity detection means 401 to 404 in each subsystem
are stroke sensors 4011 to 4011, respectively, which detect the stroke of the suspension and convert it into a corresponding electric signal.
4014, vertical acceleration sensors 4021-4024 that detect vertical acceleration under the spring and convert it into a corresponding electric signal, and force sensors 4031-4034 that detect suspension force and convert it into a corresponding electric signal.
01 to 4004 and A to digitize these signals.
/D converters 4201 to 4204. (5) Control means The control means 501 to 504 in each subsystem operate based on the suspension stroke signal output from the state quantity detection means 401 to 404, the vertical acceleration signal under the spring, and the suspension force signal. , calculates a pressure valve opening signal necessary for the suspension force to match the command signal calculated by the intelligent means, and outputs this as a control variable. (6) Driving means The driving means 601 to 604 in each subsystem receives a pressure valve opening signal as a control amount signal outputted from the control means 501 to 504, and operates based on this signal. Operates the active suspension pressure valve to control oil flow. (7) Functions and Effects The functions and effects of this embodiment having the above configuration are as follows. First, a steering angle signal and an accelerator/brake operation amount signal representing the driver's intention output from the operation amount detection means 10, and a longitudinal speed signal and a lateral speed representing the current vehicle body motion output from the behavior amount detection means 20. A signal, a vertical speed signal, a vertical position signal, a roll angular velocity signal, a roll angle signal, a pitch angular velocity signal, a pitch angle signal, a yaw angular velocity signal,
Each of the intelligent means 301 to 304 and the control means 501 to 504 uses the suspension stroke signal, unsprung vertical acceleration signal, and suspension force signal, which are state quantities for each suspension, output from the state quantity detection means in each subsystem. It is input to CPUs 701 to 704 in the subsystem. Next, intelligent means 301 ~
At step 304, command signals are sent to other subsystems based on the operation amount and behavior amount to coordinate with other suspensions and to control the attitude of the vehicle body according to desired dynamic characteristics. Compute through a network with intelligent means within. [0046] Next, control means 501 to 50
4, the intelligent means 301 to 30 suppress the road surface disturbance detected from the unsprung vertical acceleration signal.
The pressure valve opening necessary to obtain the suspension target control force θ as a command signal calculated by 4 is calculated based on the suspension characteristics, and this is output as a control amount. Next, the drive means 601 to 604 input the pressure valve opening signal as a control amount outputted from the control means 501 to 504, and calculate the actual suspension force to the intelligent means 301 to 30.
The target suspension control force is corrected to the command signal calculated by 4. [0048] In this way, by performing autonomous decentralized control on each suspension, it is possible to simultaneously improve ride comfort by suppressing road surface disturbances for each suspension independently, and to improve steering stability through attitude control as each suspension cooperates. Ru. In addition, the command signal θ calculated through the network between the intelligent means 301 to 304 in each subsystem is the one necessary to obtain the target dynamic characteristic of the vehicle motion expressed by equation (22). Since this is the minimum norm solution among the control inputs, it is a very efficient control target. Furthermore, since command signals as control targets for each subsystem are computed autonomously and decentrally in the intelligent means 301 to 304 within each subsystem, excellent fault tolerance is achieved. That is, it is assumed that among the four wheels of the suspension, for example, the second wheel causes a failure and changes to a passive state. At this time, the intelligent means 302 in the faulty subsystem S2 is activated by the other subsystems S1, S3, S4
A force sensor 403 that is generated in a passive state, rather than a control target calculated by a recurrence formula in a network with
The suspension force as the output of 2 is expressed by the command signal θ2
Output as . In addition, in the intelligent means 301, 303, 304 in other subsystems, θ
The initial value of j (j≠2) is set to 0, and the command signal θj (j≠2) is calculated according to the recurrence formula. Through such calculation, other suspension controls share the role of the failed suspension control based on the motion state of the suspension that is in a passive state due to the failure. <Second Embodiment> This embodiment is a vehicle in which the steering angles of each of the four wheels can be controlled independently, and each of the four wheel steering mechanisms is made into a subsystem, and the configuration is shown in FIG. show. In a typical four-wheel steered vehicle, the left and right front wheels and rear wheels are each steered at the same steering angle. However, if you are willing to make the steering mechanism more complicated, it is possible to make the most of the tire capacity of each wheel by independently controlling the steering angles of the four wheels. As a result, it can be expected that the vehicle's motion limit will be increased. Therefore, in this embodiment, autonomous decentralized control is performed for a vehicle whose four wheels can be steered independently. (1) Operation amount detection means The operation amount detection means 11 is a sensor 11 consisting of a steering angle sensor 101 that detects a steering angle and converts it into a corresponding electric signal.
0 and an A/D converter 121 that digitizes these signals. (2) Behavior amount detection means The behavior amount detection means 21 includes a longitudinal speed sensor 201 that detects the longitudinal speed of the vehicle body and converts it into a corresponding electric signal, and a longitudinal speed sensor 201 that detects the lateral speed of the vehicle body and converts it into a corresponding electric signal. A sensor group 210 consists of a lateral velocity sensor 202 that detects the yaw angular velocity of the vehicle body and converts it into a corresponding electrical signal, and an A/D converter 221 that digitizes these signals. ing. (3) Intelligent means The intelligent means 311 to 314 in each subsystem are as follows:
The control means 511 to 514 as well as the CPU 7
11 to 714. CPU711 ~
Intelligent means 311 to 314 in 714
inputs the steering angle signal output from the A/D converter 121 of the operation amount detection means 11. In addition, the intelligent means 311 to 314 input the longitudinal velocity signal, lateral velocity signal, and yaw angular velocity signal output from the A/D converter 221 of the behavior amount detection means 21. Based on the above input values, the intelligent means calculates the target value of the cornering force of each wheel through a network with the intelligent means in other subsystems, and controls the respective control means 511 to 511.
514 as a command signal. Next, intelligent means 311 ~
An outline of the calculation in 314 will be explained below. First, it is assumed that a vehicle whose four wheels can be steered independently is traveling at a constant speed, and that cornering force is generated by steering in a direction perpendicular to the vehicle speed. If each tire is regarded as each subsystem, and the cornering force of each wheel is the output of each subsystem, the vehicle body motion is expressed as follows. dφ/dt=f(φ) +G(φ) ・θ
…
(26) However, G(φ) = [g1(φ), g2(φ), g3(φ
), g4(φ) ]
...(27) θ=[θ1, θ2, θ3, θ4]T

…(28) φ=[φ1, φ
2]T
…
(29) Here, θi : Cornering force generated at the i-th tire φ1 : Vehicle body slip angle φ2 : Yaw angular velocity

…(30) g1(φ) = [1/mv,
(af cosφ1 +df sinφ1)/I
z]T...(31) g2(φ) = [1/mv,
(af cosφ1 −df sinφ1 )/
Iz ]T…(32) g3(φ) = [1/mv
, (-ar cosφ1 +dr sinφ1)
/Iz]T…(33) g4(φ) = [1/m
v, (-ar cosφ1 -dr sinφ1
)/Iz]T...(34) However, af: Distance between front axle centers of gravity ar: Distance between rear axle centers of gravity df: Front suspension spacing/2 dr: Rear suspension spacing/2 Iz: Yaw moment of inertia m: Vehicle mass v :Vehicle body speed Now, let us consider controlling the behavior of the vehicle body to have the dynamic characteristics expressed by the following equation. dφ/dt=f0(φ)

...(35) However, f0(φ) = [-kb tanφ1 ,-kR
(φ2 −R0)]T…
(36) Also, R0 = kvδsw (kvδ
sw≦μmax g/v)...
(37) R0 = μmax g/v (
kvδsw>μmax g/v)
...(38) However, g: Gravitational acceleration R0: Target yaw angular velocity δsw: Steering angle μmax: Maximum road friction coefficient However, this is a setting that allows a yaw angular velocity R0 proportional to the steering angle δsw to be obtained in a region where the tire characteristics have some margin, and in a limit region to prevent spin and exhibit drift-out characteristics. be. In order to obtain the characteristics of the above equations, equations (30) to (34) and (36) to (38) must be replaced by (6
) or (9) and converged θ1 ~ θ
It is necessary to realize 4 as the cornering force of each wheel. Therefore, intelligent means 311 to 31
4, θ1 to θ4 thus obtained are output as command signals. In this embodiment, since the intelligent means is implemented by a digital computer, the calculations of equations (6) to (9) are performed using discretized recurrence equations. After all, the intelligent means 311 to 31 in the i-th (i=1, 2, 3, 4) subsystem
The process in step 4 follows the flow shown below. first,
Based on the longitudinal speed signal U and the lateral speed signal V input from the behavior amount detection means 21, the vehicle speed v and the vehicle body slip angle φ1 are estimated according to the following equation. v= (U2 +V2)1/2

...(39) φ1 = tan(V
/U)
…(4
0) Next, 0 is assigned as the initial value of θi and output to intelligent means in other subsystems. Next, the estimated vehicle speed v, the vehicle slip angle φ1, the steering angle signal input from the operation amount detection means, the yaw angular velocity signal input from the behavior amount detection means, and the intelligent means in other subsystems are calculated. θj input from
(j≠i) and the current θi, formulas (30) to (34) and (36) to (38), and (6)
According to the recurrence formula that discretizes equation (9), θi
and output to intelligent means in other subsystems. Next, repeat the recurrence formula and θi
Converge. Next, the control means 511 to 5
14, output θi as a command signal and return to the first step. (4) State quantity detection means State quantity detection means 411 to 414 in each subsystem
Sensors 4101 to 4104 each consist of force sensors 4031 to 4034 that detect the ground force of each wheel and convert it into a corresponding electric signal, and A that digitizes this signal.
/D converters 4211 to 4214. (5) Control means The control means 511 to 514 in each subsystem control the yaw angular velocity signal output from the behavior quantity detection means 21 and the ground force signal output from the respective state quantity detection means 411 to 414. and intelligent means 311-314
The steering angle necessary to obtain the target cornering force is calculated based on the target cornering force as a command signal calculated by , the estimated vehicle speed, and the vehicle body slip angle, and is output as a control variable. By the way, μi (corresponding to the lateral road surface friction coefficient), which is the cornering force θi standardized by the ground force Ni of each wheel, is a function of the slip angle αi of each wheel (the angle between the tire direction and the speed direction). μy (αi)
(See Figure 6). μi ≡θi /Ni = μy (αi)
(i=1,2,3,4)
...(41) Ni: Ground contact force αi of each wheel: Slip angle of each wheel [0067] Also, the slip angle αi of each wheel is expressed as a function of the vehicle body slip angle signal φ1 and the steering angle δi of each wheel as shown in the following equation. expressed. α1 = tan-1 (( af cosφ1
+df sinφ1)φ2/v) +φ1 −δ1
...(42) α2 = tan-1 (( af
cosφ1 −df sinφ1)φ2/v) +
φ1 −δ2 …(43) α3 = tan−1(
(-ar cosφ1 +dr sinφ1)φ2
/v) +φ1 −δ3 …(44) α4 = t
an-1((-ar cosφ1-dr sin
φ1) φ2/v) +φ1 −δ4 …(45) 00
[68] In the end, the processing in the control means 511 to 514 in the i-th (i=1, 2, 3, 4) subsystem becomes the flow shown below. First, the target cornering force θi as a command signal calculated in the intelligent means 311 to 314 is converted to the ground force Ni output from the state quantity detection means 411 to 414.
Standardize it by and set it as μi. Next, using the map shown in FIG.
Find the slip angle αi corresponding to . Next, this slip angle αi, the yaw angular velocity signal φ2 output from the behavior amount detection means 2, and the intelligent means 311
Based on the vehicle body speed signal v and the vehicle body slip angle signal φ1 estimated in steps 314 to 314, the steering angle δi is calculated according to the following equation. δ1 = tan-1 (( af cosφ1
+df sinφ1)φ2/v) +φ1 −α1
...(46) δ2 = tan-1 (( af
cosφ1 −df sinφ1)φ2/v) +
φ1 −α2 …(47) δ3 = tan−1(
(-ar cosφ1 +dr sinφ1)φ2
/v) +φ1 −α3 …(48) δ4 = t
an-1((-ar cosφ1-dr sin
φ1) φ2/v) +φ1 −α4 (49) Next, the calculated steering angle δi is output as the control amount for each wheel, and the process returns to the first step. (6) Driving Means The driving means 611 to 614 in each subsystem receive the steering angle signals output from the respective control means, and drive the steering actuators based on these signals. (7) Functions and Effects The functions and effects of this embodiment having the above configuration are as follows. First, a steering angle signal representing the driver's intention outputted from the operation amount detection means 11 and a steering angle signal representing the driver's intention outputted from the operation amount detection means 21
Intelligent means detects the longitudinal speed signal, lateral speed signal, and yaw angular velocity signal representing the current vehicle body motion output from 311 to 314 and the CPUs 711 to 714 in each subsystem constituting the control means 511 to 514. Next, intelligent means 311 ~
At step 314, a command signal that is a control target for each wheel for coordinating with other tires and controlling the behavior of the vehicle body according to desired dynamic characteristics is transmitted to other sub-wheels based on the operation amount and behavior amount. Compute through a network with intelligent means within the system. Next, the control means 511 to 51
4, the intelligent means 311 is based on the ground force of each wheel and the slip angle of each wheel as state quantities of each wheel.
The steering angle of each wheel necessary to obtain the cornering force θ of each wheel as a command signal calculated by 314 is calculated, and this is output as a control amount. Next, the drive means 611 to 614 input the steering angle signal of each wheel as a control amount outputted from the control means 511 to 514, and steer the wheels to the set steering angle, thereby achieving the target cornering. Gain the Force. [0075] In this way, by autonomous distributed control of the steering angle of each wheel, complex vehicle motion is distributed and hierarchized. This allows for global motion control, including areas near the limits. Further, the command signal θ calculated through the network of intelligent means 311 to 314 in each subsystem is a control input necessary to obtain the target dynamic characteristic of the vehicle body motion expressed by equation (35). Since this is the minimum norm solution, it is a very efficient control target compared to a normal vehicle in which the two front wheels and the two rear wheels are steered at the same steering angle. Furthermore, the command signals as control targets for each subsystem are computed autonomously and decentrally in the intelligent means 311 to 314 within each subsystem, thereby achieving excellent fault tolerance. That is, when the steering mechanism of one wheel malfunctions, other subsystems take over the role of the malfunctioning wheel, and it is possible to obtain almost the same vehicle body motion as during normal operation.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of a specific example of the present invention.

FIG. 2 is a diagram showing the relationship between each intelligent means.

FIG. 3 shows the relationship between other intelligent means when one subsystem fails;

FIG. 4 is a diagram showing the configuration of the first embodiment.

FIG. 5 is a diagram showing the configuration of a second embodiment.

FIG. 6 is a diagram showing the characteristics of the slip angle and cornering force of each wheel.

Claims (1)

[Claims] [Claim 1] Operation amount detection means for detecting the amount of operation by the driver; behavior amount detection means for detecting the amount of behavior of the vehicle body;
It consists of a plurality of subsystems provided corresponding to the wheel mechanism, and each of the subsystems adjusts the desired vehicle motion through a network with other subsystems based on the detected operation amount and behavior amount. intelligent means that calculates the dynamic characteristics that each of the subsystems should follow in order to obtain the dynamic characteristics and outputs it as a command signal; a state quantity detection means that detects the state quantity within the subsystem; and the detected state quantity. and a control means for calculating and outputting a control amount necessary for the subsystem to obtain the dynamic characteristics as the command signal based on the behavior amount and the command signal from the intelligent means, and a control means for calculating and outputting a control amount necessary for the subsystem to obtain the dynamic characteristics as the command signal, An autonomous decentralized control device for a vehicle, comprising a drive means for driving a wheel mechanism.