JPH01255746A - Travel controller for vehicle - Google Patents

Abstract

PURPOSE:To cover the response delay portion of a control system by operating a character value indicating an operation inclination proper to an operator on the basis of a plurality of time succession changes which is the aggregation of the extractions of the time succession changes of an acceleration pedal stepping in speed. CONSTITUTION:The time succession change of an acceleration pedal stepping in speed detected by means of a detection means a extracted by means of an extraction means b at a time point going upstream for a predetermined time from a time point which has become the position of a voluntary stepping in stoppage, and is aggregated by means of an aggregation means c. Next, a character value indicating an operation inclination proper to an operator is calculated by means of an operation means on the basis of a plurality of time succession changes aggregated. And a command means e, on the basis of the present acceleration pedal stepping in speed and the character value, predicts that an acceleration pedal position is at a position which is right before a position that comes to a predetermined stoppage position, and outputs beforehand a predetermined characteristic change command signal to a characteristic operation means f such as an automatic transmission, a suspension device or the like. Thus, the response delay portion of a control system is covered, and the responsiveness of control can be improved.

Description

[Detailed Description of the Invention] (Industrial Field of Application) The present invention relates to a vehicle travel control device, and more particularly, the present invention relates to a vehicle travel control device, and more specifically, it employs predictive control to control electronically controlled automatic transmissions, etc. This invention relates to a travel control device.

(Prior Art) In recent years, as the demands on vehicles have become more sophisticated, there is a tendency for vehicles to be equipped with various electronic control devices, which control the speed change characteristics of automatic transmissions, the suspension characteristics of suspension devices, etc. during driving conditions. It is finely variable according to the situation, raising ride comfort and driving performance to an even higher level.

FIG. 15 is a diagram showing an example of this type of electronically controlled automatic transmission. In this device, control is performed based on vehicle speed information from vehicle speed sensor 1 and accelerator pedal depression amount information from throttle sensor 2. Refers to a shift pattern stored in advance in the unit 3, and outputs a command signal to the automatic transmission 4 to switch to a predetermined gear when the shift up point or downshift point written in this shift pattern is exceeded. There is. In other words, it is possible to improve control accuracy by preparing multiple types of shift patterns such as economy patterns and power patterns, and by setting more detailed diagrams indicating shift amplifier points and shift down points. This makes it possible to obtain shift characteristics that match the driving conditions. And, there are also those that shift the transmission to an overdrive ratio depending on the driving condition.
For example, as is known from Japanese Patent Application Laid-open No. 52-89760, overdrive is prohibited while the amount of depression of the accelerator pedal is increasing.
When the amount of depression becomes constant, the gear is shifted to the overdrive ratio.

Although not shown, an example of an electronically controlled suspension system is such that when sudden acceleration is detected from information on the amount of depression of the accelerator pedal, a command signal is output that hardens the elastic characteristics of the suspension H to prevent nose-up. There is a system that immediately stops outputting the command signal and sets the elastic characteristics to be softer when the sudden acceleration is finished and the vehicle starts moving at a constant speed to improve ride comfort (see Japanese Patent Laid-Open No. 58-30544). ).

(Problem to be Solved by the Invention) However, in this type of conventional device, various information representing the running state of the vehicle is detected, and the current state of the vehicle body indicated by this information is determined based on the characteristics. When a change is required, a command signal is output at that point.For example, in many cases automatic transmissions and suspension devices are operated by hydraulic pressure, and a predetermined delay in the hydraulic response can be avoided. Therefore, even though the command signal was output at the optimal timing, there was a certain delay before actual control was performed. In other words, the conventional device is insufficient in terms of control responsiveness, and there is room for improvement.

(Object of the Invention) Therefore, the present invention aggregates accelerator pedal operation data, grasps the driving tendency of each driver from this aggregated data, and when an actual accelerator pedal operation is performed, the driving tendency By predicting the end of depression of the accelerator pedal and outputting a command signal in advance, the purpose is to compensate for the response delay of the control system and improve control responsiveness.

(Structure of the Invention) In order to achieve the above object, the vehicle running control device according to the present invention, as shown in FIG. an extraction means for extracting the time-series changes in the accelerator pedal depression speed traced back to a predetermined time from the time when the accelerator pedal reaches an arbitrary depression stop position; C, a calculation means d for calculating a propensity value representing a driver's unique operating tendency based on a plurality of time-series changes totaled by the totalization means C, and a current accelerator pedal detected by the detection means a. a command means e that predicts that the current accelerator pedal position is in the process of reaching a predetermined stop position based on the depression speed and the propensity value, and outputs a predetermined characteristic change command signal;
and characteristic operating means f for operating the shifting characteristics of the automatic transmission mounted on the vehicle or the suspension characteristics of the suspension device in accordance with the output of the command means e.

(Function) In the present invention, the time series of the accelerator pedal depression speed detected by the detection means is aggregated, and based on this aggregated time series data, the propensity (J
M calculation is performed. When the accelerator pedal is actually operated, the propensity value is referenced, it is predicted that the accelerator pedal is on the verge of reaching the stop position, and a characteristic change command signal is output in advance.

Since the above-mentioned command signal is input to the automatic transmission and suspension Wi before the required timing, the response delay due to the rise of oil pressure etc. is almost compensated for, and as a result, the command signal is input to the automatic transmission and suspension Wi at the desired timing. Since the speed change characteristics and suspension characteristics are manipulated, control responsiveness is improved.

(Example) Hereinafter, the present invention will be explained based on the drawings.

2 to 13 show an embodiment of the vehicle travel control device according to the present invention, which is an example applied to a vehicle equipped with an electronically controlled automatic transmission.

First, the configuration will be explained. In FIG. 2, 10 is a sensor group, and the sensor group 10 includes a throttle sensor (detection means a) 10a that detects the throttle opening θ that changes with the operation of the accelerator pedal, and a steering angle AGL of the steering wheel.
The vehicle is configured to include a steering angle sensor 10b that detects the steering angle sensor 10b, and a vehicle speed sensor 10c that detects the vehicle speed vR. Reference numeral 11 denotes a clear switch provided at the driver's seat. The clear switch 11 is operated by the driver as necessary, and a clear signal S_CLR is output by this operation.

Reference numeral 12 denotes a control unit having the functions of extraction means b, aggregation means C1, calculation means d, and instruction means e, and the control unit 12 includes a detector unit (hereinafter referred to as DET-U) composed of a microcomputer, etc., and a pattern control unit 12. Equipped with an analyzer unit (hereinafter referred to as PAN-U), each of the DET-U and PAN-U has an RA
M (Random Access Me+nory)
A storage device consisting of the following is provided. Below, DET-
The storage device inside U is called DET-M, and the storage device inside PAN-U is called PAN-M. DET-M stores IHM (Individual History Map), which will be described later, mapped three-dimensionally, and PAN-M stores IRM (Individual Real Map), which will be described two-dimensionally, which will be mapped two-dimensionally. has been done.

The PAN-TJ executes a predetermined program to sequentially capture θ from the throttle sensor 10a, arranges this θ on the time axis and stores it in the IRM, and also applies a predetermined value to the time series of θ stored in the IRM. A statistical method is applied to calculate a value that specifies the tendency of the driver's accelerator work, that is, a propensity value (described in detail later), and the obtained value is transferred to the DET-U. In addition, the PAN-U takes in the already transferred value from the DET-U as needed, and while referring to this value, estimates the accelerator work based on the current θ from the throttle sensor 10a, and estimates the accelerator work in the near future. It is predicted that the depressing operation will end and the transition will be made to the return operation, and the characteristic change command signal S is generated at the predicted time. Prediction processing such as outputting D is also performed. In addition, PAN
-U clears the contents of the IRM when SCLII is output from the clear switch 11.

On the other hand, DET-U executes a predetermined program and
The IHM stores the value that specifies the tendency of the accelerator work transferred from the AN-U, and also sends the stored value back to the PAN-U in response to a request from the PAN-U.

Furthermore, DET-U monitors the prediction process of PAN-U, compares the prediction result with the actual result, evaluates it, learns the evaluation result, reflects it in the prediction process of PAN-U, and A basic shift command signal S that causes the automatic transmission to perform the shift characteristics based on the basic shift pattern in accordance with the basic shift pattern.
Output IIAsE.

13 is a speed change control device (hereinafter referred to as an A/T control unit) for an automatic transmission as a characteristic operation means f;
Is the T control unit 13 internal? It has a basic shift pattern set in advance, and selects a gear position according to the driving condition by referring to this basic shift pattern according to θ and ■, and outputs a gear position signal S SEL. Note that such a determination operation is performed while the basic shift command signal SIIAsE is being input, but when SOD is being input, overdrive is forcibly selected.

14 is an electronically controlled automatic transmission (hereinafter referred to as EA/T), and EA/T 14 selectively engages/disengages a plurality of internal friction elements according to S SEL, for example, each part of a planetary gear. is engaged/disengaged to achieve the predetermined gear position indicated by SEL.

FIG. 3 is a block diagram showing the main parts of the PAN-U. In FIG. 3, 15 is a signal generation circuit that generates a clock signal CP, 16 is an A/D converter that converts θ from the throttle sensor 10a into a digital value, 17 is a latch circuit, and 18 is a large number of stacked circuits. A stack buffer 19 has a storage area and sequentially shifts the fetched θ from area to area for each cp; reference numeral 19 is an extraction signal S; A multiplexer 21 stores differential coefficient changes of multiple θs (specifically, first-order difference values θ′) stored in the calculation memory 20, and sends this θ′ to the IRM in the PAN-M. The storing calculation unit 22 is a processing unit that executes a part of a predetermined prediction processing program, and when the result of the predetermined prediction processing is correct, the processing unit 22
An extraction signal S is output that instructs to extract θ from the time when this prediction is made until the end of depression of the accelerator pedal from the stack buffer 18.

Next, the action will be explained.

- In a control system using a hydraulic actuator or a pneumatic actuator for S, a response delay of the system due to the rising time of the oil pressure or air pressure is unavoidable.

Therefore, even if control accuracy is improved by ignoring such response delay, the desired control effect cannot be sufficiently expected. In particular, modern vehicles tend to be equipped with a variety of control devices as user demands become more sophisticated, but many of these devices use hydraulic actuators, pneumatic actuators, and other devices with delayed response. There is. Therefore, there is an urgent need to compensate for the response delay, achieve the desired control effect, and respond to user requests.

Therefore, in this embodiment, after (I) depressing the throttle pedal to accelerate to a desired vehicle speed, depressing the throttle pedal is stopped in order to shift to constant speed driving at the desired vehicle speed; ([1) constant speed driving After moving to , all you need to do is transmit as little driving force to the wheels as possible to overcome the tread resistance.
Furthermore, from the standpoint of fuel efficiency, control is required to quickly shift to a higher gear (e.g., overdrive) when driving at a constant speed.When the desired vehicle speed is reached, the control system responds The aim is to quickly switch to a higher gear without being affected by delays.

In order to achieve this aim, the inventor focused on the following points. In other words, the amount of change (change N in θ) from the time when the accelerator pedal is depressed until it is stopped (hereinafter referred to as accelerator work) is shown in Figure 4.
As can be understood from this figure, the amount of change in θ gradually decreases with time, and finally reaches the amount of change O. This point of change amount 00 is the point at which further pedaling is stopped, and before reaching this point, the amount of change draws a predetermined curve and gradually converges. When the vehicle speed reaches the vehicle speed intended by the driver, the driver stops the accelerator work and maintains the accelerator pedal depression position, or makes fine adjustments in the returning direction as the case may be.

Incidentally, FIG. 5 shows an example of impossible accelerator work. In other words, the amount of change in θ increases over time, and it is possible that the line indicating the amount of change in θ will suddenly break and reach O at the same time as the desired vehicle speed is reached, as shown in Figure 4. It draws a curve and converges to the amount of change O. In this way, the process of reaching zero change draws a predetermined curve and converges, but the shape of this curve is determined depending on the driver's muscle strength (of the right leg) and ability to control muscle strength, so driving We wondered if it would be possible to identify the trends of the curve for each person. As a result of repeated experiments with many drivers based on this idea, it was confirmed that the curves of each Inusamurai have their own characteristics, as expected. , FIG. 6 is a diagram showing a driver's curve with relatively good reproducibility. The figure shows the results of four experiments for simplicity. Four curves (-
(represented by the floor difference value θ'), the opening between the curves is large at first, and as the point A is approached, the opening becomes smaller. In the case of a driver who has good reproducibility of this width, all curves fall within a predetermined width W at a predetermined time tA before point A, and almost all other curves (not shown) also fall within W. Therefore, when a new curve appears, if this curve falls within the above W, it can be concluded with an extremely high probability that the curve will extend to point A after the tA waiting time. That is, it is possible to predict that point A will be reached at time tA. Applying this to (1) and (II) above, if the accelerator pedal is depressed further and the amount of depression falls within W, the point at which the accelerator pedal is stopped is time tA. - It can be predicted in advance, for example the operation of switching to overdrive can be started 11 hours in advance.

Therefore, even if there is a response delay in the hydraulic actuator, this delay is offset by the LA, so that the response of the control system can be improved. In addition, although this is the case of the driver with poor reproducibility shown in Fig. 7, even in this case, point A
Since almost all curves fall within W before time L[l (tB<tA), control can be started before point A, although this is disadvantageous compared to the above-mentioned driver with good reproducibility. There is no difference, but the response delay can be reduced.

The operation of this embodiment based on such a principle will be described in detail below.

FIG. 8 is a flowchart showing part of the program executed by the control unit 12. In FIG. 8, this program consists of two processing sections. That is, they are the prediction determination processing section 12a from P1 to P7 and the learning processing section 12b from P8 to P13.

i1! L! !几 B 12 a First, when the ignition key is turned on and the power is turned on to the control unit 12, this program is started by a predetermined timer interrupt, and after resetting the error flag F with Pl, the predictive judgment process starts. be done.

That is, at P2, the throttle opening θ is read, and at P3, the amount of change Δθ in the throttle opening θ is determined according to the following equation (2). , is calculated.

Δθ. , −〇. -1, -θ. ,...■However, θ
. , : current θ θ( - sad) : previous θ Next, in P4, it is determined whether the magnitude of Δθ(1, exceeds a predetermined reference value ΔθTH. This is to provide a dead zone so that the control system does not respond to minute operations, and an appropriate size is selected so that the control accuracy does not deteriorate through experiments etc. In P5, the dead zone is set so that the control system does not respond to minute operations. The magnitude Δθ(
1) is the control comparison width (AVJ0±θj
o) Determine whether it falls within the range.

Here, AV. ±θj0 is θ above and below AVJo. These AVs have a wide range of width. and θ1° are IRM
This is the result of applying a statistical method to the data stored in . That is, the structural concept of IRM is shown as shown in FIG. The IRM is a two-dimensional arrangement of a large number of storage areas, and addresses are specified in the row direction by the time axis (j) and in the column direction by the number of sampling times (i). The time axis j is set at a predetermined time to (corresponding to the point A mentioned above) and moves to the right in the figure at predetermined time intervals from 1- to ~.
1-.

These 1-, ~ are arranged by going back in time to -1
are arranged as a signal sequence on the same time axis for each sampling number (i). Storing data in the IRM is performed by the learning processing unit 12b, which will be described later, but will be explained here because it is necessary for understanding the next statistical method.

Fig. 10 is a diagram showing the change in θ due to accelerator work.
It has reached this point. 1- in time that goes back in the past than to
. . . , θ becomes smaller in the order of 1-3, . . . .

FIG. 11 is a diagram in which the first-order difference value θ' of θ is plotted on the inter-axes going back in the past, from tl to t.

The value of θ' gradually decreases as it moves toward (from the past to the future) and reaches a minimum value at to. That is, an example of the IRM shown in FIG. 9 stores such changes in θ' for each sample. Note that when storing to rRM, the contents are first stored in row 81, and in the next sample, the contents of row 31 are moved to row 32, and the empty S
New data is stored in row I. Then, the data of the next sample input after all the rows have been filled are similarly stored in the S1 column, and S1 is thereby pushed out like a ball. The data in the column will be discarded. The tendency of θ' stored in the IRM in this manner is analyzed using the following statistical method. Note that hereinafter, data in each area on the IRM will be referred to as d ij. However, i represents the number of sampling times, and j represents the time to go back. That is, all data on the IRM may be conveniently considered as a matrix.

When data is stored in all areas on IRM,
The unbiased variance value vA of the entire data is calculated according to the following equation (2).

However, m: Maximum traceback time n: Maximum number of samples Note that the above calculation (2) is executed every time new sampling is performed for the IRM. On the other hand, at the same time as the above-mentioned calculation (2), the unbiased variance value (1) for the data at each inter-column time is calculated according to the following equation (2).

(Σ dkJ)2 Next, the unbiased dispersion ratio F of VA and V is calculated according to the following equation (2).

Fj = VA/Vj...■ However, in the case of F, <1, Fj = F, /VA This indicates that the larger the value of F4, the more peculiar the tendency of the data is. Then, F obtained by the above formula ■,
Perform F test on. The F value of the F test is calculated only once according to the following formula 〇.

F=F (m-n-1, n-1, 0,05)...
... ■As a result of the F test, F and the F value can have the following two relationships.

Male ``J■Anal FJ<F th''''■苅伶F, >F In the first relationship, F is less than or equal to the F value, which indicates that the trend of the data is not unique in this case. In other words, this is a case where the accelerator work cannot be specified due to variations in data. On the other hand, in the second relation, FJ is F
In this case, a peculiar trend has appeared in the data.

In other words, JX of the value F4 is F (J is 1-, ~
1) to Jo (L.) (θ')
represents the unique habits of the driver from whom the data was extracted, so the data in this section is reliable as a model for prediction processing. Therefore, for all the data included in this section, for example, the section from 12 to t0 in FIG. 9, the average value AV is calculated according to the following formula. seek.

Σ dkJo AVJo−□・・・・・・■ The above formula 0 is executed every time the data is updated,
It will be updated to a more reliable AV;o.

On the other hand, AV. At the same time as calculating F according to the following formula 〇.

and standard deviation of F value θ,. Calculate.

That is, θ,0 is J such that FJ>F and FJ, ,<F
This is the standard deviation value of point o.

Again, in FIG. 8, Δθ (Ll is AV, at P5.

±θ,. When it falls within the range of , it corresponds to a peculiar tendency of accelerator work.
6, an error flag F (described later) is checked, and if F=1, the process proceeds to P7, where a 1-characteristic change command signal S is issued to instruct switching the gear stage of the E-A/T 14 to a higher gear stage. 0
Output. In other words, at the time of YES in P mentioned above, the accelerator is still being pressed down, but for example,
In the case of the above-mentioned driver with good reproducibility, after tA, and
Even in the case of a driver with poor reproducibility, it is predicted that he will stop pressing the accelerator pedal after time t8, so S in advance. By outputting D, control (shift control I) can be started taking into account the response delay of the hydraulic mechanism of the E-A/T 14. Therefore, at the time when the accelerator pedal is actually stopped from being pressed further, the shift control has ended or is about to end, and the responsiveness of the control is improved.

Science) υth 1st block →- By the way, since the above-mentioned prediction determination process is performed based on a statistical method, there are very few cases in which the prediction is wrong. Therefore, by performing the learning process described below, the IRM
We are trying to increase the reliability of the data stored in the . First, the learning processing section 12b has P of the prediction determination processing section 12a.
6 YES command, that is, when error flag F=1, and S at P7. When , is output, the processing moves through two routes. The error flag is set to P (described later) when it is determined that an incorrect determination has been made when the prediction determination made by the prediction determination processing unit 12a is compared with the actual result.
In this case, P is bypassed until the error flag is reset, and the A/T control unit 13 performs normal speed change control based on the basic speed change pattern. On the other hand, in the case of the route from P7, since the error flag is not set, this indicates that at least the previous prediction judgment evaluation was passed. However, even in this case, evaluation is performed to increase the reliability of the data. The evaluation of the predictive judgment is performed in two stages.

First, the first stage evaluation performed at P6 will be explained with reference to FIG. For example, assuming that the prediction determination (YES command) at P5 described above was made at time 1-2 in FIG. 10, if the prediction is correct, the amount of change in θ should be constant at to.

However, in some cases, the accelerator work may be performed such as turning down the alley again after the pedal pressure is stopped momentarily at TO, and in such cases, the above prediction is incorrect. That is, since the driver's intention is still to accelerate, if the driver switches to the front gear, the driving force will be insufficient and the desired acceleration feeling will not be obtained, resulting in a worsening of the driving feeling. Therefore, at the point when you stop pressing the accelerator pedal (point A)
), a predetermined time TF is defined as the monitoring period, and this T,
If the amount of change in θ between
It is determined that it has stabilized. In other words, if it is settled at point F, the first stage evaluation is passed, and t0~ shown in Figure 10.
Since θ of 2... is reliable as sampling data, it is stored in the IRM as P. On the other hand, if it does not settle at point F, the actual accelerator work that is the target of prediction is temporary, and in this case it is not suitable as sampling data, so it is discarded by PIG, and the prediction judgment processing unit 12a described above Shift processing to .

Next, the second stage evaluation will be explained. First, the aforementioned AV.

-, the estimated accelerator opening θ(1) is calculated from the average of the predicted values expressed as a sequence of numbers, and this θ. , and the actual accelerator opening θ. Based on , the following equation is executed to find the error area E between the curves drawn by both.

E-1(θ., -θ., 1dt...■The concept of E calculation is shown in Fig. 12. In Fig. 12, the changes in the actual throttle opening θ, t, time t=Q
is within a predetermined reference width ε (e.g. 2deg), and furthermore, the time TG (e.g. 2sec) from t=Q
If it is still within ε even after t=Q, it is said to have stabilized at point G, and from t=Q to time a (for example, a=0
．． 5 to l sec), θ. , and θ. Find the area between the curves drawn by , and let this area be E. In other words, the smaller E is, the more θ. ) = θ. , approach
This indicates that the prediction judgment is highly accurate. In addition,
If θ(tl exceeds the range of ε before point G, the evaluation described below is not performed and the counter C5 (not shown) is set to +
Increment by 1. Further, when performing the evaluation described later, both the counter C1 and the counter C2, which is also not shown, are incremented by +1. These 01 and C2
is reset every time the prediction determination processing unit 12a performs a predetermined number N of prediction determinations. The 02 count value is an index indicating whether or not the value of E has been sampled to a sufficient extent for evaluation.If the following formula (■) is satisfied, it is determined that evaluation is impossible due to insufficient samples, and the A/T control unit 13 Basic shift command signal S B
ASE is output and normal gear shift operation based on the basic gear shift pattern is performed.

C2<N, <N...■ However, N: C,, the prediction judgment number No. of the prediction judgment processing section 12a that resets C2: If the formula 0 is not satisfied on the predetermined numerical value, the A final evaluation of the prediction judgment is performed in P1□. First, the average value E′ of E every predetermined number of times
is calculated, and this E' is compared with a predetermined numerical value e. If the comparison result is E'>e, the error is large, so the prediction judgment is incorrect and an error flag F is sent in PI3. On the other hand, if E'>e is not, the prediction judgment is correct and an error is sent in PI4. The flag is reset, and in either case, the process is transferred to the prediction determination processing section 12a. In this manner, in the learning processing unit 12b, if the result of the first evaluation is correct, the data at that time is stored in the IRM, and thereafter, the final pass/fail of the prediction judgment is determined by the second evaluation. Therefore, even if a change trend appears in the accelerator work of a particular driver, that data is stored and updated in the TRM, and future prediction judgments are made based on the new data. Change trends in the workpiece will be learned. As a result, highly flexible predictive judgments are made that take into account factors such as driver fatigue.

In the Konoyo Unigi Example, time-series changes in the throttle opening θ are aggregated and stored in the IRM, and a statistical method is applied to the data in the IHM to obtain a value that indicates the pattern tendency of the driver's specific accelerator work. Find (AV, .±θ, .),
This value is compared with the actual θ, and if the pattern corresponds to a unique accelerator work pattern, an operation to change the gear stage of the EA/T 14 is started in advance. Therefore, when the accelerator pedal is actually stopped from being pressed further, the gear has already been changed or is about to be changed, which eliminates the response delay of the control system and improves the driving feeling of the vehicle. be able to.

In addition, AV has a learning function. ±θ,. The system is constantly updated to improve reliability, so it can flexibly respond not only to changes in drivers, but also to changes in the accelerator work pattern of the same driver due to fatigue, etc. can do.

Furthermore, AV. ±θ,. The value of is stored in the IHM having a three-dimensional conceptual structure shown in FIG. In FIG. 13, the IHM is addressed in three directions. That is, the direction addressed by the average value V3F of the vehicle speed 3 per unit time, the direction addressed by the integral value 5AGt of the steering angle AGL, and Δθ35. is the direction addressed by the average value Δθ'. Storing to IHM is performed as follows. For example, a new AV for a given pro.

When storing ±04°, the previously stored AV.

±θ,. and the averaged value is stored. Then, this stored value is a combination of 5AGL, VSF, Δθ', and one block is addressed, AV in that block. ±θ,. is read out, sent to the PAN-U, and used for processing by the prediction determination processing section 12a. At this time of reading, the previously read block number is stored in a predetermined memory, and when a new block is addressed, this new block number is changed from the previously stored block number. The virtual distance between both blocks is calculated. If D is larger than the predetermined value d, it indicates that the data such as SAG vsp and Δθ' have changed significantly. In this case, for example, the driving environment has changed from -i road to expressway. It has changed to a scratched edge. Therefore, in this case, the reliability will decrease if predictive judgment is made based on the previous data, so the basic shift command signal S BASE is not output to the A/T control unit 3 until reliable data is accumulated. to perform normal shift control based on the basic shift pattern.

In addition, in the 3rd to/T control unit, when the basic shift command signal S IIAsE is input, the above-mentioned operation is performed, and when the sampling start signal (not shown) from the PAN-U is input, the normal shift operation is performed. and waits for the characteristic change command signal SOO. Note that even after the predetermined time tX has elapsed, the characteristic change command signal S remains unchanged. If 0 is not input, the inhibition is canceled and normal shift control is performed.

If Soo is input within tX, tpr O, 1 second after S(l is input (however, tp, < Q,
5 seconds), a gear shift signal S SEL is output to shift to overdrive. Note that if the return of the throttle pedal is large even within tX, SSEL is output and the shift is made to overdrive.

In addition, the IHM shown in FIG. 13 and the IHM shown in FIG.
Since the data within is unique to each driver, it is preferable that the data be stored in a portable storage device such as a memory card and retained by each driver.

Furthermore, switches are installed on the ignition key, seat belt mechanism, seat slide, bank mirror, etc., and the signal from these switches is used to detect when the driver has changed.
It is also possible to initialize the IRM or IRM to speed up the startup of the control.

Figure 14 shows an example in which the present invention is applied to an electronically controlled suspension device. In FIG. 14, 30 is, for example, the vehicle height H.
31 is a throttle sensor (detection means a) that detects throttle opening θ, 32 is a control unit similar to that of the first embodiment, and 33 is a suspension control device. The control unit 32 is the same as in the first embodiment.
It is composed of U and PAN-U, and grasps the pattern tendency of accelerator work with the same effect as the first embodiment,
Predict when to stop pressing the accelerator pedal and
Output. Therefore, the control section 32 has the functions of extraction means b, aggregation means C1 calculation means d, and instruction means e. The suspension control device (characteristic operation means f) 33 receives a control signal SR, which changes the overall length of the suspension device 34R134L or increases or decreases the damping force based on information such as vehicle height H and vehicle speed Vn (not shown).
Output SL.

In such a configuration, when controlling to change the overall length of the suspension or change the damping force of the shock absorber 34R134L using hydraulic or pneumatic pressure, the required control timing, for example, when you stop pressing the accelerator pedal, is determined. As a timing, the control signals SA and S11 can be given in advance prior to this timing. Therefore, in this case as well, the response delay of the control system can be eliminated, and the riding comfort can be improved.

(Effect) According to the present invention, accelerator pedal operation data is aggregated, the driving tendency of each driver is grasped from this aggregated data, and when an actual accelerator pedal operation is performed, the driving tendency is determined based on the driving tendency. It is possible to predict when the accelerator pedal will end and output a command signal in advance.

Therefore, the response delay of the control system can be compensated for, and the responsiveness of the control can be improved.

[Brief explanation of the drawing]

Fig. 1 is a basic conceptual diagram of the present invention, Figs. 2 to 13 are diagrams showing a first embodiment of the invention, Fig. 2 is an overall configuration diagram thereof,
Fig. 3 is a block diagram of the main part, Fig. 4 is a graph showing the change in θ due to the accelerator work, Fig. 5 is a graph showing the impossible change in θ due to the accelerator work, and Fig. 6 is a graph showing the change in θ due to the accelerator work. A graph showing a curve created by a driver with good reproducibility, FIG. 7 a graph showing a curve created by a driver with poor reproducibility, FIG. 8 a flowchart showing the process flow, and FIG. 9 showing the IRM. The conceptual diagram, Figure 10 is a graph showing the data stored in the IRM, and Figure 11 is the graph showing the data stored in the IRM.
12 is a graph showing its integral range, FIG. 13 is a conceptual diagram showing its IHM, and FIG. 14 is a second embodiment of the present invention. FIG. 15 is an overall configuration diagram showing a conventional example. 10a...throttle sensor (detection means a),
12...Control unit (extraction means b, aggregation means C1 calculation means d, command means e), 13...A/T control unit (characteristic J
fi operation means f), 31...Throttle sensor (detection means a), 3
2... system: 111 copies (extraction means C, aggregation means C)
, calculation means d, command means e), 33...The suspension control'tM
l ('P? gender manipulation means r).

Claims (1)

[Scope of Claims] a) Detection means for detecting the operation speed of the accelerator pedal; and b) A detection means for detecting the operation speed of the accelerator pedal, and b) the accelerator pedal operation speed that is traced back to a predetermined time from the time when the accelerator pedal is depressed and reaches an arbitrary depression stop position. an extraction means for extracting series changes; c) aggregation means for aggregating the time series changes extracted by the extraction means; d) driver-specific information based on the multiple time series changes aggregated by the aggregation means. a calculation means for calculating a propensity value representing an operation tendency; and e) a calculation means for calculating a propensity value representing an operation tendency; command means for predicting that the vehicle is in the process of and outputting a predetermined characteristic change command signal; A vehicle running control device comprising: a characteristic operation means.