JPH0471933A - Travel control device for vehicle - Google Patents

Abstract

PURPOSE:To allow a comfortable constant speed travel coinciding with actual drive characteristics by reflecting the past travel data of a driver's own vehicle upon speed control, using a distance from a forward vehicle and time series values of speed deviation and acceleration of the driver's own vehicle within the predetermined time as control parameters. CONSTITUTION:A distance between a forward vehicle and a driver's own vehicle is detected with a distance sensor 10 such as a laser radar device at every predetermined time, and the output thereof is inputted to a pre-processing circuit 20, together with output from a steering angle sensor 16 and a torque sensor 18. In addition, each parameter operated or processed for saving in the aforesaid pre-processing circuit 20 is inputted to a neural network 22 comprising a hierarchical structure of input, intermediate and output layers, and converting and outputting a signal through the predetermined rule according to a total of input signals weighted with the predetermined value (coupling factor). An acceleration or deceleration control magnitude is thereby controlled, according to time series values within the predetermined time for controlling a throttle actuator 24, a brake actuator 26 and a gear change controller 28.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a vehicle travel control device, and particularly to a vehicle travel control device that causes a vehicle to travel at a constant speed according to the past travel history of the vehicle.

[Prior Art] Conventionally, control devices have been developed that cause a vehicle to travel at a constant speed at a set speed in order to improve driving operability, particularly on expressways.

As this type of device, for example, Japanese Patent Application Laid-Open No. 60-21543
A vehicle travel control device disclosed in Publication No. 2 is known. In this vehicle running control device, the purpose of this device is to maintain the own vehicle speed at the set vehicle speed or the following distance at a safe following distance without the need for the vehicle driver to operate the accelerator. A preceding vehicle detection means is provided to detect the presence or absence of a preceding vehicle within the vehicle, and when there is a preceding vehicle, the vehicle speed is controlled so that the following distance becomes a safe distance, and when there is no preceding vehicle, the vehicle speed is controlled to be either the own vehicle speed or the set vehicle speed. It runs under control.

[Problems to be solved by the invention] However, in controlling such conventional devices,
This control is based only on the current vehicle speed and following distance to maintain the own vehicle speed at a set speed or the following distance to a safe distance, and does not necessarily match the actual driving characteristics of the driver and may not be comfortable. There was a problem in that running could not be guaranteed.

That is, for example, consider a situation in which a preceding vehicle changes lanes from its own lane to an adjacent lane, and the situation suddenly changes from a situation where there is a preceding vehicle to a situation where there is no preceding vehicle. In this case, the control system switches from vehicle speed control based on the safe inter-vehicle distance to vehicle speed control based on the set vehicle speed, but for a normal vehicle driver, although there is currently no vehicle in front, there was a vehicle in front until a few seconds ago, and the current The driver recognizes that his own vehicle speed was set in response to the situation a few seconds ago, so he gradually operates the throttle to gradually accelerate the vehicle and set it to a comfortable vehicle speed. Conventional devices do not recognize the presence of a preceding vehicle until several seconds ago, so they must instantly match the vehicle speed to the set vehicle speed (and perform rapid acceleration control).

The present invention has been made in view of the above-mentioned conventional problems, and its purpose is to perform comfortable constant-speed driving that matches the actual driving characteristics by controlling the vehicle speed in consideration of the past driving history of the vehicle. The object of the present invention is to provide a vehicle travel control device that can perform the following functions.

[Means for Solving the Problems] In order to achieve the above object, the vehicle travel control device according to the present invention includes an inter-vehicle distance detection means for detecting the inter-vehicle distance between the vehicle in front and the own vehicle, and a means for detecting the inter-vehicle distance between the vehicle in front and the own vehicle. A vehicle speed detection means for detecting, an acceleration detection means for detecting the acceleration of the own vehicle, a calculation means for calculating a speed deviation which is a difference between the detected vehicle speed and a predetermined target vehicle speed, a detected inter-vehicle distance, Time-series values of acceleration and calculated speed deviation within a predetermined time are manually input, and the time within a predetermined time is calculated using a neural network formed by connecting multiple layers of neural units that output according to the weighted sum of the human input. The present invention is characterized by having a calculation means for calculating an acceleration/deceleration control amount according to the series value, and a control means for controlling the speed of the own vehicle so as to reach a target vehicle speed according to the calculated acceleration/deceleration control amount.

[Function] The vehicle running control device of the present invention has such a configuration, and is capable of controlling the distance between the vehicle in front and the vehicle ahead, the speed deviation which is the difference between the own vehicle speed and the target vehicle speed, and the acceleration of the own vehicle within a predetermined time. By using the time series values of , as control parameters, past driving history is reflected in the control.

For example, if the vehicle in front changes lanes to the adjacent lane,
The change appears as a change in the time-series data of the distance between vehicles, and the past driving history is recognized.

By processing these time-series values using a neural network that does not require complicated preprocessing of input data, M1 speed control according to past driving history can be performed easily and quickly.

[Embodiments] Hereinafter, preferred embodiments of the vehicle travel control device according to the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of the configuration of this embodiment. A distance sensor 10 such as a laser radar device is attached, for example, near the bumper of the own vehicle, and detects the inter-vehicle distance between the preceding vehicle and the own vehicle at every predetermined time Δ. Then, the detected inter-vehicle distance ML is sequentially sent to the preprocessing circuit 20.

On the other hand, the vehicle speed sensor 12 and the acceleration sensor 14 detect the speed ■ and acceleration g of the own vehicle at predetermined time intervals Δ, respectively.
These data are sequentially sent to the preprocessing circuit 20.

Furthermore, the steering angle sensor 16 and the torque sensor 18 detect the steering angle θS and drive torque TD of the own vehicle, respectively, and similarly, the detected values are sequentially sent to the preprocessing circuit 20. Note that the driving torque may be calculated based on the engine rotation speed, engine load signal, or gear ratio.

The preprocessing circuit 20 has a storage unit and an arithmetic processing unit (not shown), and stores data within a predetermined time among the data of inter-vehicle distance L1, vehicle speed V, and acceleration g sent from each sensor every Δ. The information is sequentially stored in the storage unit.

This predetermined time is determined according to the following detection signals: inter-vehicle distance L1 vehicle vehicle speed 1 acceleration g For example, in the case of inter-vehicle pressure 1lItL, this predetermined time is determined as N Δt (NL is 0 or a natural number), and the current inter-vehicle distance is From pressure ML (0) to NL·Δ, the previous inter-vehicle distance L (NL) is stored.

Regarding vehicle speed V and acceleration g, each predetermined time is N.
v ・Δt (Nv is O or a natural number) and N ・Δt
(N is O or a natural number), 8g From the current vehicle speed v (0), Nv conflict Δ is the previous vehicle speed V (Nv)
and from the current acceleration g(0) to NgΔ the previous acceleration g(
N) will be stored.

Here, regarding the vehicle speed V, there is also a speed deviation ΔV "" V from the target vehicle speed V set by the vehicle driver.

V is sequentially calculated and stored in the arithmetic processing unit.

In this embodiment, the acceleration sensor 14 is used to detect the acceleration g, but instead of providing a separate sensor, the acceleration may be detected.

The acceleration g detected by differentiating the speed V from the acceleration sensor 14 or the speed sensor 12 of course indicates the acceleration of the vehicle, but it is different when the vehicle is running on a flat road, on an uphill road or on a downhill road. The driving torque required to achieve that acceleration g differs depending on when the vehicle is traveling. In other words, when controlling the throttle, brake, etc. using the detected acceleration g, even if the same driving torque is applied, the effect on vehicle running when the vehicle is running on an uphill road, a downhill road, or a flat road. Therefore, when controlling vehicle travel based on this acceleration g, it is necessary to judge whether this acceleration g is an acceleration on a flat road, or an acceleration on a climbing road or a descending road.

Therefore, in this embodiment, the estimated acceleration g.

A new physical quantity is introduced. This estimated acceleration g.

is an estimated acceleration assuming that the vehicle is traveling on a flat road, and is calculated using the drive torque TD and the vehicle speed V as shown below.

g -C-T -C-V -C31D2 c,,c2. c3: Constant Then, by using this estimated acceleration go, for example, when the vehicle is running on an uphill road, the actual acceleration is low even with the same driving torque, so g-go <0, and conversely, when the vehicle is running on a downhill road, g g o > 0, so acceleration g and estimated acceleration g. It is possible to determine whether the vehicle is currently running on the uphill road or downhill road based on the size relationship between the two.

In this way, inter-vehicle distance L1 acceleration g1 vehicle speed V and target vehicle speed ■. After the speed deviation ΔV, the difference between the acceleration g and the estimated acceleration g, and the steering angle θS1 drive torque TD are calculated or stored in the preprocessing circuit 20, these parameters are output to the neural network 22 as a calculation means. Ru.

Below, using FIG. 2, this neural network 22
We will explain the processing performed in .

As is well known, a neural network has a hierarchical structure of an input layer, an intermediate layer, and an output layer, and each layer is composed of a group of neural units. The neural unit is a unit that converts and outputs according to a certain rule according to the sum of inputs to which a predetermined weighting (coupling coefficient) WIj is added. Various functions are used as this rule, but in this example, when the total human power "net" is net-ΣWI jI +, f=1/il+exp [-(net+a)]l, where α is a constant We decided to use the 51g1oid function. The value range of this function is 0 to 1, and as the input value becomes larger, it approaches 1, and as the input value becomes smaller, it approaches O.

Now, each parameter from the preprocessing circuit 20, namely, inter-vehicle distance L(0) to L(NL), speed deviation Δv(0)Nv
(Nv), acceleration g(0) to g(N), target speed V. , vehicle speed ■, difference between acceleration g and estimated acceleration ggo
One steering angle θS is input to each neural unit constituting the human layer of the neural network shown in FIG. Further, the current throttle opening degree θTH and brake operation amount θB are also manually inputted to this human power level in order to be fed back.

After each parameter is input to each neural unit in the input layer in this way, a predetermined conversion process is performed in the intermediate layer and output layer, which are made up of a group of similar neural units. That is, if the output value of the i-th neural unit in the human layer (D) input to the j-th neural unit in the intermediate layer is 1, and the weighting or connection coefficient at this time is w, then this j-th The total sum net- of the input values input to the th neural unit is net, -ΣWlj°Ii, and the output is y-=f (net,) J J using the sigmoid function as described above. Then, the above-mentioned processing is performed in all the neural units existing in the intermediate layer, and the output value is inputted to each neural unit in the output layer.

The same conversion process as in the intermediate layer is also performed in the neural unit of the output layer. In other words, the first neural unit in the intermediate layer that is input to the jth neural unit located in the output layer
The output from the th neural unit is yl, and the weighting at this time is W8. ″ and 1
1J, the total sum of input values input to this j-th neural unit, net, becomes net - ΣW... @ y J IJ I, and the output value O1 at this time is O - f (net , -) J J . Note that the weighting wl from the input layer to the hidden layer and the weighting W, j from the hidden layer to the output layer are set such that the difference between the actual output value from the output layer and the desired output value is reduced. Adjust in advance by learning. In other words, each weighting may be determined by a skilled driver actually driving the vehicle.

In addition, when converting the output value from the sum of input values in each neural unit, instead of calculating using the conversion function f each time, the sum value manually entered in advance in ROM etc. and the output value at that time are The conversion process may be performed by storing the data as a master knob and reading it from this ROM without performing arithmetic processing.

By using a neural network in this way, a set of output values corresponding to the - set of human image information can be obtained, but each neural unit that makes up the output layer corresponds to a predetermined state. Based on the output value from this output layer, a control signal ΔθT11 indicates how much the throttle actuator should be driven, a control signal Δθ3 indicates how much the brake actuator should be driven, and a control signal indicates where the gear ratio of the transmission should be set. To〇.

is determined.

Then, each of these control signals Δ, 11, Δθ8, Δθ
are respectively output to the throttle actuator 24, brake actuator 26, and shift controller 28 as control means, and are the throttle opening θTl+, the brake operation amount θ8, and the amount necessary to set the vehicle speed to the target vehicle speed.
The gear ratio θ is output and controlled.

In this way, in this embodiment, the inter-vehicle distance L1, the speed deviation ΔV, and the acceleration g are treated as time series data, and this time series data is treated as a reflection of the vehicle's travel history and processed by the neural network. Rather than performing control according to the vehicle's driving situation at the moment, it also takes into account the vehicle's past driving history and controls the current driving of the vehicle, which more closely matches the actual driving interval of the driver. It becomes possible to perform control, and more comfortable follow-up driving can be performed.

[Effects of the Invention] As explained above, according to the vehicle running control device according to the present invention, since the running of the vehicle is controlled in consideration of the running history of the vehicle, it is possible to respond more flexibly to various driving situations. This has the effect of allowing comfortable follow-up driving.

[Brief explanation of the drawing]

FIG. 1 is a configuration block diagram of an embodiment of a vehicle travel control device according to the present invention, and FIG. 2 is an explanatory diagram of a neural network in the same embodiment. Distance sensor Vehicle speed sensor Acceleration sensor Steering angle sensor Torque sensor Preprocessing circuit Neural network Throttle actuator Brake actuator Shift controller

Claims (1)

[Scope of Claims] Inter-vehicle distance detection means for detecting the distance between the preceding vehicle and the own vehicle; vehicle speed detection means for detecting the speed of the own vehicle; acceleration detection means for detecting the acceleration of the own vehicle; a calculation means for calculating a speed deviation which is the difference between the calculated vehicle speed and a predetermined target vehicle speed; Calculating means for calculating an acceleration/deceleration control amount according to time series values within a predetermined time using a neural network formed by connecting multiple layers of neural units that output according to a weighted sum, and the calculated acceleration/deceleration. 1. A vehicle travel control device comprising: control means for controlling the speed of the own vehicle so that the speed of the host vehicle reaches a target vehicle speed according to a control amount.