Autonomous vehicle for automatically/autonomously running on route of travel and its method using fuzzy control

• US 5,122,957 A
• Filed: 02/27/1990
• Issued: 06/16/1992
• Est. Priority Date: 02/28/1989
• Status: Expired due to Term

• Current Assignee: Nissan Motor Co., Ltd.
• Original Assignee: Nissan Motor Co., Ltd.
• Inventors: Hattori, Akira
• Primary Examiner(s): Lall, Parshotam S.
• Assistant Examiner(s): Auchterlonie, Thomas S.
• Application Number:
  US07/485,910
• Time in Patent Office: 840 Days
• Field of Search: 364/424.02, 364/424.01, 364/516, 382/10, 382/22, 382/25, 382/1
• US Class Current: 701/27
• CPC Class Codes: B62D 15/025 Active steering aids, e.g. helping the driver by actively influencing the steering system after environment evaluation B62D1/28 takes precedence
• CPC Class Codes: G05D 1/0221 involving a learning process
• CPC Class Codes: G05D 1/0246 using a video camera in combination with image processing means
• CPC Class Codes: Y10S 706/90 Fuzzy logic

Abstract

An autonomously running vehicle and method for autonomously running the vehicle on a route of travel are disclosed in which using a plurality of control rules defining predetermined steering controlled variables according to magnitude classifications of previously classified deviations on steering operations and membership functions indicating certainties corresponding to respective classifications of actual deviations (offset), the certainties corresponding to each magnitude classification are correlated to the control rules to determine the steering controlled variable. In the autonomously running vehicle, the vehicle'"'"'s running scenes are automatically detected on the basis of the image data derived by an image processing unit. Reference membership functions are varied (constricted or reduced) according to any one of the running scenes so that the membership functions are automatically set according to the running scenes. Control rules are previously set only to the reference membership functions. Hence, setting operations of control rules and membership functions through simulations may be carried out only for the reference membership functions.

Claims

1. An autonomous vehicle, comprising:

a) first means for imaging a road surface on which the vehicle runs, having an image plane and an associated image data generating means;

b) second means for detecting a plurality of predetermined features contained within the data output by the first means and generating characterizing point data corresponding to said plurality of predetermined features;

c) third means for classifying the image data output by said first means as a running scene among a group of predetermined running scenes on the basis of the characterizing point data generated by the second means;

d) fourth means for selecting at least one among a group of predetermined reference membership functions according to the classification of the running scene, said membership function indicating a certainty of correspondence between a measured steering deviation quantity and a member of a group of predetermined deviation magnitudes, said measured steering deviation quantity being a difference between a measured vehicle steering angle and a predetermined vehicle steering angle;

e) fifth means for measuring a steering angle of the vehicle;

f) sixth means for selecting, based on a reference membership function selected by said fourth means and a measured steering deviation quantity operated on by said selected function, at least one among a group of predetermined control rules;

g) seventh means for calculating a steering controlled variable of the vehicle on the basis of the measured steering angle of the vehicle, characterizing point data generated by the second means, the membership function selected by the fourth means, and a control rule selected by the sixth means; and

h) eighth means for controlling an operation of a mechanism controlling the steering angle of the vehicle according to the calculated steering controlled variable.

2. An autonomous vehicle as set forth in claim 1, wherein the predetermined features include guide lines along the road surface and an obstacle present on the road surface.

3. An autonomous vehicle as set forth in claim 2, wherein said predetermined group of running scenes includes an obstacle running scene and the third means has means for determining, based on characterizing point data generated by the second means, whether an obstacle is present on the road surface corresponding to the image data from the first means, whether a distance (l) from the vehicle to the obstacle is shorter than a predetermined distance (α

), and based on said obstacle determination and said distance determination, indicates if the image output by said first means is the obstacle running scene.

4. An autonomous vehicle as set forth in claim 3, wherein the third means calculates the distance (l) using the following equation:

space="preserve" listing-type="equation">(x-xv).sup.2 +y.sup.2 =l.sup.2

, wherein y is a distance along first axis parallel to a forwarding direction of the vehicle, x is a distance from a reference point to the obstacle along a second axis perpendicular to the first axis of y, and xv is a distance along said second axis from said reference point to a point on the vehicle.

5. An autonomous vehicle as set forth in claim 4, wherein said predetermined group of running scenes includes a straight running road scene and a curved running road scene and the third means has means for determining, based on one or both of the determined distance being equal or greater than the predetermined distance and no obstacle being determined within the image, whether said guide lines are substantially straight and, based on said determination, indicates whether the image output by the first means is the straight running road scene or the curved road running scene.

6. An autonomous vehicle as set forth in claim 5, wherein the second means outputs a group of points including (x₀, y₀), (x₁, y₁), (x₂, y₂), and (x₃, y₃), corresponding to an image on said image plane of a left guide line along the road surface and the third means determines whether the vehicle is running on a straight road based on a gradient function defined as f(i)=(y_i -y₀)/(x_i, x₀) where i--1, 2, and 3, and based on f(1)-f(3) and f(2)-f(3).

7. An autonomous vehicle as set forth in claim 6, which further comprises ninth means for storing the reference membership function with respect to deviation quantities Δ

X, Δ

θ

on a position of the vehicle and angle of the vehicle with respect to the running road.

8. An autonomous vehicle as set forth in claim 7, wherein the membership function on the positional deviation quantity is such that five kinds of magnitude classification deviation quantities X:

"-B", "-M", "S", "+M", and "+B" are defined, with B being large, M being intermediate, and S being small, and a membership value μ

_x (0 to

     1) indicates a certainty accomodating each classification.

9. An autonomous vehicle as set forth in claim 8, wherein maximum evaluation values at the straight road running scene are set as reference values for maximum evaluation values ±

MaxΔ

X in both directions.

10. An autonomous vehicle as set forth in claim 9, wherein the membership function on the angle quantity is such that five kinds of magnitude classification deviation quantities θ

;

"-B", "-M", "S", "+M", and "+B" are defined and maximum evaluation values ±

MaxΔ

θ

are set.

11. An autonomous vehicle as set forth in claim 10, wherein the membership functions represent the certainties of accommodations for the respective classifications at intersections of respective triangular shaped functions corresponding to input deviation quantities Δ

X,Δ

Θ

.

12. An autonomous vehicle as set forth in claim 11, wherein the fourth means selects the membership function according to the identified running scene, and changes the maximum evaluation values MaxΔ

X, MaxΔ

Θ

using a predetermined shape of the reference membership function.

13. An autonomous vehicle as set forth in claim 12, wherein the fourth means changes the maximum evaluation values MaxΔ

X₁, MaxΔ

Θ

₁ when the third means identifies a straight running road scene, MaxΔ

X₂, MaxΔ

Θ

₂ when the third means identifies a curved running road scene, MaxΔ

X₃, MaxΔ

Θ

₃ when the third means identifies a obstacle running road scene.

14. An autonomous vehicle as set forth in claim 13, wherein the control rule is selected by the sixth means according to 25 kinds of combinations of the deviation quantities Δ

X, Δ

Θ

on the position and angle based on a rule in the form of "if Δ

X is X and Δ

Θ

is Θ

then Δ

S=A (X, Θ

)".

15. An autonomous vehicle as set forth in claim 14, wherein the seventh means calculates a displacement instruction Δ

S_T of the steering angle on a route data using the following equation;

space="preserve" listing-type="equation">Δ

S.sub.T =drfΘ

.sub.A +(1-drf)Θ

.sub.B -Θ

,

wherein Θ

denotes the measured steering angle, drf denotes a function value defined such that 1 is indicated when a remaining distance r to a target point B is a constant distance r₀ and varies from 1 to 0 as the remaining distance r becomes smaller than the constant distance r₀ (where drf=min{1, r/r₀ }).

16. An autonomous vehicle as set forth in claim 15, wherein the seventh means automatically sets an intermediate target point C from the vehicle speed to a predetermined distance L_m set according to a vehicle speed with respect to the measured steering angle of the vehicle and calculates a lateral direction deviation quantity Δ

X and a deviation quantity Δ

Θ

of the steering angle with respect to a target position X_c and target posture Θ

_c at the target point C using the following equation;

space="preserve" listing-type="equation">L=L.sub.0 +0.25V

space="preserve" listing-type="equation">Δ

X-X.sub.C -X

space="preserve" listing-type="equation">Δ

0=Θ

.sub.c -Θ

, wherein L₀ denotes a constant distance, X denotes a predicted position at the target point C based on the current measured steering angle.

17. An autonomous vehicle as set forth in claim 16, wherein the seventh means derives the membership values μ

_x, μ

.sub.θ

of certainties corresponding to each classification "-B", "-M", "S", "+M" and "+B" with respect to each deviation quantity Δ

X, Δ

θ

.

18. An autonomous vehicle as set forth in claim 17, wherein the seventh means calculates a weight average of the control rules selected by the sixth means based on a membership value indicated by a membership function selected by said fourth means and for the displacement instruction Δ

S_E for the fuzzy control S_g.

19. An autonomous vehicle as set forth in claim 18, wherein the seventh means derives a sum of the displacement instructions Δ

S_T and Δ

S_g ;

Δ

S_T +Δ

S_g.

20. An autonomous vehicle as set forth in claim 19, wherein the eight means derives the steering control of the vehicle so that the displacement Δ

S is given to a present steering angle Θ

.

21. An autonomous vehicle as set forth in claim 10, wherein the membership functions are expressed in discrete data and the discrete data of the membership values μ

are searched according to the deviation quantities α

(Δ

X, Δ

θ

).

22. An method for autonomously operating a vehicle, comprising the steps of:

a) imaging a road surface on which the vehicle runs, and generating an associated image data;

b) detecting a plurality of predetermined features contained within the image data output by step a) and generating characterizing point data corresponding to said plurality of predetermined features;

c) classifying the image data output by said step a) as a running scene among a group of predetermined running scenes on the basis of the characterizing point data generated by step b);

d) selecting at least one among a group of predetermined reference membership functions according to the classification of the running scene, said membership function indicating a certainty of correspondence between a measured steering deviation quantity and a member of a group of predetermined deviation magnitudes, said measured steering deviation quantity being a difference between a measured vehicle steering angle and a predetermined vehicle steering angle;

e) measuring a steering angle of the vehicle;

f) selecting, based on a reference membership function selected by step d) and a measured steering deviation quantity operated on by said selected reference membership function, at least one among a group of predetermined control rules;

g) calculating a steering controlled variable of the vehicle on the basis of the measured steering angle of the vehicle, characterizing point data generated by step b), the membership function selected by step d), and a control rule selected by step f); and

h) controlling an operation of a mechanism controlling the steering angle of the vehicle according to the calculated steering controlled variable.