Process for detecting the road conditions ahead for motor vehicles

• US 6,300,865 B1
• Filed: 04/22/1999
• Issued: 10/09/2001
• Est. Priority Date: 05/08/1996
• Status: Expired due to Fees

• Current Assignee: DaimlerChrysler AG (Mercedes-Benz Group AG)
• Original Assignee: DaimlerChrysler AG (Mercedes-Benz Group AG)
• Inventors: Fechner, Thomas, Hach, Rainer
• Primary Examiner(s): TONG, NINA C
• Application Number:
  US09/180,564
• Time in Patent Office: 901 Days
• Field of Search: 340/436, 340/435, 340/901, 340/903, 340/904, 180/167, 180/168, 180/169, 180/170, 342/70, 342/71, 342/72, 367/909
• US Class Current: 340/436
• CPC Class Codes: B62D 1/28 non-mechanical , e.g. following a line or other known markers
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01S 2013/93271 in the front of the vehicles
• CPC Class Codes: G05D 1/0257 using a radar radar systems designed for anti-collision purposes between land vehicles or between land vehicle and fixed obstacles G01S13/931
• CPC Class Codes: Y10S 367/909 Collision avoidance

Abstract

A process for recognizing the course of the road ahead for a motor vehicle shall be provided, in which the velocity of the vehicle is determined and the position and the velocity of objects located ahead are determined by rigid search antennas oriented in the direction of travel via a radar sensor system, and which makes possible a simple, forward-looking recognition of the course of the road by special evaluation of the measured position and velocity data available. The roadside-specific targets are filtered out of the fixed target by the threshold value comparison of the fixed target amplitudes, and the current distances between the vehicle and the side of the road are determined in discrete angle ranges by ordering filtration. The distances from the side of the road, estimated for each angle range, are sent as characteristics to a classifier for the prediction of the curve type lying ahead, and they are used as support values for a nonlinear regression to obtain a curvature parameter of the curve.

Claims

1. A process for recognizing the course of the road ahead for a motor vehicle, the process comprising the steps of;

determining the velocity of the motor vehicle;

determining the position and velocity of objects located ahead of the vehicle by a radar sensor system with rigidly oriented antennas;

separating fixed objects out of the objects located ahead of the vehicle, said separating being performed in a plurality of discrete angle ranges;

determining current distances between a side of the road and the motor vehicle from the fixed objects and using the determined current distances as characteristics for a nonlinear regression;

determining the objects which are specific of the side of the road from among the fixed objects in the discrete angle ranges by a threshold value comparison of radar target amplitudes;

filtering, with an ordering filter, the object with a shortest distance out of the roadside-specific objects determined in each of the discrete angle ranges over a plurality of consecutive measurement cycles;

using the filtered object from said ordering filter to determine the distance between the side of the road and the motor vehicle in each of the discrete angle range; and

using the objects determined in the discrete angle ranges as characteristics for an adaptive prediction of a curve type using a curve classifier.

2. The process in accordance with claim 1, wherein a model-based Kalman filter, whose model is based on a priori knowledge of dynamics of courses of roads, is used to smoothen said distances between the side of the road and the motor vehicle.

3. The process in accordance with claim 2, wherein a correction signal, which is obtained by comparison of a current curve type of the motor vehicle with a curve type which is predicted by the curve classifier and is time-delayed corresponding to the current velocity of the motor vehicle is sent continuously to the curve classifier.

4. The process in accordance with claim 2, wherein the curve classifier is formed by a neuronal network.

5. The process in accordance with claim 1, wherein a correction signal, which is obtained by comparison of a current curve type of the motor vehicle with a curve type which is predicted by the curve classifier and is time-delayed corresponding to the current velocity of the motor vehicle is sent continuously to the curve classifier.

6. The process in accordance with claim 5, wherein the curve classifier is formed by a neuronal network.

7. The process in accordance with claim 1, wherein the curve classifier is formed by a neuronal network.

8. The process in accordance with claim 7, wherein weights of the neuronal network are preset once by an off-line adaptation based on example situations.

9. The process in accordance with claim 7, wherein the weights of the neuronal network, predetermined on a basis of example situations before being put into operation, are corrected by the correction signal during operation.