Adaptive anti-collision method for vehicle

• US 9,254,824 B2
• Filed: 12/30/2013
• Issued: 02/09/2016
• Est. Priority Date: 12/30/2013
• Status: Active Grant

• Current Assignee: Automotive Research & Testing Center
• Original Assignee: Automotive Research & Testing Center
• Inventors: Ko, Ming-Kuan, Yao, Chi-Chun, Huang, Shih-Chieh
• Primary Examiner(s): Khatib, Rami
• Assistant Examiner(s): NESLEY, TIMOTHY ALAN
• Application Number:
  US14/143,279
• Publication Number:
  US 20150183410A1
• Time in Patent Office: 771 Days
• Field of Search: 701 36- 49, 340901-905
• US Class Current: 1/1
• CPC Class Codes: B60T 2201/022 Collision avoidance systems
• CPC Class Codes: B60T 2220/02 Driver type; Driving style; Driver adaptive features
• CPC Class Codes: B60T 7/22 initiated by contact of vehicle, e.g. bumper, with an external object, e.g. another vehicle , or by means of contactless obstacle detectors mounted on the vehicle
• CPC Class Codes: B60T 8/174 characterised by using special control logic, e.g. fuzzy logic , neural computing
• CPC Class Codes: B60T 8/17558 specially adapted for collision avoidance or collision mitigation road vehicle drive control systems for collision avoidance otherwise than by controlling a particular sub-unit B60W30/09
• CPC Class Codes: B60T 8/58 responsive to speed and another condition or to plural speed conditions

Abstract

An adaptive anti-collision method for vehicles has steps of creating multiple driving patterns with each driving pattern corresponding to a vehicle speed, a safe distance and a braking distance parameter, such as longer safe distance configured for faster vehicle speed, and higher vehicle speed or shorter safe distance for different road condition, acquiring dynamic information, such as vehicle speed or acceleration, of the vehicle using sensors on the vehicle, combining the dynamic information and drivers'"'"' driving behavior to determine a driving pattern through a statistical analysis and a neural network, adjusting control parameters of the vehicle according to the driving pattern for an electronic control unit of the vehicle to issue an alert or activate a braking action according to the driving pattern. Accordingly, the anti-collision method can be adapted to different vehicle speed, road condition and drivers'"'"' driving habits for adjusting the safe distance and the braking system.

Claims

1. An adaptive anti-collision method for vehicles, wherein each vehicle having multiple sensors and a control unit, the sensors acquire a vehicle speed and a measured distance between the vehicle and an obstacle in the front, the control unit is electrically connected to an electronic control unit (ECU) of the vehicle and the sensors, and the adaptive anti-collision method is performed by the control unit and comprises steps of:

creating multiple driving patterns and indices or alert levels corresponding to the driving patterns, wherein each driving pattern has a vehicle speed, a safe distance and a reference braking distance related parameter, and the driving pattern at the vehicle speed corresponds to a comparison relationship between the measured distance and the safe distance and one of four safe levels;

analyzing the vehicle speed and the measured distance acquired by the sensors to determine one of the driving patterns, wherein actual driving conditions of the vehicle are statistically analyzed to acquire a driving state of a driver of the vehicle and classify into a corresponding driving pattern associated with the driver so as to generate one of the indices or the alert levels corresponding to the driving pattern and vary control parameters of the vehicle according to the classified driving patterns, wherein the index generated according to the driving patterns is calculated by

2. The method as claimed in claim 1, wherein the driving pattern of the driver is analyzed with a neural network.

3. The method as claimed in claim 2, wherein the neural network is calculated by a self-organizing map (SOM) algorithm, and the SOM algorithm includes steps of:

inputting an initial weighting matrix (IW);

introducing an input training vector;

calculating a distance and a winning node, wherein the winning node has the shortest distance between an output unit (j) and the input training vector;

calculating a weighting value correction matrix (Δ

W↓

j), wherein the weighting value correction matrix is a difference between a weighting matrix at next time and a weighting matrix at present; and

updating the IW;

wherein

a radius of a neighboring center (q) is expressed by;

R<sup>n+1=λ

×</sup>

Rⁿ

where

R is a radius of a neighboring center (q); and

λ

is a radius adjustment factor;

a neighboring distance is expressed by;

dis_jq=∥

r_j−

r_q∥

=√

{square root over ((x_j−

x_q)²+(y_j−

y_q)²)}{square root over ((x_j−

x_q)²+(y_j−

y_q)²)}

where

dis_jq is a distance between the output unit (j) and the neighboring center (q);

r_j is a topology position (x_j, y_j) of the output unit (j); and

r_q is the topology position (x_q, y_q) of the neighboring center (q);

a neighboring function is expressed by;

4. The method as claimed in claim 3, wherein the driving pattern is classified into two types respectively corresponding to a safe level and a slightly dangerous level, the safe level is classified when the measured distance is greater than the safe distance, and the slightly dangerous level is classified when the measured distance is less than the safe distance.

5. The method as claimed in claim 4, wherein the driving pattern is further classified into two additional types respectively corresponding to a very safe level and a very dangerous level, the very safe level is classified when the measured distance is greater than double of the safe distance, the very dangerous level is classified when the measured distance is less than two thirds of the safe distance, and the slightly dangerous level is classified when the measured distance is within two thirds of the safe distance and the safe distance.

6. The method as claimed in claim 1, wherein after the measured distance and the vehicle speed detected by the sensors are analyzed to classify into a driving pattern, the measured distance between the obstacle and the vehicle is further updated, and if the measured distance is less than the index of the driving pattern of the driver, an alert is issued to warn the driver.

7. The method as claimed in claim 6, wherein after the alert is issued to warn the driver, if the measured distance is still less than the index and the driver has not responded to avoid the obstacle, the control unit automatically activates a braking system of the vehicle to stop the vehicle.

8. The method as claimed in claim 1, wherein the driving pattern is classified into two types respectively corresponding to a safe level and a slightly dangerous level, the safe level is classified when the measured distance is greater than the safe distance, and the slightly dangerous level is classified when the measured distance is less than the safe distance.

9. The method as claimed in claim 8, wherein the driving pattern is further classified into two additional types respectively corresponding to a very safe level and a very dangerous level, the very safe level is classified when the measured distance is greater than double of the safe distance, the very dangerous level is classified when the measured distance is less than two thirds of the safe distance, and the slightly dangerous level is classified when the measured distance is within two thirds of the safe distance and the safe distance.