Path prediction system and method

• US 6,675,094 B2
• Filed: 09/07/2001
• Issued: 01/06/2004
• Est. Priority Date: 09/08/2000
• Status: Expired due to Term

• Current Assignee: Valeo Radar Systems, Inc. (Valeo SE)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Firda, John M., Russell, Mark E., Delcheccolo, Michael Joseph
• Primary Examiner(s): NGUYEN, THU V
• Application Number:
  US09/949,295
• Publication Number:
  US 20020049539A1
• Time in Patent Office: 851 Days
• Field of Search: 701/1, 701/26, 701/116, 701/119, 701/121-122, 701/300-302, 342/454, 342/455, 342/458, 342/461, 340/903, 340/901
• US Class Current: 701/301
• CPC Class Codes: B60W 2050/0052 Filtering, filters
• CPC Class Codes: B60W 2520/14 Yaw
• CPC Class Codes: B60W 50/0097 Predicting future conditions
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01S 2013/932 using own vehicle data, e.g. ground speed, steering wheel direction
• 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: G05D 1/027 comprising intertial navigation means, e.g. azimuth detector inertial navigation G01C21/16; inertial navigation combined with non-inertial navigation instruments G01C21/165

Abstract

There is disclosed herein a method for detecting objects in a predicted path of a host vehicle moving on a highway lane. The method is for use in a system including a forward looking sensor, preferably a radar system, for providing range, angle and velocity data for objects within a field of view in front of the host vehicle, measuring systems for providing velocity and yaw rate data for the host vehicle, and a processing system responsive to the forward looking sensor and the measuring systems. The method includes the unordered steps of (a) calculating an estimated path of the host vehicle based on its velocity and yaw rate; (b) calculating estimated paths for each of the objects; (c) determining the lateral distance of each object from the predicted path of the host vehicle; and (d) classifying each object as either in or out of the highway lane of the host vehicle. The method includes processes for detecting lane changes by the host vehicle and for generating alternative path hypotheses in response to target deviations from the predicted paths.

Claims

1. A system for detecting objects in a predicted path of a vehicle moving on a highway lane, comprising:

a forward looking sensor for providing range, angle and velocity data for objects within a field of view in front of said vehicle;

measuring systems for providing velocity and yaw rate data for said vehicle; and

a processing system responsive to said forward looking sensor and said measuring systems to calculate an estimated path of said vehicle based on its velocity and yaw rate while said vehicle traverses the highway lane, and to calculate estimated paths for each of said objects sensed within the field of view in front of said vehicle, and to determine a lateral distance of each object sensed within the field of view in front of said vehicle from the predicted path of said vehicle while said vehicle traverses the highway lane, and classifying each object sensed within the field of view in front of said vehicle as either in or out of the highway lane of said vehicles, wherein the predicted path of said vehicle is generated by correlating the estimated path of said vehicle and said objects and fusing the paths of said objects as a weighted average.

2. The object detecting system in accordance with claim 1 wherein said forward looking sensor comprises a radar system.

3. The object detecting system in accordance with claim 1 wherein said yaw rate measuring system comprises a gyrocompass.

4. In a system for use in a vehicle, said system including a forward looking sensor for providing range, angle and velocity data for objects within a field of view in front of said vehicle, measuring systems for providing velocity and yaw rate data for said vehicle, and a processing system responsive to said forward looking sensor and said measuring systems, a method for detecting objects in a predicted path of said vehicle while moving on a highway lane, said method comprising the steps of:

a. calculating an estimated path of said vehicle based on its velocity and yaw rate while said vehicle traverses the highway lane;

b. calculating estimated paths for each of said objects sensed within the field of view in front of said vehicle;

c. determining a lateral distance of each object sensed within the field of view in front of said vehicle from the predicted path of said vehicle; and

d. classifying each object sensed within the field of view in front of said vehicle as either in or out of the highway lane of said vehicle while said vehicle traverses the highway lane, wherein the predicted path of said vehicle is generated by correlating the estimated path of said vehicle and said objects and fusing the paths of said objects as a weighted average.

5. The method in accordance with claim 4 wherein said method is repeated between 10 and 20 times per second.

6. In a system for use in a host vehicle, said system including a forward looking radar system for providing range, angle and velocity data for targets within a field of view in front of said vehicle, measuring systems for providing velocity and yaw rate data for said vehicle, and a processing system responsive to said forward looking radar system and said measuring systems, a method for detecting targets in a predicted path of said host vehicle while moving on a highway lane, said method comprising:

(a) calculating a host vehicle path estimate from said velocity and yaw rate data for said host vehicle while said host vehicle traverses the highway lane;

(b) propagating target position histories with said host vehicle path estimate;

(c) propagating target positions forward with longitudinal and lateral target position state vectors;

(d) calculating polynomial curve fits for host vehicle and target position state vectors;

(e) generating said predicted path by correlating said estimated path of said host vehicle and target paths and then fusing the target paths as a weighted average;

(f) comparing target cross range positions to said predicted path of said host vehicle on the highway lane and classifying targets as in-lane or out-of-lane with respect to the highway lane of said host vehicle;

(g) receiving updated data from said forward looking radar system; and

(h) repeating steps a through g continuously.

7. The method in accordance with claim 6 wherein calculating a host vehicle path estimate includes applying a 2-state Kalman filter to track curvature parameters and propagate a host position vector forward in range.

8. The method in accordance with claim 6 wherein propagating target position histories includes calculating a previous position of the target relative to a longitudinal axis of said host vehicle, filling a vector of history points, and dropping one or more points included in the vector of history points as a value of the longitudinal position state vector drops below zero.

9. The method in accordance with claim 6 wherein calculating a the polynomial curve fit for said host vehicle and target position state vectors includes applying a 2^nd order polynomial curve fit to target history data and host vehicle curvature rate data.

10. The method in accordance with claim 9 wherein said polynomial curve has a form

11. The method in accordance with claim 6 wherein calculating a polynomial curve fit for said target position state vectors includes propagating data forward from target position and velocity estimates in the longitudinal and lateral dimensions.

12. The method in accordance with claim 11 wherein said polynomial curve has the a form

13. The method in accordance with claim 12 wherein generating said predicted path by correlating host vehicle and target paths and then fusing the target paths as a weighted average further comprises:

adjusting fusion weights depending on the resolution of an accepted alternate path hypothesis, or

using the host vehicle derived path if an alternate path hypothesis has been rejected and there is only a single target, or

adjusting fusion weights depending on the resolution of an rejected alternate path hypothesis where there is more than one target, or

applying default, range-based fusion weights where no alternate path hypothesis has been rejected;

computing the weighted average of the target paths; and

extracting the coefficients of said polynomial.

14. The method in accordance with claim 6 wherein generating said predicted path includes testing for a highway lane change by said host vehicle.

15. The method in accordance with claim 14 wherein generating said predicted path includes comparing the unfiltered yaw rate measurement of said host vehicle to a previously updated path heading coefficient multiplied by twice the velocity of said host vehicle.

16. The method in accordance with claim 15 wherein a highway lane change by said host vehicle is confirmed by:

comparing the integrated yaw rate to each target heading coefficient to see if it is equal and opposite;

confirming that the majority of targets are moving in the opposite direction of said host vehicle; and

noting that said opposite direction motion has occurred at two consecutive data updates from said forward looking radar system.

17. The method in accordance with claim 6 wherein generating said predicted path includes calculating default target fusion weights.

18. The method in accordance with claim 17 wherein said default target fusion weights are assigned a value of one for the range between said host vehicle and the present position of said target, and monotonically decreasing values beyond said target to the end of the range of said forward looking radar system.

19. The method in accordance with claim 6 wherein comparing target cross range positions to said predicted path includes comparing the target positions to within one-half highway lane width of the predicted path to determine in-lane classification.

20. The method in accordance with claim 6 wherein generating said predicted path by correlating host vehicle and target paths and then fusing the target paths as a weighted average further comprises the process of setting an alternate path hypothesis when the target furthest in range deviates from the predicted path, said alternate path hypothesis continuing to follow the predicted path.