Satellite based on-board vehicle navigation system including predictive filtering and map-matching to reduce errors in a vehicular position

• US 6,615,135 B2
• Filed: 05/24/2001
• Issued: 09/02/2003
• Est. Priority Date: 05/24/2001
• Status: Expired due to Term

• Current Assignee: Northrop Grumman Systems Corporation (Northrop Grumman Corporation)
• Original Assignee: PRC, Inc. (Arthur D. Little, Inc.)
• Inventors: Davies, F. Bryan
• Primary Examiner(s): Nguyen, Tan Q.
• Assistant Examiner(s): Tran, Dalena
• Application Number:
  US09/863,883
• Publication Number:
  US 20020177950A1
• Time in Patent Office: 831 Days
• Field of Search: 701/213, 701/215, 701/207, 701/201, 701/205, 701/210, 701/224, 701/209, 701/214, 701/200, 342/357.13, 342/357.01, 342/357.03, 342/357.02, 342/357.07, 340/988, 340/990, 340/995
• US Class Current: 701/213
• CPC Class Codes: G01C 21/30 Map- or contour-matching
• CPC Class Codes: G01S 19/50 whereby the position solution is constrained to lie upon a particular curve or surface, e.g. for locomotives on railway tracks
• CPC Class Codes: G01S 5/011 Identifying the radio environment

Abstract

An accurate position of a mobile receiver is determined using a received positional signal having relatively low and high frequency noise components corrupting the positional accuracy of the positional signal. A low frequency error associated with a detected turn of the mobile receiver is derived and applied as a low frequency correction factor to positions of the receiver derived after the turn is detected. Resulting corrected positions are applied along with associated high frequency correction factors representative of the high frequency noise component to a predictive filter to derive predicted positions substantially free of the low and high frequency noise components.

Claims

1. A method of rapidly associating a position of a mobile receiver (MR) with a navigable route accessible in a map database, the map database including information representing predetermined navigable routes, comprising:

seeding a predictive filter with an initial value indicative of an erroneous current position value of the MR displaced a predetermined minimum distance from a known approximate position of the MR to establish an initial gain of the predictive filter;

receiving a positional signal over m time periods and deriving position values of the MR over the m time periods using the received signal;

deriving predicted position values of the MR over the m time periods by predictively filtering the position values with the predictive filter; and

determining a probable position of the MR at an end of the m time periods using a current predicted position value, the probable position coinciding with a navigable route of where the MR is probably positioned at the end of the m time periods.

2. The method as recited in claim 1, wherein the positional signal comprises a plurality of GPS satellite signals wherein each of the positional signals includes errors to reduce the positional accuracy of each of the positional signals and wherein the mobile receiver is a GPS receiver.

3. The method as recited in claim 2, comprising

identifying candidate navigable routes in the map database within a predetermined distance of the current predicted position value; and

deriving a weighted distance between each of the candidate navigable routes and the current predicted position value using a shortest distance and a difference in direction between each of the candidate navigable routes and the MR, the probable position coinciding with a position on one of the candidate navigable routes associated with a minimum weighted distance.

4. The method as recited in claim 3, wherein said step of deriving a weighted distance includes the step of:

adding a variable distance to each shortest distance, the variable distance progressively increasing from a minimum distance to a maximum distance with an increasing difference in the direction between the associated candidate navigable route and the MR.

5. The method as recited in claim 4, wherein the variable distance is a sinusoidally varying function of the difference in the direction between the associated candidate navigable route and the MR.

6. The method as recited in claim 4, further including the step of storing the predicted position values in a time ordered list of predicted position values, and wherein said step of deriving a weighted distance for each of the candidate navigable routes includes the step of comparing the direction of each of the candidate navigable routes to an average direction of the MR derived using the predicted position values stored in the time ordered list.

7. The method as recited in claim 2, wherein the predictive filter is a Kalman filter, wherein the predetermined minimum distance is in the approximate range of 0.5 to 3 kilometers, and wherein m is less than 6.

8. A method of accurately determining the position of a mobile receiver (MR) based upon a received positional signal having relatively low and high frequency noise components corrupting the positional accuracy of the received signal, comprising:

(a) detecting a turn of the MR using the received positional signal and deriving a first positional offset associated with the turn and representative of the low frequency noise component;

(b) deriving a position value of the MR using the received positional signal after detecting the turn, and correcting the position value with the first positional offset associated with the turn to derive a corrected position value; and

(c) applying the corrected position value to a filter along with a second positional offset associated with the corrected position value and representative of the high frequency noise component to derive a filtered position value substantially free of the low and high frequency noise components.

9. The method of claim 8, wherein said steps (a) through (c) are executed without applying a dead reckoning algorithm that uses an external sensor other than a satellite signal receiver.

10. The method of claim 8, wherein said steps (a) through (c) are repeated continuously to derive subsequent filtered position values substantially free of the low and high frequency noise components.

11. The method of claim 8, comprising

deriving position values using the received positional signal;

predictively filtering the position values with a predictive filter to derive predicted position values; and

detecting the turn and deriving a predicted turn position using the predicted position values, the first positional offset representing a distance between the predicted turn position and a probable turn position of where the MR probably turned on a navigable route represented in a map database and selected from the map database using a map matching algorithm.

12. The method as recited in claim 11, wherein the positional signal comprises a plurality of GPS satellite signals wherein at least one of the positional signals includes the low frequency and high frequency noise component, wherein the mobile receiver is a GPS receiver, and wherein the predictive filter is a Kalman filter.

13. The method as recited in claim 12, wherein the second positional offset is applied to the Kalman filter as a positional variance error to control a Kalman gain of the Kalman filter.

14. The method as recited in claim 12, wherein the high frequency noise component includes noise frequencies in the approximate range of 2 Hz or more, and the low frequency noise includes noise caused by at least one of Selective Availability and multipath reflections.

15. The method of claim 8, wherein the second positional offset represents a shortest distance between the corrected position value and a probable navigable route coinciding with where the MR is probably located, the probable navigable route being determined by map matching a predicted position value derived by the predictive filter after the turn of the MR is detected to the navigable route.

16. A method of accurately determining the position of a mobile receiver (MR) based upon a received positional signal having low and high frequency components corrupting the positional accuracy of the received signal, comprising:

(a) receiving the positional signal over time intervals T₁ through T_m, where T represents the duration of each time interval and the subscript represents the position of each time interval in the sequence of time intervals T₁ through T_m;

(b) deriving position values of the MR at time intervals T₁ through T_m using the received positional signal;

(c) predictively filtering the position values at time intervals T₁ through T<sub>m−

1</sub> with a predictive filter to derive predicted position values of the MR at time intervals T₁ through T<sub>m−

1;</sub>

(d) deriving a first positional error indicative of a low frequency noise component reducing the accuracy of the position values and predicted position values at time intervals T₁ through T_m using predicted position values T₁ through T<sub>m−

1</sub>;

(e) correcting the position value at time interval T_m with the first positional error to derive a corrected position value at time interval T_m; and

(f) predictively filtering the corrected position value at time interval T_m to derive a predicted position of the MR at time interval T_m, the predicted position at time interval T_m being substantially free of the low frequency noise component.

17. The method as recited in claim 16, wherein the positional signal comprises a plurality of GPS satellite signals wherein at least one of the positional signals includes the low frequency and high frequency noise component, and wherein the mobile receiver is a GPS receiver.

18. The method as recited in claim 17, comprising

detecting a turn and determining a predicted turn position of the MR using the predicted position values at T₁ through T_{m− 1};

associating the predicted turn position with a probable turn position on a navigable route represented in a map database; and

calculating the first positional error as a distance between the predicted and probable turn positions.

19. The method as recited in claim 17, including the further steps

deriving a current velocity and acceleration of the MR using the corrected position value and at least one of the position values of the MR at time intervals T₁ through T_m−

1;

validating the corrected position value using a comparison between the current velocity and acceleration of the MR and a vehicle dynamics rule base including valid ranges for a velocity and an acceleration of the MR; and

supplying an averaged corrected position value to the predictive filter instead of the corrected position value when the corrected position value determined invalid by the validating step.

20. The method as recited in claim 16, wherein steps (a) through (f) are repeated for each subsequent time interval.

21. The method of claim 16, wherein said steps (a) through (f) are executed without applying a dead reckoning algorithm that uses an external sensor other than a GPS receiver.

22. The method as recited in claim 16, further including the steps of

receiving the positional signal and deriving a position value of the MR at a time interval T_m+1 using the positional signal;

correcting the position value at time interval T_m+1 with the first positional error to derive a corrected position value at time T₁₊₁;

deriving a second positional error indicative of a high frequency noise component reducing the accuracy of the position value at time interval T_m+1, the second positional error value being determined using the corrected position value at time T_m+1 and the predicted position value at T_m;

applying the second positional error to the predictive filter at time interval T_m+1 as a positional variance error to drive a gain of the predictive filter; and

predictively filtering the corrected position value at time interval T_m+1 with the predictive filter to derive a predicted position value of the MR at time interval T_m+1, the predicted position at time interval T_m+1 being substantially free of the low and high frequency noise components.

23. The method as recited in claim 22, wherein said step of deriving a second positional error includes the step of calculating a distance between the corrected position value at time T_m+1 and a probable navigable route represented in a map database on which the MR is probably located.

24. The method as recited in claim 23, further including the steps of

identifying candidate navigable routes in the map database within a predetermined distance of the predicted position value at time interval T_m;

and

deriving a weighted distance between each of the candidate navigable routes and the predicted position value at time interval T_m using a shortest distance and a difference in direction between each of the candidate navigable routes and the MR, the probable position coinciding with a position on one of the candidate navigable routes associated with a minimum weighted distance.

25. The method as recited in claim 24, wherein the high frequency noise component includes noise having a frequency of at least 1/T.

26. The method as recited in claim 25, wherein the duration of each time interval is in the approximate range of 0.5 to 2 seconds.

27. The method as recited in claim 26, wherein the low frequency noise component is caused by at least one of selective availability and multipath reflections.

28. A method of accurately determining the position of a mobile receiver (MR) based upon a received positional signal having relatively low and high frequency components corrupting the positional accuracy of the received signal, comprising:

receiving the positional signal and deriving position values of the MR using the positional signal;

applying the position values to a predictive filter to derive predicted position values of the MR;

detecting a turn of the MR using the predicted position values and deriving a first positional offset associated with the turn and representative of the low frequency noise component;

correcting position values with the first positional offset to derive corrected position values after detecting the turn;

deriving a second positional offset between each of the corrected position values and a navigable route coinciding with where the MR is probably located for each of the corrected position values, the second positional offset being representative of the high frequency noise; and

applying each of the corrected position values along with the second positional offset associated with each of the corrected position values to the predictive filter to derive predicted position values substantially free of the low frequency and high frequency noise.

29. An article, comprising:

at least one sequence of machine executable instructions;

a medium bearing the executable instructions in machine readable form,

wherein execution of the instructions by one or more processors causes the one or more processors to;

seed a predictive filter with an initial value indicative of an erroneous current position value of a mobile receiver (MR) displaced a predetermined minimum distance from a known approximate position of the MR to establish an initial gain of the predictive filter;

receive a positional signal over m time periods and derive position values of the MR over the m time periods using the received signal;

derive predicted position values of the MR over the m time periods by predictively filtering the position values with the predictive filter; and

determine a probable position of the MR at an end of the m time periods using a current predicted position value, the probable position coinciding with a navigable route of where the MR is probably positioned at the end of the m time periods.

30. An article, comprising:

at least one sequence of machine executable instructions;

a medium bearing the executable instructions in machine readable form,

wherein execution of the instructions by one or more processors causes the one or more processors to;

(a) detect a turn of a mobile receiver (MR) using a received positional signal having a positional accuracy corrupted by relatively low and high frequency noise components, and derive a first positional offset associated with the turn and representative of the low frequency noise component;

(b) derive a position value of the MR using the received positional signal after detecting the turn, and correct the position value with the first positional offset associated with the turn to derive a corrected position value; and

(c) apply the corrected position value to a filter along with a second positional offset associated with the corrected position value and representative of the high frequency noise component to derive a filtered position value substantially free of the low and high frequency noise components.

31. A computer architecture, comprising:

seeding means for seeding a predictive filter with an initial value indicative of an erroneous current position value of a mobile receiver (MR) displaced a predetermined minimum distance from a known approximate position of the MR to establish an initial gain of the predictive filter;

receiving means for receiving a positional signal over m time periods and for deriving position values of the MR over the m time periods using the received signal;

deriving means for deriving predicted position values of the MR over the m time periods by predictively filtering the position values with the predictive filter; and

determining means for determining a probable position of the MR at an end of the m time periods using a current predicted position value, the probable position coinciding with a navigable route of where the MR is probably positioned at the end of the m time periods.

32. A computer architecture, comprising:

(a) detecting means for detecting a turn of a mobile receiver (MR) using a received positional signal having a positional accuracy corrupted by relatively low and high frequency noise components, and for deriving a first positional offset associated with the turn and representative of the low frequency noise component;

(b) deriving means for deriving a position value of the MR using the received positional signal after detecting the turn, and correcting means for correcting the position value with the first positional offset associated with the turn to derive a corrected position value; and

(c) applying means for applying the corrected position value to a filter along with a second positional offset associated with the corrected position value and representative of the high frequency noise component to derive a filtered position value substantially free of the low and high frequency noise components.

33. A computer system, comprising:

a processor; and

a memory coupled to the processor, the memory having stored therein sequences of instructions, which, when executed by the processor, cause the processor to perform the steps of;

seeding a predictive filter with an initial value indicative of an erroneous current position value of a mobile receiver (MR) displaced a predetermined minimum distance from a known approximate position of the MR to establish an initial gain of the predictive filter;

receiving a positional signal over m time periods and deriving position values of the MR over the m time periods using the received signal;

deriving predicted position values of the MR over the m time periods by predictively filtering the position values with the predictive filter; and

determining a probable position of the MR at an end of the m time periods using a current predicted position value, the probable position coinciding with a navigable route of where the MR is probably positioned at the end of the m time periods.

34. A computer system, comprising:

a processor; and

a memory coupled to the processor, the memory having stored therein sequences of instructions, which, when executed by the processor, cause the processor to perform the steps of;

(a) detecting a turn of a mobile receiver (MR) using a received positional signal having a positional accuracy corrupted by relatively low and high frequency noise components, and deriving a first positional offset associated with the turn and representative of the low frequency noise component;

(b) deriving a position value of the MR using the received positional signal after detecting the turn, and correcting the position value with the first positional offset associated with the turn to derive a corrected position value; and

(c) applying the corrected position value to a filter along with a second positional offset associated with the corrected position value and representative of the high frequency noise component to derive a filtered position value substantially free of the low and high frequency noise components.