Method of Kalman filtering for estimating the position and velocity of a tracked object

• US 5,051,751 A
• Filed: 02/12/1991
• Issued: 09/24/1991
• Est. Priority Date: 02/12/1991
• Status: Expired due to Fees

• Current Assignee: the united states of america as represented by the secretary of the navy
• Original Assignee: the united states of america as represented by the secretary of the navy
• Inventors: Gray, Roger L.
• Primary Examiner(s): Hellner, Mark
• Application Number:
  US07/654,111
• Time in Patent Office: 224 Days
• Field of Search: 342/107, 342/106, 342/75, 342/354, 235/412, 235/413, 364/516
• US Class Current: 342/107
• CPC Class Codes: G01S 13/723 by using numerical data
• CPC Class Codes: H03H 17/0257 KALMAN filters

Abstract

A method of Kalman filtering for estimating the position and velocity of aracked object is provided. A Kalman filter is initialized with at least position and velocity error states in an inertial computational frame. Sensor measurements are used to develop a measured line-of-sight vector to the object. Matrix transformations are used to analytically rotate the sensor measurements into a measurement frame. The measurement frame is defined as having one axis pointing towards the estimated relative position of the object. The use of the measurement frame allows the method to be adaptable to any line-of-sight/computational frame geometry. Since statistical correlation of the measurements is not present in the measurement frame, the number of computations at each filter update is reduced.

Claims

1. In a tracking system having a sensor mounted on a vehicle for sensing the presence of an object, a method of Kalman filtering for estimating the position and velocity of the object, comprising the steps of:

a. initializing a Kalman filter with at least position and velocity error states;

b. providing trajectory models of the vehicle and object;

c. establishing a measurement frame of reference for the vehicle and object using the trajectory models wherein the measurement frame is defined as having one axis pointing towards the estimated relative position of the object;

d. receiving sensor measurements to develop at least a measured line-of-sight vector to the object;

e. transforming the sensor measurements into the measurement frame;

f. updating the Kalman filter with the transformed sensor measurements for estimating the position and velocity of the tracked object;

g. correcting the trajectory models based on the estimated position and velocity;

h. resetting the position and velocity error states to zero;

j. propagating the trajectory models and the filter'"'"'s covariance matrix to the next time step; and

k. performing steps (c) through (j) as an iterative Kalman filter process.

2. A method as in claim 1 further including the step of explicitly measuring range to the object.

3. A method as in claim 1 wherein said step of initializing includes the steps of:

a. providing a priori position and velocity error state information;

b. receiving a first measurement from the sensor, the first sensor measurement including relative cross-range position information about the object;

c. constructing the Kalman filter'"'"'s covariance matrix to reflect the a priori position error state information and the relative cross-range position information from the first sensor measurement; and

d. correcting the initial position estimate in accordance with the first sensor measurement.

4. A method as in claim 1 wherein said step of initializing includes the steps of:

a. providing a priori position and velocity error state information;

b. receiving a first measurement from the sensor to establish the measured line-of-sight vector to the object, the first sensor measurement including relative cross-range position information about the object;

c. constructing a frame of reference having one axis along the measured line-of-sight vector;

d. constructing the Kalman filter'"'"'s covariance matrix relative to the constructed frame of reference to reflect the a priori position error state information and the relative cross-range position information from the first sensor measurement;

e. analytically rotating the covariance matrix into the Kalman filter'"'"'s inertial computational frame; and

f. correcting the trajectory models in accordance with the first sensor measurement.

5. A method as in claim 1 wherein the Kalman filter'"'"'s state vector and covariance matrix are in an inertial computational frame of reference and wherein said step of updating includes the step of computing the Kalman filter'"'"'s measurement matrix using a transformation matrix to relate the measurement frame to the inertial computational frame.

6. In a tracking system having a sensor mounted on a vehicle for sensing the presence of an object, a method of Kalman filtering for estimating the position and velocity of the object, comprising the steps of:

a. providing the tracking system with initial estimates of velocity error state information;

b. initializing a Kalman filter with a priori position and velocity error state information;

c. receiving a first measurement from the sensor, the first sensor measurement including relative cross-range position information about the object;

d. constructing the Kalman filter'"'"'s covariance matrix to reflect the a priori position error state information and the relative cross-range position information from the first sensor measurement; and

e. correcting the initial position estimate in accordance with the first sensor measurement;

f. deriving a measurement frame of reference for the vehicle and object wherein the measurement frame is defined as having one axis pointing towards the estimated relative position of the object;

g. receiving sensor measurements to develop at least a measured line-of-sight vector to the object;

h. transforming the sensor measurements into the measurement frame; and

i. updating the Kalman filter with the transformed sensor measurements for estimating the position and velocity of the tracked object.

7. In a tracking system having a sensor mounted on a vehicle for sensing the presence of an object, a method of Kalman filtering for estimating the position and velocity of the object, comprising the steps of:

a. providing the tracking system with initial estimates of position and velocity of the object;

b. initializing a Kalman filter with a priori position and velocity error state information;

c. receiving a first measurement from the sensor to establish the measured line-of--sight vector to the object the first sensor measurement including relative cross-range position information about the object;

d. constructing a frame of reference having one axis along the measured line-of-sight vector;

e. constructing the Kalman filter'"'"'s covariance matrix relative to the constructed frame of reference to reflect the a priori position error state information and the relative cross-range position information from the first sensor measurement;

f. analytically rotating the covariance matrix into the Kalman filter'"'"'s inertial computational frame;

g. correcting the initial position estimate in accordance with the first sensor measurement;

h. deriving a measurement frame of reference for the vehicle and object wherein the measurement frame is defined as having one axis pointing towards the estimated relative position of the object;

i. receiving sensor measurements to develop at least a measured line of sight vector to the object;

j. transforming the sensor measurements into the measurement frame; and

k. updating the Kalman filter with the transformed sensor measurements for estimating the position and velocity of the tracked object.

8. In a tracking system having a sensor mounted on a vehicle for sensing the presence of an object, a method of Kalman filtering for estimating the position and velocity of the object, comprising the steps of:

a. referencing the Lakman filter'"'"'s state vector and covariance matrix in an inertial computational frame of reference;

b. providing the tracking system with initial estimates of position and velocity of the object;

c. initializing the Kalman filter with at least position and velocity error states;

d. deriving a measurement frame of reference for the vehicle and object wherein the measurement frame is defined as having one axis pointing towards the estimated relative position of the object;

e. receiving sensor measurements to develop at least a measured line-of-sight vector to the object;

f. transforming the sensor measurements into the measurement frame;

g. updating the Kalman filter with the tansformed sensor measurements and computing the Kalman filter'"'"'s measurement magic using a transformation matrix to relate the measurement frame to the inertial computational frame;

h. providing trajectory models of the vehicle and object;

i. correcting the trajectory models based on the estimated position and velocity error states;

j. resetting the position and velocity error states to zero; and

propagating the trajectory models and the filter'"'"'s covariance matrix to the next time step.