Method for the probabilistic estimation of noisy measurements

• US 6,188,350 B1
• Filed: 09/01/1999
• Issued: 02/13/2001
• Est. Priority Date: 09/02/1998
• Status: Expired due to Fees

• Current Assignee: DaimlerChrysler Rail Systems GmbH (Bombardier Incorporated)
• Original Assignee: ABB Daimler-Benz Transportation (Technology) GmbH
• Inventors: Gründer, Thomas, Neumerkel, Dietmar
• Primary Examiner(s): Sotomayor, John B.
• Application Number:
  US09/387,460
• Time in Patent Office: 531 Days
• Field of Search: 342/104, 342/89, 342/98, 342/99, 342/108, 342/115, 342/159, 342/162, 342/168, 342/195
• US Class Current: 342/104
• CPC Class Codes: G01D 1/14 giving a distribution function of a value, i.e. number of times the value comes within specified ranges of amplitude
• CPC Class Codes: G01D 3/032 affecting incoming signal, e.g. by averaging; gating undesired signals
• CPC Class Codes: G01D 3/08 with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown

Abstract

In a method for the probabilistic estimation of measurements, based on a measurement signal in which an interference signal is superimposed on the value to be measured, the measurement signal is sampled at specified chronological intervals. A defined measurement range associated with the value to be measured is divided into discrete values and a model is formed of a process on which the measurement signal is based with discrete states that correspond to the discrete values of the measurement range. In the model, a probability value of the occurrence of each state is assigned for each sampling time, and the value to be measured is determined on the basis of the probability value of at least one state in this model. In addition, for each state of the model at a sampling time, a probability for the state to remain in its current state is determined, as well as a probability for the state to change to another state by the next sampling time. On the basis of the value of the measurement signal sampled in a current sampling time, the probability value for the occurrence of the states in the preceding sampling time and the probability for each state to remain unchanged and to change to another state between the two sampling times, the probability values for the occurrence of the states of the model for the current sampling time are recalculated.

Claims

1. A method for the probabilistic estimation of noisy measurements on the basis of a measurement signal in which a noise signal is superimposed on the value to be measured, said method comprising the steps of:

associating a defined measurement range with the value to be measured;

sampling the measurement signal at specified chronological intervals;

dividing the measurement range into discrete values;

forming a model of a process on which the measurement signal is based with discrete states that correspond to the discrete values of the measurement range;

and whereby at each sampling time, assigning a probability value of the occurrence of each state to each state, determining the value to be measured on the basis of the probability value of at least one state, assigning a probability for the state to remain unchanged to each state, assigning a probability for the state in question to change to another state by the next sampling time, and recalculating based on the sample value of the measurement signal at the current sampling time the probability values for the occurrence of the states in the preceding sampling time, the probabilities for each state to remain unchanged and to change to another state between the two sampling times, and the probability values for the occurrence of the states of the model for the current sampling time are recalculated.

2. The method as claimed in claim 1, wherein the probability values associated with the sampled value of the measurement signal for the occurrence of the states at a sampling time are determined on the basis of a Gaussian function.

3. The method as claimed in claim 1, wherein the value to be measured is determined on the basis of the highest probability value for the occurrence of a state in the respective sampling time.

4. The method as claimed in claim 1, wherein the value to be measured is determined on the basis of the sum weighted with the probabilities for the occurrence of a plurality of states in the respective sampling time.

5. The method as claimed in claim 1, wherein the measurement signal is represented by an analog signal.

6. The method as claimed in claim 1, wherein the measurement signal is generated by a speed sensor of a railway vehicle.

7. The method as claimed in claim 6, wherein the measurement signal is generated by a speed-of-rotation sensor (gyro sensor) for the movement of a railway vehicle.

8. The method as claimed in claim 6, wherein the value measured is used to determine the curvature of rails used as track for railway vehicles.

9. The method as claimed in claim 2, wherein the value to be measured is determined on the basis of the highest probability value for the occurrence of a state in the respective sampling time.

10. The method as claimed in claim 2, wherein the value to be measured is determined on the basis of the sum weighted with the probabilities for the occurrence of a plurality of states in the respective sampling time.

11. The method as claimed in claim 2, wherein the measurement signal is represented by an analog signal.

12. The method as claimed in claim 3, wherein the measurement signal is represented by an analog signal.

13. The method as claimed in claim 4, wherein the measurement signal is represented by an analog signal.

14. The method as claimed in claim 2, wherein the measurement signal is generated by a speed sensor of a railway vehicle.

15. The method as claimed in claim 3, wherein the measurement signal is generated by a speed sensor of a railway vehicle.

16. The method as claimed in claim 4, wherein the measurement signal is generated by a speed sensor of a railway vehicle.

17. The method as claimed in claim 5, wherein the measurement signal is generated by a speed sensor of a railway vehicle.

18. The method as claimed in claim 14, wherein the measurement signal is generated by a speed-of-rotation sensor (gyro sensor) for the movement of a railway vehicle.

19. The method as claimed in claim 15, wherein the measurement signal is generated by a speed-of-rotation sensor (gyro sensor) for the movement of a railway vehicle.

20. The method as claimed in claim 7, wherein the value measured is used to determine the curvature of rails used as track for railway vehicles.