Method for the probabilistic estimation of noisy measurements

Details Method for the probabilistic estimation of noisy measurements Method for the probabilistic estimation of noisy measurements

1999-09-01

2001-02-13

Sotomayor, John B. (Department: 3662)

Communications: directive radio wave systems and devices (e.g.,

Determining velocity

C342S089000, C342S115000, C342S159000, C342S162000

Reexamination Certificate

active

06188350

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to a method for the probablistic estimation of noisy measurements.
2. Brief Description of the Prior Art
To remove superimposed noise from measurement signals, one method of the prior art uses filtering with a high-pass, low-pass, band-pass or band-stop filter. In digital signal processing, FIR (finite impulse response) or IIR (infinite impulse response) filters are used, although depending on the design of the system, there can be significant delays in the reception of the signal. This delay can only be reduced at the expense of the filtering action.
An additional method of the prior art to eliminate noise and interference from measurement signals is to use a model of the noisy system, so that the interference signal without a useful signal is generated and can be subtracted from the measurement signal, as a result of which only the useful signal is received. This method is generally very complex, time-consuming and expensive, because it requires an accurate model of the noisy system.
The object of the invention is therefore a method for the probabilistic estimation of noisy measurements on the basis of a measurement signal in which a noise or interference signal is superimposed on the value to be measured, whereby a defined measurement range is associated with the value to be measured and the measurement signal is sampled at specified chronological intervals, and in which a suppression of the interference signal can be accomplished in a relatively simple manner with a minor delay and with high efficiency.
The invention teaches that this object can be accomplished by the features disclosed in the characterizing portion of claim
1
. Advantageous refinements of the method claimed by the invention are disclosed in the subclaims.
SUMMARY OF THE INVENTION
The invention teaches that the measurement range is divided into discrete values, a model of a process on which the measurement signal is based is formed with discrete states that correspond to the discrete values of the measurement range, in which at each sampling time, a probability value of the occurrence of each state is assigned to each state, whereby the value to be measured is determined on the basis of the probability value at least one state, that at each sampling point, a probability for the state to remain unchanged is assigned to each state, as well as a probability for the state in question to change to another state by the next sampling time, and that on the basis of the sampled value of the measurement signal at a current sampling time, the probability values for the occurrence of the states in the preceding sampling time and the probabilities 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 at the current sampling time are recalculated, and the result is a good adaptability to the stochastic and deterministic characteristics of the process and of the sensor, as well as a good adaptability to any desired filter action.


REFERENCES:
patent: 3611369 (1971-10-01), Maguire
patent: 4622556 (1986-11-01), Bryant et al.
patent: 5165051 (1992-11-01), Kumar
patent: 5539412 (1996-07-01), Mendelson
patent: 5546084 (1996-08-01), Hindman
patent: 5808579 (1998-09-01), Rademacher
patent: 5900835 (1999-05-01), Stein
patent: 5943661 (1999-08-01), Katz
patent: 6011507 (2000-01-01), Curran et al.
patent: 6031881 (2000-02-01), Weill et al.
patent: 6104336 (2000-08-01), Curran et al.
patent: 2709184A1 (1995-02-01), None
patent: 566218 (1977-11-01), None
patent: 896766B (1982-01-01), None
patent: 902241B (1982-01-01), None
“A multipath data association tracker for over-the-horizon radar”, Pulford, G.W.; Evans, R.J., Aerospace and Electronic Systems, IEEE Transactions on, vol.: 34 Issue: 4, Oct. 1998, pp: 1165-1183.
“Determination of windspeed within a weather storm using airborne Doppler radar”, Deshpande, M.D.; Staton, L., Southeastcon '91., IEEE Proceedings of, 1991, pp. 508-519 vol. 1.

Affiliated with

Also associated with

Rating

Method for the probabilistic estimation of noisy measurements does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Method for the probabilistic estimation of noisy measurements, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for the probabilistic estimation of noisy measurements will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2569901