Signal processing apparatus and method

• US 5,146,417 A
• Filed: 03/07/1990
• Issued: 09/08/1992
• Est. Priority Date: 03/22/1989
• Status: Expired due to Fees

• Current Assignee: Gec-Ferranti Defence Systems Limited (Leonardo S.p.A.)
• Original Assignee: Gec-Ferranti Defence Systems Limited (Leonardo S.p.A.)
• Inventors: Watson, Norman F.
• Primary Examiner(s): Dixon, Joseph L.
• Application Number:
  US07/489,708
• Time in Patent Office: 916 Days
• Field of Search: 364/565, 364/566, 364/571.01, 364/571.02, 364/571.05, 341/122, 341/123, 73/517 R, 324/160, 324/166
• US Class Current: 702/85
• CPC Class Codes: G01P 7/00 Measuring speed by integrating acceleration measuring travelled distance by double integration of acceleration G01C21/16

Abstract

A method for correcting an integral signal representing the integral of an input signal derived from an accelerometer to eliminate the effects of errors in the input signal comprises sampling one or more signals which bear a functional relationship with the errors and subjecting the samples to a correction algorithmic process. The processed results are integrated and used to correct the input signal integral. Relatively few samples may be taken at below the Nyquist rate at random or pseudorandom intervals to provide an acceptable quantitive estimate of the error in the integrated input signal.

The samples are spaced throughout the integration period by defining window intervals and taking one or more samples randomly in each window interval. Apparatus to practice the method include an integrator for integrating the input signal and a sample and hold circuit for sampling the input signal. Sample signals are subjected to the correction algorithmic process under the control of a control circuit arranged to provide output control signals that define sampling intervals at a high frequency rate, sampling window intervals and the integration time. An accumulator receives and stores processed samples with the output of the accumulator applied to a correction circuit to correct errors in the input signal which after integration is also applied to the correction circuit, the output of the correction circuit representing the corrected signal.

Claims

1. Apparatus for processing an analog input signal by integrating the analog input signal with respect to time thereof for an integration period and correcting errors in the integrated signal due to errors in the analog input signal comprising control means (26) for defining the integration period and generating a plurality of control signals, the number of control signals generated being smaller than that required by the Nyquist sampling theorem for a predetermined highest frequency component of the input signal, a sampling circuit for sampling the input signal, a gate for applying the control signals to the sampling means such that said sampling circuit derives sampled signals in response to said control signals that bear a functional relationship with errors in the input signal, means for processing the sampled signals and subjecting the sampled signals to an algorithmic process representative of said functional relationship to quantify the error in the input signal represented by the sampled signals, means for integrating the processed sample results, means for applying the control signals to the means for integrating the processed sample results to quantity the effect of errors on the integrated analog input signal and means for modifying the integrated input signal in accordance with the quantified effect to correct errors in the integrated input signal.

2. Apparatus as claimed in claim 1 in which the control means is operable to define the integration period as a succession of sampling window intervals at a repetition rate greater than a Nyquist sampling rate for the lowest frequency component of interest of the sampled signal and said control signals are generated for each window interval.

3. Apparatus as claimed in claim 1 in which the control means is operable to define a plurality of high frequency intervals occurring at least the Nyquist sampling rate for the highest frequencies of interest of the signal to be sampled and randomly or pseudorandomly select ones of said high frequency intervals at which to produce said control signals.

4. Apparatus as claimed in claim 3 in which the control means is operable to define the integration period and to select a predetermined number of said high frequency intervals in the integration period.

5. Apparatus as claimed in claim 3 in which the control means is operable to cause the sampling means to produce said signal samples at the high frequency rate and apply only the samples for processing that coincide with the control signals.

6. Apparatus as claimed in claim 3 in which the control means includes a random or pseudorandom signal generator and means to condition the generator to select said high frequency intervals at which control signals are produced.

7. Apparatus as claimed in claim 3 in which arranged to sample and process the input signal for said algorithmic process and in which the sampling and processing means is responsive to the control means to take sequential samples of the integrated signal as pairs separated by one of said high frequency intervals and derive the rate of change in sample value between each sample of the pair as representing a sample of the input signal within said short time interval.

8. Apparatus as claimed in claim 1 arranged to sample and process the input signal for said algorithmic process and in which the sampling and processing means are responsive to the control means to take sequential samples of the integrated signal as pairs separated by a short time interval and derive the rate of change in sample value between each sample of the pair as representing a sample of the input signal within said short time interval.

9. Apparatus as claimed in claim 1 arranged to sample and process more than one signal for said algorithmic process and in which the control means is arranged to process samples taken simultaneously from each of the signals.

10. Apparatus as claimed in claim 1 in which the control means is operable to define the integration period as a succession of window intervals at a repetition rate greater than the Nyquist sampling rate for the lowest frequency component of interest of the sampled signal and said control signals are generated for each window interval and in which the control means is operable to define a plurality of high frequency intervals occurring at least the Nyquist sampling rate for the highest frequencies of interest of the signal to be sampled and randomly or pseudorandomly select ones of said high frequency intervals at which to produce said control signals.

11. Apparatus as claimed in claim 10 in which arranged to sample and process the input signal for said algorithmic process and in which the sampling and processing means is responsive to the control means to take samples of the integrated signal as pairs separated by a short time interval and derive the rate of change in sample value between each sample of the pair as representing a sample of the input signal within said short time interval.

12. Apparatus as set forth in claim 1 wherein said analog input signal is the output of an accelerometer.

13. Apparatus as set forth in claim 1 further including means for integrating the analog input signal and applying the integrated input signal to the means for modifying the integrated input signal.

14. Apparatus as set forth in claim 1 wherein the means for modifying the integrated analog input signal is a subtractor having a first input, a second input and an output, means connecting the integrated input signal to the first input, means connecting the integrated processed sample results to the second input, the output being a signal corresponding to the integrated input signal with errors corrected in accordance with the quantified effect derived by said means for integrating the processed sample results.