Loop integration control system
OwnerEugene J. Pollockinventor
Life of this patent
Estimating the term…
Abstract
An In-Loop Integration Control System which is capable of producing consistent and accurate descriptions and predictions of any dynamic process, on line and in real time. This operation is performed by, first, integrating the best estimate of the highest derivative to produce an estimate of the expected next measurement and then comparing this measurement to the actual measurement; and second incorporating into the prediction any systematic error corrections. By changing the highest derivative on which the integration is performed, the total state vector describing the process is improved and all components are correctly timed.
Claims14
An In-Loop Integration Control System comprising means for receiving and storing a state vector measurement input, said state vector measurement being defined in terms of time, position, velocity and acceleration, means operatively connected to said receiving and storing means for receiving said acceleration term of said state vector measurement and for computing therefrom an estimate of a next actual measurement, means for providing systematic error terms for use within said system, means operatively connected between said systematic error term providing means and said computing means for combining the outputs therefrom thereby producing a corrected estimate of the next actual measurement, means operatively connected to said combining means for storing said corrected estimate of said next actual measurement, means for providing the next actual measurement, means operatively connected between said next actual measurement providing means and said storing means for comparing the outputs therefrom thereby producing sequential error measurements, means operatively connected to said comparing and error measurement producing means for comparing two of said sequential error measurements and producing an output therefrom, means operatively connected to said comparing means for generating an adaptive gain signal from said output, means operatively connected to said adaptive gain signal generating means for combining said adaptive gain signal with a plurality of polynomial equations thereby producing a plurality of adaptive gain multipliers, means operatively connected between said comparing and error measurement producing means and said combining and multiplier producing means for computing correction terms therefrom, means operatively connected between said correction term computing means and said input storing and receiving means for combining said correction terms with said input state vector measurement to produce an updated state vector measurement which replaces said input state vector measurement in said system and said system continually comparing said corrected estimate of said next actual measurement with said next actual measurement maintaining the difference therebetween at zero.
A method of producing consistent and accurate descriptions and predictions of a dynamic process comprising the steps of:
- 1. receiving and storing a state vector measurement input, said measurement being defined in terms of time, position, velocity and accelerations;
- 2. computing from said acceleration term of said input state vector measurement an estimate of a next actual measurement;
- 3. combining systematic error terms with said estimate of a next actual measurement and producing a corrected estimate of the next actual measurement;
- 4. comparing the corrected estimate of the next actual measurement with the next actual measurement in order to produce sequential error measurements;
- 5. generating an adaptive gain signal from said sequential error measurements;
- 6. producing a plurality of adaptive gain multipliers from said adaptive gain signal;
- 7 computing correction terms from said sequential error measurements and said adaptive gain multipliers;
- 8. combining said correction terms with said input state vector measurement to produce an updated state vector measurement;
- 9. replacing said input state vector measurement with said updated state vector measurement;
- and
- 10. repeating said method until the comparison between said corrected estimate of said next actual measurement and said next actual measurement is zero.
Dependencies are inferred from the claim text.