US5971579A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Unit and method for determining gains a of PID controller using a genetic algorithm
Assignee: Samsung Electronics Co Ltd
•
Filing: 1997-04-08
•
Granted: 1999-10-26
•
Expires: 2017-04-08
Abstract
Abstract A unit and a method for determining gains of a PID (Proportion, Integral, Differential) controller using a genetic algorithm. A genetic algorithm tuning unit decodes an output value from an object system extracted by a feature extracting unit, and generates a first population including a random binary string. Then, the first population is normalized in a given ratio to measure a PID gain converting fitness with respect to the first population. Thereafter, the normalized first population is compared with the decoded value to determine whether it is a user's desired value. If the desired value is obtained, the genetic algorithm tuning unit generates PID gain control signals with respect to the resulting value, which are supplied to the PID controller to tune PID gains.
Inventors
- 👤 Young-Sang Kim
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-01 12:14
CPC Classifications
G; G05; G05B; G05B11/00; G05B11/01; G05B11/36; G05B11/42; G05B13/00; G05B13/02; G05B13/0265; G05B13/04
Claims Analysis
Total Claims: 17
Claim 1
Independent Claim
114 words
A unit for determining gains of a PID (Proportion, Integral, Differential) controller using a genetic algorithm, comprising: reference setting value inputting means for inputting reference data to determine a gain of an object system; feature extracting means for determining a difference between the reference data input through the reference setting inputting means and output data of the object system to extract a feature of the object system; genetic algorithm tuning means for tuning PID gains in response to the feature output by the feature extracting means; and PID controlling means for generating gain control signals with respect to the object system relative to the tuned PID gains output from the genetic algorithm tuning means.
Claim 2
Dependent Claim
114 words
A unit for determining gains of a PID (Proportion, Integral, Differential) controller using a genetic algorithm, comprising: reference setting value inputting means for inputting reference data to determine a gain of an object system; feature extracting means for determining a difference between the reference data input through the reference setting inputting means and output data of the object system to extract a feature of the object system; genetic algorithm tuning means for tuning PID gains in response to the feature output by the feature extracting means; and PID controlling means for generating gain control signals with respect to the object system relative to the tuned PID gains output from the genetic algorithm tuning means.
Claim 3
Dependent Claim
48 words
A unit according to claim 1, wherein the feature of the object system comprises a difference between the reference data and the output of the object system, and an integral value, differential value, rising time, settling time, overshot, and damping ratio of the output of the object system.
Claim 4
Dependent Claim
94 words
A unit according to claim 1, wherein the genetic algorithm tuning means comprises: random number generating means for generating a first population to be compared with the feature, to control the PID gains, said random number generating means comprising an amplifier for amplifying an input noise, a comparator for comparing the amplified noise of the amplifier with a reference signal to produce an output signal relative to an amplitude of the input noise, and signal converting means for converting the output signal from the comparator into a binary string, to generate the first population.
Claim 5
Dependent Claim
59 words
A unit according to claim 1, wherein the feature extracting unit stores an integral value, differential value, rising time, settling time, overshot, and damping ratio of the object system, and selects the integral value, differential value, rising time, settling time, overshot, and damping ratio based upon the difference between the reference data and the output of the object system.
Claim 6
Independent Claim
99 words
A method for determining gains of a PID (Proportion, Integral, Differential) controller using a genetic algorithm, comprising the steps of: decoding a feature of an output value from an object system; generating a first population including a random binary string; normalizing the first population in a given ratio to measure PID gain converting fitness with respect to the first population; and comparing the normalized first population with the decoded output value, determining whether the normalized first population is a desired value, and generating PID gain control signals with respect to the resulting value if the desired value is obtained.
Claim 7
Independent Claim
99 words
A method for determining gains of a PID (Proportion, Integral, Differential) controller using a genetic algorithm, comprising the steps of: decoding a feature of an output value from an object system; generating a first population including a random binary string; normalizing the first population in a given ratio to measure PID gain converting fitness with respect to the first population; and comparing the normalized first population with the decoded output value, determining whether the normalized first population is a desired value, and generating PID gain control signals with respect to the resulting value if the desired value is obtained.
Claim 8
Dependent Claim
112 words
A method according to claim 4, wherein said step of normalizing a first population comprises the second population steps of: determining the first population using a fitness function, and extracting a plurality of populations having values more than a predetermined value; determining the second population using the fitness function until the determined second population value is obtained as a fixed value, and generating a third population having the fixed value; crossing over each other adjacent populations among the third population by bits of a predetermined length and generating a fourth population; and mutating respective bits of the fourth population to generate a fifth population, the fifth population being the normalized first population.
Claim 9
Independent Claim
73 words
A method of generating gain control signals representative of gains of a PID (Proportion, Integral, Differential) controller, comprising the steps of: determining a difference between reference data for determining a gain of an object system and output data of the object, to extract a feature of the object system; tuning PID gains in response to the feature; and generating the gain control signals for the object system based upon the tuned PID gains.
Claim 10
Independent Claim
73 words
A method of generating gain control signals representative of gains of a PID (Proportion, Integral, Differential) controller, comprising the steps of: determining a difference between reference data for determining a gain of an object system and output data of the object, to extract a feature of the object system; tuning PID gains in response to the feature; and generating the gain control signals for the object system based upon the tuned PID gains.
Claim 11
Dependent Claim
65 words
A method according to claim 7, wherein the step of determining a difference comprises the steps of: pre-storing an integral value, differential value, rising time, settling time, overshot, and damping ratio of the output value; and selecting the integral value, differential value, rising time, settling time, overshot, and damping ratio based upon the difference between the reference data and the output of the object system.
Claim 12
Dependent Claim
100 words
A method according to claim 8, wherein the step of tuning PID gains comprises the steps of: decoding the feature of the object system; generating a first population including a random binary string based upon the decoded output; normalizing the first population in a given ratio to measure PID gain converting fitness with respect to the first population; and comparing the normalized first population with the decoded output, determining whether the normalized first population is a desired value, and generating the gain control signals with respect to the normalized first population if the normalized first population is the desired value.
Claim 13
Dependent Claim
102 words
A method according to claim 9, wherein the step of normalizing a first population comprises the steps of: randomly determining a first population using a fitness function to extract values more than a given value; determining the second population using the fitness function until the determined second population value is a fixed value; generating a third population having the fixed value; crossing over each other adjacent populations among the third population by a predetermined bit length to generate a fourth population; and mutating respective bits of the fourth population to generate a fifth population, the fifth population being the normalized first population.
Claim 14
Dependent Claim
102 words
A method according to claim 8, wherein the step of normalizing a first population comprises the steps of: randomly determining the first population using a fitness function to extract values more than a given value; determining a second population using the fitness function until the determined second population value is a fixed value; generating a third population having the fixed value; crossing over each other adjacent populations among the third population by a predetermined bit length to generate a fourth population; and mutating respective bits of the fourth population to generate a fifth population, the fifth population being the normalized first population.
Claim 15
Dependent Claim
96 words
A method according to claim 7, wherein the step of tuning PID gains comprises the steps of: decoding the feature; generating a first population including a random binary string based upon the decoded output; normalizing the first population in a given ratio to measure PID gain converting fitness with respect to the first population; and comparing the normalized first population with the decoded output, determining whether the normalized first population is a desired value, and generating the gain control signals with respect to the normalized first population if the normalized first population is the desired value.
Claim 16
Dependent Claim
102 words
A method according to claim 12, wherein the step of normalizing a first population comprises the steps of: randomly determining a first population using a fitness function to extract values more than a given value; determining the second population using the fitness function until the determined second population value is a fixed value; generating a third population having the fixed value; crossing over each other adjacent populations among the third population by a predetermined bit length to generate a fourth population; and mutating respective bits of the fourth population to generate a fifth population, the fifth population being the normalized first population.
Claim 17
Dependent Claim
102 words
A method according to claim 7, wherein the step of normalizing a first population comprises the steps of: randomly determining the first population using a fitness function to extract values more than a given value; determining a second population using the fitness function until the determined second population value is a fixed value; generating a third population having the fixed value; crossing over each other adjacent populations among the third population by a predetermined bit length to generate a fourth population; and mutating respective bits of the fourth population to generate a fifth population, the fifth population being the normalized first population.
References Cited (Backward Citations)
US4669040A
Cited
US4864490A
Cited
US5245528A
Cited
US5311421A
Cited
US5295061A
Cited
US5649062A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2002-10-25 | 2003-04-26 | 2003-10-26 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2006-10-25 | 2007-04-26 | 2007-10-26 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2010-10-25 | 2011-04-26 | 2011-10-26 | $7700.00 | Window Closed / Lapsed |