Iterative matrix processor based implementation of real-time model predictive control
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G05B-013/02
G06F-019/00
G06F-007/38
G06F-007/52
출원번호
UP-0876440
(2007-10-22)
등록번호
US-7844352
(2011-01-31)
발명자
/ 주소
Vouzis, Panagiotis
Bleris, Leonidas
Arnold, Mark G.
Kothare, Mayuresh V.
출원인 / 주소
Lehigh University
대리인 / 주소
McNees Wallace & Nurick, LLC
인용정보
피인용 횟수 :
33인용 특허 :
17
초록▼
A system for embedding real-time Model Predictive Control (MPC) in a System-on-a-Chip (SoC) devices is provided. In the system, a microprocessor is connected to an auxiliary unit or application-specific matrix coprocessor. The microprocessor can control the operation of the MPC algorithm, i.e., carr
A system for embedding real-time Model Predictive Control (MPC) in a System-on-a-Chip (SoC) devices is provided. In the system, a microprocessor is connected to an auxiliary unit or application-specific matrix coprocessor. The microprocessor can control the operation of the MPC algorithm, i.e., carry out the tasks of input/output for the MPC algorithm, initialize and send the appropriate commands to auxiliary unit and receive back the optimal control moves or instructions from auxiliary unit. The auxiliary unit can operate as a matrix coprocessor by executing matrix operations, e.g. addition, multiplication, inversion, etc., required by the MPC algorithm.
대표청구항▼
What is claimed is: 1. A system comprising: a plant controlled by real-time model predictive control; an iterative matrix processor configured to perform computations on vectors and matrices stored in a logarithmic format; a general purpose processor in communication with the plant and the iterativ
What is claimed is: 1. A system comprising: a plant controlled by real-time model predictive control; an iterative matrix processor configured to perform computations on vectors and matrices stored in a logarithmic format; a general purpose processor in communication with the plant and the iterative matrix processor to transfer information between the plant and the iterative matrix processor, the general purpose processor being configured to issue a sequence of instructions to the iterative matrix processor to implement the model predictive control; and wherein each instruction of the sequence of instructions initiates a series of element-by-element operations in the logarithmic format on scalars, vectors and matrices stored in the iterative matrix processor. 2. The system of claim 1 wherein: the iterative matrix processor is configured to execute matrix by vector multiplication, scalar by vector multiplication, matrix-element zeroing, matrix-element negation, matrix-element reciprocal, vector-reciprocal square, vector-reciprocal cube, vector sum, vector multiply-accumulate, or matrix multiply-accumulate; and the matrix-element reciprocal operation results in a negation in the logarithmic format, the vector-reciprocal square operation results in a shift and negation in the logarithmic format, the vector-reciprocal cube operation results in a shift, add and negation in the logarithmic format. 3. The system of claim 2 wherein the iterative matrix processor further comprises a pipelined logarithmic processor, the pipelined logarithmic processor is configured to execute matrix by vector multiplication, scalar by vector multiplication, vector sum, vector multiply-accumulate, matrix multiply-accumulate, matrix-element reciprocal, vector-reciprocal square, and vector-reciprocal cube. 4. The system of claim 3 wherein the iterative matrix processor delays computation with a later portion of a matrix while computing with an earlier portion of the matrix in response to a size of the matrix being small compared to a depth of the pipelined logarithmic processor. 5. The system of claim 4 wherein the computation delay is a predetermined number of clock cycles related to the depth of the pipelined logarithmic processor. 6. A controller to implement a real-time model predictive control algorithm, the controller comprising: a microprocessor to execute a portion of the model predictive control algorithm; and an iterative matrix processor configured to receive an instruction from the microprocessor to perform at least one matrix calculation of the model predictive control algorithm, and the iterative matrix processor being configured to operate using a logarithmic number system. 7. The controller of claim 6 wherein the iterative matrix processor is configured to execute a single multiply-accumulate operation at each clock cycle. 8. The controller of claim 6 wherein the iterative matrix processor is configured to execute multiple multiply-accumulate operations at each clock cycle. 9. The controller of claim 6 wherein the iterative matrix processor is configured to execute one-bit of a single multiply-accumulate operation at each clock cycle. 10. The controller of claim 6 wherein the iterative matrix processor is configured to execute at least one of matrix by vector multiplication, scalar by vector multiplication, matrix-element zeroing, matrix-element negation, matrix-element reciprocal, vector-reciprocal square, vector-reciprocal cube, vector sum, vector multiply-accumulate, or matrix multiply-accumulate. 11. A model predictive controller comprising: a general purpose processor to execute a model predictive control algorithm; an iterative matrix processor in communication with the general purpose processor, the iterative matrix processor being configured to execute at least one matrix calculation required by the model predictive control algorithm in an iterative process in response to receiving an instruction from the general purpose processor; and wherein the iterative matrix processor performs the at least one matrix calculation using a logarithmic number system. 12. The model predictive controller of claim 11 wherein the general purpose processor comprises a microprocessor, a memory device, at least one input/output connection and at least one bus interconnecting the microprocessor, the memory device and the at least one input/output connection. 13. The model predictive controller of claim 11 wherein the iterative matrix processor comprises a control unit, a plurality of data memory units and a logarithmic number unit. 14. The model predictive controller of claim 13 wherein the control unit is a finite state machine. 15. The model predictive controller of claim 13 wherein the logarithmic number unit is a pipelined arithmetic logic unit. 16. The model predictive controller of claim 11 wherein the general purpose processor and the iterative matrix processor are configured to operate using a word length of up to 32 bits. 17. The model predictive controller of claim 11 wherein the general purpose processor provides the iterative matrix processor with required matrices for optimizing the model predictive control algorithm at a start of the model predictive control algorithm, and the iterative matrix processor stores the required matrices for later use in optimizing the model predictive control algorithm. 18. The model predictive controller of claim 17 wherein the general purpose processor provides the iterative matrix processor with a sequence of commands to optimize the model predictive algorithm and the iterative matrix processor provides the general purpose processor with a control command.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (17)
Keeler, James D.; Hartman, Eric J.; Godbole, Devendra B.; Piche, Steve; Arbila, Laura; Ellinger, Joshua; Ferguson, II, R. Bruce; Krauskop, John; Kempf, Jill L.; O'Hara, Steven A.; Strauss, Audrey; Te, Automated method for building a model.
Ishii Junichi (Katsuta JPX) Ohnishi Hiroshi (Hitachi JPX) Kurihara Nobuo (Hitachiota JPX), Control system for an automotive vehicle having apparatus for predicting the driving environment of the vehicle.
Rizzoni Giorgio (Upper Arlington OH), Methods and apparatus for performing combustion analysis in an internal combustion engine utilizing ignition voltage ana.
Keith,Steven; Parker,Robert S.; Harvey,Noel G.; Pettis,Ronald J.; DeNuzzio,John D.; Vonk,Glenn, System and method for initiating and maintaining continuous, long-term control of a concentration of a substance in a patient using a feedback or model-based controller coupled to a single-needle or .
Keeler James David (Austin TX) Havener John Paul (Austin TX) Godbole Devendra (Austin TX) Ferguson ; II Ralph Bruce (Austin TX), Virtual emissions monitor for automobile and associated control system.
Drees, Kirk H.; Wenzel, Michael J.; Turney, Robert D., Building control systems with optimization of equipment life cycle economic value while participating in IBDR and PBDR programs.
Drees, Kirk H.; Wenzel, Michael J.; Turney, Robert D., Building management system with electrical energy storage optimization based on statistical estimates of IBDR event probabilities.
Wenzel, Michael J.; Lenhardt, Brett M.; Drees, Kirk H., Electrical energy storage system with battery power setpoint optimization based on battery degradation costs and expected frequency response revenue.
Wenzel, Michael J.; Drees, Kirk H., Electrical energy storage system with battery power setpoint optimization using predicted values of a frequency regulation signal.
Wenzel, Michael J.; Drees, Kirk H.; ElBsat, Mohammad N., Electrical energy storage system with variable state-of-charge frequency response optimization.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.