Method and system for combining feedback and feedforward in model predictive control
View Patent ↗A method and system for combining a feedback control and a feedforward control in a linear MPC to minimize effect of model uncertainty. An externally computed feedforward signal, which is more accurate and reliable, can be utilized in association with the MPC, A steady state relation between system parameters can be determined in order to compute the feedforward signal for a set of actuators associated with a non-linear system. A feedback MPC controller can then be designed. A state observer can be configured as an unknown input observer to estimate the effect of the feedforward signal. A strategy for manipulating the constraints of the MPC feedback signal can be implemented. A resulting control action for the actuators can be provided as a sum of corresponding feedback and feedforward signal while ensuring the constraints satisfaction.
1. A method for integrating a feedforward control into a feedback model-based predictive controller, said method comprising:
determining a relation between parameters with respect to a non-linear system to compute a feedforward signal for at least one actuator associated with said non-linear system;
manipulating a plurality of constraints with respect to a model-based predictive feedback controller as a function of a feedforward signal, wherein the plurality of constraints include a constraint envelope that is defined about and is a function of the feedforward signal;
generating a control action as a sum of a feedback signal and said feedforward signal; and
transmitting said control action to said at least one actuator associated with said non-linear system, wherein said feedback signal in association with said feedforward signal minimize an effect of uncertainty with respect to disturbances.
2. The method of claim 1 further comprising configuring said model-based predictive feedback controller with measured signals as measured disturbances that parameterize said feedforward signal.
3. The method of claim 1 further comprising configuring a state observer as an unknown input observer to estimate an effect of said feedforward signal to be utilized in concert with said model-based predictive feedback controller.
4. The method of claim 1 wherein said feedforward signal comprises a nonlinear function.
5. The method of claim 1 further comprising utilizing a dynamic filter to improve said feedforward signal and achieve an enhanced transient response with respect to said non-linear system.
6. The method of claim 1 further comprising manipulating said plurality of constraints in order to limit an actuator signal associated with said control action so that said control action does not violate a prescribed time-varying constraint with respect to said at least one actuator.
7. The method of claim 1 further comprising estimating said feedforward signal as an unknown disturbance by an unknown input observer.
8. The method of claim 7 further comprising implementing said unknown input observer as a part of a system state observer.
9. The method of claim 7 wherein said unknown input observer comprises a Kalman filter.
10. The method of claim 1 wherein said manipulating said plurality of constraints is expressed as a function of a plurality of relevant parameters.
11. The method of claim 1 wherein said model-based predictive feedback controller comprises a plant model linearized along a trajectory defined by said feedforward signal.
12. The method of claim 1 further comprising manipulating said plurality of constraints in order to limit said feedback signal associated with said control action so that said control action does not violate a prescribed time-varying constraints with respect to said measurements thereof.
13. A method for integrating a feedforward control into a feedback model-based predictive controller, said method comprising:
determining a relation between parameters with respect to a non-linear system in order to compute a feedforward signal for at least one actuator associated with said non-linear system
manipulating a plurality of constraints with respect to a model-based predictive feedback controller as a function of the feedforward signal, said feedforward signal comprising a non-linear function, wherein the plurality of constraints include a constraint envelope that is a function of the feedforward signal;
generating a control action as a sum of a feedback signal and said feedforward signal; and transmitting said control action to said at least one actuator associated with said non-linear system, wherein said feedback signal in association with said feedforward signal are configured to reduce an effect of uncertainty with respect to a disturbances.
14. A system for integrating a feedforward control into a feedback model-based predictive controller, said system comprising:
an input configured to receive information from one or more sensors;
a controller coupled to said input, the controller configured to:
determine a relation between parameters with respect to a non-linear system,
compute a feedforward signal for at least one actuator associated with said non-linear system;
manipulate a plurality of constraints with respect to a model-based predictive feedback controller as a function of the feedforward signal, wherein the plurality of constraints include a constraint envelope that is defined about the feedforward signal;
generate a control action as a sum of a feedback signal and said feedforward signal using, at least in part, information received from the one or more sensors; and
an output coupled to the controller for outputting said control action for reception by said at least one actuator.
15. The system of claim 14 wherein said model-based predictive feedback controller is configured with measured signals as measured disturbances that parameterize said feedforward signal.
16. The system of claim 14 wherein the controller further comprises a state observer configured as an unknown input observer to estimate an effect of said feedforward signal in order to thereby implement a strategy for manipulating a plurality of constraints with respect to a model-based predictive feedback controller.
17. The system of claim 16 wherein:
said unknown input observer estimates said feedforward signal as an unknown disturbance.
18. The system of claim 14 wherein the controller further comprises a dynamic filter and a state observer, wherein said dynamic filter improves said feedforward signal to achieve an enhanced transient response with respect to said non-linear system.
19. The system of claim 14 wherein said plurality of constraints are manipulated in order to limit an actuator signal associated with said control action so that said control action does not violate a prescribed time-varying constraint with respect to said at least one actuator.
20. The system of claim 14 wherein:
manipulating said plurality of constraints is expressed as a function of a plurality of relevant parameters; and
said model-based predictive feedback controller comprises a plant model linearized along a trajectory defined by said feedforward signal.