Abstract: A generic, self-tuning control system deployable in a wide variety of systems is achieved from a digital system effectuating a matrix chain. Each matrix in the chain yields an output based upon two or more inputs. The outputs of the various matrices making up the chain are used as inputs to other matrices in the chain. Measured inputs are state variables of the system under control. Some inputs that are generated by other matrices are state variables. Consequently, the set of values generated by the matrices results from the state variables describing the system at a given point in time. Initially, the matrices exhibit a quality referred to as symmetry. This quality serves to cause the system under control to transition to desired state. In addition to the controlling force exerted by virtue of the symmetry of the matrices, a learning algorithm may be introduced to the matrix chain.
Abstract: A generic, self-tuning control system deployable in a wide variety of systems is achieved from a digital system effectuating a matrix chain. Each matrix in the chain yields an output based upon two or more inputs. The outputs of the various matrices making up the chain are used as inputs to other matrices in the chain. Measured inputs are state variables of the system under control. Some inputs that are generated by other matrices are state variables. Consequently, the set of values generated by the matrices results from the state variables describing the system at a given point in time. Initially, the matrices exhibit a quality referred to as symmetry. This quality serves to cause the system under control to transition to desired state. In addition to the controlling force exerted by virtue of the symmetry of the matrices, a learning algorithm may be introduced to the matrix chain.