Structured and Reduced Dimension Explicit Linear Quadratic Regulators for Systems with Constraints

摘要

It is studied how system structure can be utilized to derive reduced dimension multi-parametric quadratic programs that lead to sub-optimal explicit piecewise linear feedback solutions to the state and input constrained LQR problem. This results in a controller of lower complexity and associated computational advantages in the online implementation. At heart of the methods are state space projections using the singular value decomposition.