12.4 Controllability and observability concepts

State-space analysis dives into controllability and observability, key concepts for understanding system behavior. These ideas help engineers determine if a system can be manipulated to reach desired states and if internal states can be inferred from outputs.

Controllability and observability matrices are essential tools in this analysis. By examining these matrices, engineers can identify limitations in system control and state estimation, guiding the design of more effective control systems and state observers.

Controllability and Observability

Fundamental Concepts of System Analysis

• Controllability determines ability to manipulate system states using inputs
• Observability assesses capability to infer internal states from outputs
• Both concepts crucial for designing effective control systems
• Controllable system allows steering to desired state within finite time
• Observable system enables reconstruction of complete state from output measurements

Mathematical Foundations and Conditions

• Kalman rank condition provides mathematical criteria for controllability and observability
• Controllability rank condition: rank of controllability matrix must equal number of state variables
• Observability rank condition: rank of observability matrix must equal number of state variables
• Full rank matrices indicate complete controllability or observability
• Rank deficiency suggests limitations in system control or state estimation

System Realization and Optimization

• Minimal realization represents system with minimum number of state variables
• Achieves most efficient state-space model without losing input-output behavior
• Involves removing uncontrollable or unobservable states from system model
• Simplifies system analysis and controller design
• Balances model complexity with system performance requirements

Controllability and Observability Matrices

Controllability Matrix Construction and Analysis

• Controllability matrix formed by $[B \quad AB \quad A^2B \quad ... \quad A^{n-1}B]$
• A represents state matrix, B input matrix, n number of state variables
• Matrix dimensions: n x nm (n states, m inputs)
• Full rank controllability matrix indicates complete state controllability
• Rank deficiency reveals uncontrollable modes or states in system

Observability Matrix Formulation and Interpretation

• Observability matrix constructed as $[C^T \quad A^TC^T \quad (A^T)^2C^T \quad ... \quad (A^T)^{n-1}C^T]^T$
• C represents output matrix, A state matrix, n number of state variables
• Matrix dimensions: np x n (p outputs, n states)
• Full rank observability matrix signifies complete state observability
• Rank deficiency indicates presence of unobservable states or modes

State Feedback and Estimation

State Feedback Control Design

• State feedback involves using full state information to control system
• Feedback gain matrix K determines control law: u = -Kx
• Pole placement technique used to design K for desired closed-loop dynamics
• Ackermann's formula provides method for calculating K in single-input systems
• Linear Quadratic Regulator (LQR) optimizes feedback gains for performance criteria

State Estimation Techniques

• State estimation reconstructs full state vector from limited output measurements
• Luenberger observer design for deterministic systems
• Kalman filter optimal for stochastic systems with known noise characteristics
• Observer gain matrix L determines estimation dynamics
• Dual problem to state feedback, uses transpose of system matrices
• Separation principle allows independent design of feedback and estimation