Choose a Control Design Approach

Simulink^® Control Design™ lets you design and tune many types of control systems in Simulink. There are also deployable PID autotuning tools that let you tune your controller in real time against a physical plant.

Design in Simulink

Simulink Control Design provides several approaches to tuning Simulink blocks, such as Transfer Fcn and PID Controller blocks. Use the following table to determine which approach best supports what you want to do.

	Model-Based PID Tuning	Classical Control Design	Multiloop, Multiobjective Tuning
Supported Blocks	PID Controller Discrete PID Controller PID Controller (2DOF) Discrete PID Controller (2DOF)	Linear Blocks (see What Blocks Are Tunable?)	Any blocks; only some blocks are automatically parameterized (See How Tuned Simulink Blocks Are Parameterized)
Architecture	1-DOF and 2-DOF PID loops	Control systems that contain one or more SISO compensators	Any structure, including any number of SISO or MIMO feedback loops
Control Design Approach	Automatically tune PID gains to balance performance and robustness	Graphically tune poles and zeros on design plots, such as Bode, root locus, and Nichols Automatically tune compensators using response optimization (Simulink Design Optimization™), LQG synthesis, or IMC tuning	Automatically tune controller parameters to meet design requirements you specify, such as setpoint tracking, stability margins, disturbance rejection, and loop shaping (see Tuning Goals)
Analysis of Control System Performance	Time and frequency responses for reference tracking and disturbance rejection	Any combination of system responses	Any combination of system responses
Interface	Graphical tuning using PID Tuner (see Introduction to Model-Based PID Tuning in Simulink) Tuning of plants that do not linearize (see Frequency-Response Based Tuning) Programmatic tuning using pidtune (see PID Controller Design at the Command Line)	Graphical tuning using Control System Designer	Graphical tuning using Control System Tuner Programmatic tuning using slTuner (see Programmatic Tuning)

Real-Time PID Autotuning

The real-time PID autotuning tools in Simulink Control Design let you deploy an automatic tuning algorithm as a stand-alone application for PID tuning against a physical plant. Real-time PID autotuning lets you tune a PID controller to achieve a specified bandwidth and phase margin without a parametric plant model or an initial controller design.

The real-time PID autotuning algorithm can tune PID gains in Simulink PID Controller blocks or in your own custom PID blocks. You can tune against your physical plant with or without Simulink in the loop. Deploying the real-time PID autotuning algorithm requires a code-generation product such as Simulink Coder™.

For more information, see When to Use PID Autotuning.

Advanced Control Design

This table shows a summary of advanced techniques available in Simulink Control Design. This table is helpful for identifying where these techniques are applicable.

Method	How it Works	Advantages	Risks	Typical Applications	When to Use
Disturbance Compensation	Observer-based estimates of disturbance and subtracts them from control input.	Simple concept, improves tracking	Requires accurate disturbance model estimation, may amplify noise with high observer bandwidth	Motion control, robotics, electromechanical drives.	When model uncertainty is high and real-time disturbance rejection is critical.
Active Disturbance Rejection Control	Estimates and cancels total disturbances (internal uncertainty + external disturbance) using an Extended State Observer (ESO).	Robust to model uncertainty, minimal plant modeling needed, adaptive to disturbances.	Need to tune ESO gains, observer noise sensitivity, stability analysis less standard.	Motion control, robotics, electromechanical drives.	When model uncertainty is high and real-time disturbance rejection is critical.
Model Reference Adaptive Control	Adjusts controller parameters online so system output tracks a reference model.	Guarantees asymptotic tracking under assumptions, systematic Lyapunov-based design.	Requires persistency of excitation, risk of instability if adaptation is too aggressive.	Aerospace, robotics, systems with changing dynamics.	When system dynamics vary with operating conditions and a reference model is available.
Sliding Mode Control	Drives system state to a “sliding surface” and maintains motion on it using discontinuous control law.	Strong robustness to matched uncertainties and disturbances, simple design.	Chattering (high-frequency switching), may excite unmodeled dynamics.	Motor drives, aerospace, uncertain nonlinear systems.	When robustness to uncertainty is key and actuators can tolerate fast switching.
Iterative Learning Control	Learns control corrections over repeated trials of the same task by updating input each iteration.	Very high accuracy after sufficient repetitions, simple update laws.	Requires repetitive tasks, sensitive to unmodeled dynamics or noise accumulation.	Robotics, batch processes, repetitive manufacturing processes.	When the same trajectory/task repeats and performance must improve each trial.
Control Barrier Functions	Enforce safety constraints by restricting control inputs to maintain system states within a safe set.	Formal safety guarantees, compatible with other controllers (QP formulation).	May conflict with performance objectives, requires accurate constraint modeling.	Autonomous driving, robotic safety constraints.	When safety-critical operation is essential and constraints must always be respected.
Passivity-Based Control	Uses passivity/dissipative properties to design controllers ensuring stability via energy shaping.	Strong stability guarantees, modularity (interconnections preserve passivity).	May be conservative.	Power systems, teleoperation, mechanical systems.	When interconnection stability is critical (e.g., multi-agent, networked, energy systems).
Virtual Reference Feedback Tuning	Direct data-driven tuning identifies controller parameters by fitting to desired closed-loop behavior without explicit plant model.	No explicit system ID required, fast offline tuning, uses experimental data.	Requires good excitation data, performance depends on data quality, limited to chosen controller structure.	PID tuning, systems where plant modeling is hard.	When you want data-driven controller tuning without building a detailed plant model.
Extremum Seeking Control	Online, model-free optimization: perturbs input with small oscillations, measures output, adjusts control to drive system toward performance extremum (min/max).	No plant model needed, adapts to time-varying conditions, handles nonlinear unknown systems.	Need to specify learning rate, can be sensitive to noise depending on design parameters, requires persistent excitation.	Maximum power point tracking (MPPT), antilock braking.	When the objective is performance optimization in real time, especially with uncertain or time-varying plants.