Page 105 - Handbook of Biomechatronics

Model-Based Control of Biomechatronic Systems 101

2 MODEL-BASED CONTROL DESIGN

Control systems can be categorized as either open loop or closed loop,
depending on their structure. A closed-loop control system regulates control
actions based on the information received through a feedback loop. There is
no feedback loop in an open-loop system; thus, no further control action
adjustments can be made. Both of these control systems, depending on their
design methodology, can be categorized into model-based or error-based
controllers. Model-based controllers exploit a physical or nonphysical model
to estimate system dynamics and predict the system’s response to a control
action. This category includes optimal, robust, and nonlinear control
methods. Error-based controllers use only an error signal (the difference
between the desired and actual trajectories) to control the system. Classic
proportional-integral-derivative (PID), sliding mode, and fuzzy controllers
are the most well-known controllers in this category. Since model-based
control methods can consider physiological constraints and often out-
perform their counterparts, we will focus on model-based control methods
in this section.

2.1 Model-Based Open-Loop Control
An open-loop control system is a control system in which the system output
does not influence the control actions [shown in Fig. 1A]. In an open-loop
control system, a sequence of control actions is precomputed and stored in a
feedforward controller, then executed when a trigger is activated. Once the
control action is initiated, it cannot be adjusted based on the system response
or external loads acting on the system. Feedforward controllers are usually
used when there is no feedback available or the interaction with the

Fig. 1 Schematic representation of (A) an open-loop control system and (B) a closed-
loop control system.