ControlForge Control Toolbox Software

State of the Art Robust Control

Control

Figure 1: ControlForge functional flow. Blue boxes indicate interfaces which are provided for optional user defined functions.

ControlForge incorporates some of the most efficient and reliable modern control design approaches available:

• Sensitivity Weighted Linear Quadratic Gaussian (SWLQG)
• Multiple Model Linear Quadratic Gaussian (MMLQG)

These techniques have proven themselves on a wide range of systems, including structural, structural/acoustic, and flight vehicle systems. For example, SWLQG and MMLQG controllers achieved performance on the Middeck Active Control Experiment (MACE), an articulated multibody testbed which flew on STS-67 in 1995, which was limited only by sensor resolution. Algorithms in ControlForge were applied to a 200 state, 64 input/64 output active mirror to achieve stable wavefront shape control. ControlForge has also been applied to structural/acoustic systems. For example, acoustic attenuation of 4.88dB in the interior of a strain-actuated launch vehicle shroud mockup has been demonstrated.

Frequency Domain Design Environment

ControlForge is a frequency domain control design environment for achieving performance and robustness requirements in the presence of realistic model errors, disturbance and sensor noise levels, and real time computational constraints. Design in the frequency domain allows straightforward incorporation of steady state or transient response criteria.

Screen Shot

Figure 2: Menu entry into analysis routines (upper left), data and plant model definition window (lower left) and plots defining MIMO performance and stability for MACE MBP control.

Further, performance and stability can be accurately predicted, prior to implementation, based solely on measured data. Since the analysis is performed directly on the measured data, ControlForge offers a powerful means to assess and refine the closed loop response in the face of modeling error. As a result, designs produced in the ControlForge environment can be stably implemented and achieve the desired performance at the first hardware implementation.

High Level Graphical Interface

ControlForge presents frequency domain design information in a form which allows direct evaluation and tuning of the loops which govern closed loop performance and stability. Loops are evaluated on the true system as represented by the measured data, as well as on the design model. Graphical handles into the frequency and parameter weighting functions, which govern the loop shapes, are provided. The resulting direct control of performance and robustness reduces or eliminates design iterations. The immediate benefits are to allow the controls designer to concentrate on the design problem at hand, as the package handles details at the coding level. As a result the first control design to be implemented may be the last, as it may meet all performance requirements.

Integrated Design Environment

ControlForge is designed to provide a true design environment for a wide range of controlled systems, with attendent design models, robustness requirements, and implementation constraints. The user has the ability to modify existing design and analysis tools (implemented as Matlab functions), or to define and integrate proprietary modeling, control design and analysis code. Analysis and synthesis in the discrete and continuous domains is transparently supported. The structure of ControlForge facilitates code reuse: both of user legacy code, and also by the capability to generate a commented Matlab script file which captures the design iteration cycle in a form which can be documented, distributed, or modified.

    Top

For more information:

Please contact Chris Ludlow (Products), or call: 781-306-0609 x227