Introduction

MechaWare was designed to provide the advanced controls engineer the ability to design, implement, and load user-defined control algorithms onto a ZMP-SynqNet controller¹. The user defines and designs a new control algorithm in the user-friendly Matlab/Simulink workspace. This environment provides an easy-to-use graphical interface to define the control algorithm and also takes advantage of Matlab's data analysis capabilities.

To address more complex control system problems, particularly the issues of system resonance, vibration control and improved settling time performance, a new control topology has been introduced—MechaWare. MechaWare is made up of time proven basic building blocks used in MEI's PID and PIV algorithms. In addition, Matrix computation blocks have been added to facilitate observer-based state feedback and disturbance observers.

MechaWare allows the user to graphically define a combination of basic building blocks into a custom algorithm (or control law). The Matlab/Simulink workspace provides the user with the flexibility of a graphical user interface for algorithm definition and access to the full functionality of Matlab. The user is also able to directly load the control algorithm into the ZMP-SynqNet controller and gives the non-software engineer the ability to downoad custom control solutions onto the controller.

MechaWare can collect and process feedback from all axes on the ZMP and issue commands at the same time, which facilitates and simplifies the design of MIMO (multi in multi out) systems.

1 - Although it is possible to use MechaWare on an XMP controller, it has limited functionality due to lower processor speed and smaller memory. Therefore, no double precision math blocks will be supported by the XMP. Practical implementation on the XMP will be constrained to low axis count and simple control laws.

Back to Table of Contents | Next