Multi-sensor based control of a flexible link : a graphical simulation and animation environment

In robotics applications requiring long reach and/or lightweight manipulators, such as those involving outer space or hazardous environments, the flexibility of these structures becomes a complicating factor in their use. In this thesis, a unique computer simulation environment is developed for the modeling, design, and control of flexible robotics structures. The simulation environment provides an effective tool, both in terms of cost and time, to complement flexible structure experimental platforms. A mathematical model of the flexible structure, representing its mechanical properties and the dynamic characteristics of its motion, is implemented as a set of computer algorithms. The motion of the structure is accurately simulated numerically, and the results of the simulation are used to generate a three-dimensional, real-time animation of the beam undergoing motion. Sensors and actuators commonly employed in the control of flexible structures are modeled and simulated as well, including a hub torque motor with tachometer and shaft encoder, a camera for visualization of the tip position, and piezoelectric sensors and actuators. An advanced graphical user interface allows all mechanical parameters of the structure as well as all controller parameters and sensor/actuator locations to be instantly modified by the experimenter, and the results of these changes can be quickly observed through the animation. The accuracy and validity of the modeling and implementation algorithms is established by correlation of simulation results with the behavior of physical experimentation platforms. Several examples are used to show how the simulation environment can be used to verify and optimize an overall flexible structure system, as well as predict unexpected behavior and interactions. Non-ideal effects are also modeled and included in the simulation results.