A test bed framework for Actuator Management Operating Software (AMOS)

Traditionally, actuator control has been performed using conventional techniques generally concerned with the steady state output of a single parameter such as speed or torque. However, such methods do not account for the nonlinear dynamics that the actuator often experiences in the real world. In robotic assemblies, motion is dictated solely by actuators. In light of this, the Robotic Research Group at the University of Texas, has spent the past twenty years working to develop an intelligent actuator. Among the chief goals of this device is its ability to overcome faults through the reallocating of resources, identify the need for maintenance, and operate at maximum output in consideration of the demands from its work environment. Research efforts have included a proposed multi-sensor environment, the development of performance maps, a methodology for performing Condition Based Maintenance (CBM), and a multi- tier structure for redundant resources that can be used to avoid fault. In order to actualize the works of RRG researchers, the Software Test Bed has been created. Through the use of a field programmable gate array and realtime controller, software programs perform intelligent control, primarily by providing feed-forward data to a conventional controller. Data fed to the controller is selected from an archive of sensor readings based upon their ability to satisfy performance demands as they pertain to maximum performance, CBM, and fault tolerance. Using decision-making algorithms, the software sets priorities among these intelligent domains as they relate to the application and the system level. The material presented here is intended to give an overview of the Software Test Bed including its inception, present setup, and potential for intelligent control. Examples of past RRG research that can be applied on the test bed has been presented along with a proposed development framework. Comparisons have also been made between the RRG research and that performed by external groups in order to strengthen the overall implementation of this project. Using the information gathered by reviewing research, decision-making algorithms have provided the ability to resolve conflict among the multiple control requirements