In recent years, the requirement for embedded controllers has been steadily on the increase due to the development of computers, communication and consumptive products. In order to overcome the pressure of short life-time and rapid changing rate of products, however, related industries have been engaged in improving their design to shorten the time to market and to lower the cost. Therefore, a new concept, rapid controller prototyping, has been proposed to solve the problem with the development of controllers. However, most of the proposals are expensive and restricted in the choice of software and hardware, and aren¡¦t adequate for research. To solve the problems above, this paper has chosen MATLAB/Simulink, Real-Time Workshop, a software widely used by the academic community, as tools to study rapid controller prototyping. Here, we also focus on the establishment of the developing platform. Then we design, simulate and test the controller on the platform established. Finally, we apply the controller to the physical plant. By doing so, this paper expects to solve current problems in the development of controllers.

<p>Matlab Simulink is a commonly used tool in the design process of control systems. To further take advantage of the Matlab Simulink models it is desirable to translate them for realtime use together with the possibility to read/write physical signals. </p><p>Real-Time Workshop is an extension to Simulink that automatically generates code from a model to a variety of target platforms. RTAI and RTLinux are hard realtime operating systems, making use of Linux. </p><p>To make automatically generated code run on RTAI and RTLinux an adaptation of the generation of code is necessary. </p><p>To control, for example, an automotive engine a data acquisition card with an appropriate device driver is required. Comedi, an open source project, provides a number of device drivers for data acquisition cards. </p><p>The developed system makes use of Simulink, Real-Time Workshop, RTAI or RTLinux, and the standard data acquisition card NI 6035E using a Comedi device driver. The Simulink models may be executed at frequencies up to 50 kHz on ordinary PC hardware. </p><p>The evaluation of the system consisted of measuring the interrupt latency of the used motherboard's bus, measuring computation times running Simulink models with known complexity, running models developed at Vehicular Systems and a comparison of interfacing Simulink/Real-Time Workshop between RTAI and RTLinux. </p><p>The recommended realtime operating system is RTAI due to the open source community support of it as a target platform for Real-Time Workshop.</p>

Matlab Simulink is a commonly used tool in the design process of control systems. To further take advantage of the Matlab Simulink models it is desirable to translate them for realtime use together with the possibility to read/write physical signals. Real-Time Workshop is an extension to Simulink that automatically generates code from a model to a variety of target platforms. RTAI and RTLinux are hard realtime operating systems, making use of Linux. To make automatically generated code run on RTAI and RTLinux an adaptation of the generation of code is necessary. To control, for example, an automotive engine a data acquisition card with an appropriate device driver is required. Comedi, an open source project, provides a number of device drivers for data acquisition cards. The developed system makes use of Simulink, Real-Time Workshop, RTAI or RTLinux, and the standard data acquisition card NI 6035E using a Comedi device driver. The Simulink models may be executed at frequencies up to 50 kHz on ordinary PC hardware. The evaluation of the system consisted of measuring the interrupt latency of the used motherboard's bus, measuring computation times running Simulink models with known complexity, running models developed at Vehicular Systems and a comparison of interfacing Simulink/Real-Time Workshop between RTAI and RTLinux. The recommended realtime operating system is RTAI due to the open source community support of it as a target platform for Real-Time Workshop.

Approved for public release, distribution is unlimited / This thesis is part of an ongoing research conducted at the Naval Postgraduate School to achieve the autonomous shipboard landing of Unmanned Aerial Vehicles (UAV). Two main problems are addressed in this thesis. The first is to establish communication between the UAV's ground station and the Autonomous Landing Flight Control Computer effectively. The second addresses the design and implementation of an autonomous landing controller using classical control techniques. Device drivers for the sensors and the communications protocol were developed in ANSI C. The overall system was implemented in a PC104 computer running a real-time operating system developed by The Mathworks, Inc. Computer and hardware in the loop (HIL) simulation, as well as ground test results show the feasibility of the algorithm proposed here. Flight tests are scheduled to be performed in the near future. / Lieutenant Junior Grade, Mexican Navy