[attachment=15439]
Abstract:
Requirements placed on modern high-performance aircraft have resulted in inherently unstable airframe designs. Increased speed and altitude among other causes have led to larger aerodynamic forces and moments which cannot be physically controlled by the pilot alone. As a result, modern flight mixed-signal control systems (FCS) must be designed to account for these instabilities and to provide satisfactory dynamic responses while minimizing the pilot's workload. Accuracy for inertial navigation components must be improved to provide increased control system performance optimization, and coupled with GPS technology. In addition, Microelectromechanical systems (MEMS) provide anew revolution in miniaturization and component accuracy. Coupled with microelectronics, MEMS integrates both sensors and actuators on a silicon chip. MEMS should provide inexpensive, reliable, and rugged inertial navigation components.
Currently the design process for mixed-signal integrated circuit design is a manual design approach with over 70 percent of the design cycle being spent in the analog domain. Digital designs are automated through the use of the VHDL hardware description language, while the analog counterparts will be automated using the newly developed VHDL-AMS extension. Quantitative Feedback Theory (QFT) design provides a robust design methodology for mixed-signal design for command and control applications.
Introduction:
The QFT approach provides stability of the design to device and plant uncertainty, and disturbances. In this example, the VHDL-AMS language is used to describe the electronic elements of the discrete control system including the QFT digital compensator and filter. The QFT VHDL-AMS design, is tuned for performance to the bounds of uncertainty, the parameter tolerances, and the sampling time of the A/D and D/A.
Future small UAVs will require enhanced capabilities like seeing and avoiding obstacles, operating unpredicted flight conditions, interfacing with payload sensors, tracking moving Targets, and cooperating with other manned and unmanned systems. Cross-platform commonality to simplify system integration and training of personnel is also desired. A small guidance, navigation, and control system has been developed and tested. It employs Field Programmable Gate Array (FPGA) and Digital Signal Processor (DSP) technology to satisfy the requirements for more advanced vehicle behavior in a small package. Having these two processors in the system enables custom vehicle interfacing and fast sequential processing of high-level control algorithms. This paper focuses first on the design aspects of the hardware and the low-level software. Discussion of flight test experience with the system controlling both an unmanned helicopter and an 11-inch ducted fan follow.
All software and FPGA configuration data are stored in non-volatile Flash memory. When power is supplied to the board, a small boot program is loaded and executed in the DSP. The boot file initializes the DSP hardware and successively configures the FPGA from Flash, loads software for the soft core processor in the FPGA, and finally loads the main DSP application into SDRAM and/or internal memory. From then on, the soft core CPU on the FPGA initiates every flight control cycle by sending the most recent sensor data to the DSP, effectively becoming the system driver. Efficient internal and external communication is essential in a multi-processor system. The DSP communicates with the FPGA through a 32-bit parallel bus with a maximum throughput of 250 Mbps. Multiple, parallel First-In First-Out (FIFO) components inside the FPGA enable direct communication with the DSP from dedicated hardware components in the FPGA, avoiding information bottlenecks and reducing time delays. The FIFOs also act as buffers between the different components, which helps ensure data integrity. As shown in Figure 4, some information bypasses the softcore CPU and enters the DSP directly through hard-mapped UARTs in the FPGA to reduce the workload on the smaller softcore processor. The flexibility the FPGA allows for future addition of communication interface logic, image sensor interfaces, Ethernet, USB, and encryption or other protocols. The guidance, navigation, control and communication software has a platform-independent structure. All the hardware or operating system dependent software has been pulled into a separate software library, called the Standard Vehicle Interface Library (SVIL). Data communication ports that are created in SVIL handle all communication in and out of this flight controller software. Figure 5 depicts the overall software architecture of the complete flight controller system with the FCS20 at its base. This allows the flight controller code to be common across a number of different operating systems and hardware choices.