Utah State University
Pointing and jitter control of space-based optical sensors requires an accurate, high-fidelity model of the system dynamics. The Space Dynamics Laboratory (SDL) has created a rapid-prototyping methodology for these control systems which is based on a family of industry-standard software tools. It can be quickly iterated due to its software-driven nature. A key component of this approach is the Adams software from MSC. Features that have been developed specifically for optical pointing systems include: integrated end-to-end modeling, co-simulation, and 3-D visualization of structural and system dynamics, controls, and optics; automatic C-code synthesis from block diagrams; real-time hardware-in-the-loop (HIL) testing; and a customizable GUI to monitor testing and change control parameters "on the fly."
First, the system to be controlled is modeled using 3-D CAD software. Finite element analysis is then used to generate the modal neutral files for the flexible-body components of the system. The CAD model and modal neutral files are then transferred to Adams, where rigid and flexible models of the components and joints are used in a graphical 3-D simulation and visualization of the multi-body dynamics. Adams is used to generate either a non-linear model of the plant or a linearized state-space model for the control system software. The Adams plant model is embedded as a block inside of the block-diagram-based simulation of the complete closed-loop control system, the feedback sensors are modeled in block-diagram form, and the control algorithms are designed using classical and modern control theory. The uncontrolled and controlled responses of the system are simulated and graphically animated.
Automatic code-generation software converts the block diagram to optimized C code and then compiles, links, downloads, and runs the model in real time on a target PC that contains a real-time operating system (RTOS) and off-the-shelf I/O interface cards. The simulation blocks representing subsystem components such as sensors and actuators are replaced with real hardware. A customizable graphical user interface (GUI) on the host PC is used to monitor outputs and change controller parameters "on the fly" with animated sliders, buttons, switches, and gauges. Iterations to the controller design are made through the GUI. Connections between the real hardware and the simulation program are formed and edited graphically. Rapid iterations to the controller design are made until an acceptable performance is achieved. If changes need to be made to the original block diagram, only seconds are required to edit the diagram and generate new real-time code. Through this process, the block diagram model of the complete system is transformed from a virtual model of the system to an integrated hardware assembly. The complete code is tested and verified on the full system. Rapid iterations through the entire sequence of tasks can still be easily made at this point.
SDL has also developed a powerful customization of the Adams package that is unique in the industry, enabling the "marriage" of optical ray tracing analysis with multi-body dynamic simulation. The algorithm computes and displays the paths and intersection points of reflected and refracted optical rays as the optical surfaces move dynamically. Motions of the surfaces can be arbitrary in all 6 degrees of freedom. In quasi-real time, the user can watch the ray trace move and the resultant image quality metric change due to motion of the optical elements. This approach enables the user to more quickly understand and visualize the situation, and also reduces the chances of error that arise when two codes have to be used (a static ray trace code and a dynamic simulation code) to analyze the system.
Successful applications of this approach by our team members for trade studies, simulation, modeling, and control system design include:
- Missile steering systems (PK, SICBM, AMRAAM, Shuttle)
- Mechanisms for the Russian American Observational Satellites (RAMOS) and Space Based Infrared System (SBIRS)-Low
- Six-Degree of Freedom (DOF) trajectory analysis and optimization
- Attitude control system for Skipper and USUSat
- Inflation control and pointing system for solar thermal propulsion
- Michelson interferometer control
- Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) scan mirror control
- SOFIE pointing control system
SOFIE: SOFIE (Solar Occultation for Ice Experiment) is a 16-channel radiometer that was launched into a polar orbit on the NASA Aeronomy of Ice in the Mesosphere (AIM) spacecraft in April 2007, and continues to provide high quality data. Figures 1 and 2 show the SOFIE hardware. SOFIE stares at the sun during sunrise and sunset and profiles 5 gaseous species and polar mesospheric cloud extinction. The original design of SOFIE featured a pointing control system (PCS). The purpose of the PCS was to lock onto and precisely track the top edge of the sun in the presence of spacecraft disturbance motions, and to periodically scan the solar disk for calibration. The main components are a 2-axis fast steering mirror (FSM) and a 1-million pixel sun sensor. Coarse and fine tracking algorithms report the positions of the solar centroid and edges to the controller. The control law uses this information to command the FSM. SDL developed the control laws for the inner FSM feedback loop and the outer PCS feedback loop. The high-fidelity multi-body dynamic model of SOFIE using the Adams software includes flexible- and rigid-body representations of all components, as well as dynamic ray tracing. Figure 3 shows the Adams model, and Figure 4 shows the control system block diagram simulation with the Adams model embedded. A late change in launch loads necessitated the development of a soft-mount vibration isolation system (VIS) for the FSM, which presented an additional challenge to the performance and stability of the system. One design iteration resulted in a 750-Hz instability, and analysis showed the source to be the flexibility of the VIS bottom plate. The subsequent plate redesign resulted in good stability margins. Extensive testing has shown stable pointing performance of about 2 arcsec, compared to the requirement of 15 arcsec. SDL compared the simulation to actual test data obtained from tracking the sun. The error between the simulation and the data is about 1%. It is interesting to note that the control algorithm and gains that were finally implemented were exactly the same as those developed and presented at the SOFIE critical design review about a year earlier. Due to an overtest condition during pre-launch testing, the FSM was damaged and ceased to function properly. Since analyses of the spacecraft pointing capability predicted excellent performance and stability, the suggestion was made to replace the FSM with a fixed mirror and use only the spacecraft to do the instrument pointing. This approach would give up some science goals but could still meet the majority of the requirements. Therefore, an in-depth analysis of the jitter was needed to verify this new direction. SDL also performed this analysis, since their methodology combines structural dynamics with dynamic ray tracing to determine the motion of the boresight on the focal plane array (FPA) in the presence of disturbances, such as spacecraft momentum wheel imbalance. The analytical prediction of 9.6 arcsec compared favorably to the actual on-orbit jitter data of 6.8 arcsec.
|Steven R. Wassom, Ph.D., P.E.|
Figure 1. SOFIE Payload
Figure 2. Schematic of SOFIE
Figure 3. Adams Model of SOFIE Pointing Control System
Figure 4. SOFIE Pointing Control Block Diagram