Estimation of the parameters of sampled-data systems by stochastic approximation
Parameter estimation of sampled data control systems by stochastic approximation
SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Parameter estimation of sampled data control systems by stochastic approximation
Upgrade of data acquisition and controls systems software at Johnson Space Center's Space Environment Simulation Laboratory (SESL) involved the definition, evaluation and selection of a system communication architecture and software components. A brief discussion of the background of the SESL and its data acquisition and controls systems provides a context for discussion of the requirements for each selection. Further framework is provided as upgrades to these systems accomplished in the 1990s and in 2003 are compared to demonstrate the role that technological advances have had in their improvement. Both of the selections were similar in their three phases; 1) definition of requirements, 2) identification of candidate products and their evaluation and testing and 3) selection by comparison of requirement fulfillment. The candidates for the communication architecture selection embraced several different methodologies which are explained and contrasted. Requirements for this selection are presented and the selection process is described. Several candidates for the software component of the data acquisition and controls system are identified, requirements for evaluation and selection are presented, and the evaluation process is described.
The determination of an error criterion which will give a sampling rate for adequate performance of linear, time-invariant closed-loop, discrete-data control systems was studied. The proper modelling of the closed-loop control system for characterization of the error behavior, and the determination of an absolute error definition for performance of the two commonly used holding devices are discussed. The definition of an adequate relative error criterion as a function of the sampling rate and the parameters characterizing the system is established along with the determination of sampling rates. The validity of the expressions for the sampling interval was confirmed by computer simulations. Their application solves the problem of making a first choice in the selection of sampling rates.
Understanding and diagnosing the state of a particle accelerator requires navigating high-dimensional control system data, often involving hundreds of interdependent parameters. We propose a novel multimodal embedding framework that jointly learns representations of machine states from both numerical control system readouts and natural language descriptions. This enables the translation of complex machine conditions into human-readable summaries while maintaining fidelity to the underlying physical system. The obtained embeddings are subsequently adapted to an open-weights large language model via cross-attention conditioning. We demonstrate a first implementation trained on European XFEL machine state data. This work covers the embedding model architecture, training methodology, and presents initial examples demonstrating the model's capabilities in action. Due to the general concept of machine state, the model can be easily adapted to other facilities and control system environments.
Quality control monitoring system for satellite telemetry data information systems
Special purpose computer for implementing programmable digital filter used in sampled data control systems
Discrete filter for determining autocorrelation functions of internal disturbances in sampled data control system during operation
The thermal control surfaces experiment (TCSE) was the most complex hardware system aboard the Long Duration Exposure Facility (LDEF). The TCSE system consists of a scanning spectroreflectometer that measured test samples mounted on a rotatable carousel assembly. A microprocessor based data system controlled all aspects of TCSE system operation. Power was provided by four primary batteries. Flight measurement and housekeeping data were stored on a tape recorder for postflight analysis. The TCSE is a microcosm of complex electro-optical payloads being developed by NASA, DoD, and the aerospace community. The TCSE provides valuable data on the performance of these systems in space. The TCSE flight system and its excellent performance on the LDEF mission are described. A few operational anomalies were encountered and are discussed. Initial post-flight tests show that the TCSE system remains functional although some degradation in the optical measurements were observed. The results of these tests are also presented.
The following work details a study into real-time failure adaptive control allocation method for powered descent vehicle systems. The motivation for this work is to enable future human and robotic missions utilizing a powered descent system to tolerate engine failures in flight without the loss of crew or assets. This study is conducted using a six degree-of-freedom trajectory simulation of a PDV (Powered Descent Vehicle) experiencing either a loss of thrust or an engine stuck full on failure scenario. Sequential least squares in the frequency domain is used on-board to process inertial measurement unit (IMU) data and generate an estimate of the PDV plant model, which is then fed to the guidance and control system. Data used by the sequential least squares method is generated from an in-flight maneuver. The work herein focuses on determining a maneuver that is least impactful to the PDV trajectory and enables a suitable plant model estimate. A 1.5-second-long maneuver with an amplitude of 5 percent throttle is determined to provide suitable data for the sequential least squares method to estimate a plant model. A PDV implementing this method can adapt to a single engine failure and continue to reach its touchdown conditions.
A new approach to digital control system design is applied to the analysis and design of a practical onboard digital attitude control system for a class of spinning vehicles characterized by a rigid body and two connected flexible appendages. The approach used is to design a continuous-data control system that will provide a satisfactory system response. Then, using the digital redesign method, a digital controller with onboard digital computer is designed to provide a digital control system whose states are similar to those of the continuous system at sampling instants. The simplicity of application of this approach is indicated by an example in which spinning Skylab parameters are used to substantiate the conclusions.
An approach to digital control system design is applied to the analysis and design of a practical onboard digital attitude control system for a class of spinning vehicles characterized by a rigid body and two connected flexible appendages. The approach used is to design a continuous-data control system that will provide a satisfactory system response. Then, using the digital redesign method, a digital controller with onboard digital computer is designed to provide a digital control system whose states are similar to those of the continuous system at sampling instants. The simplicity of application of this approach is indicated by example. The example, using spinning Skylab parameters, is used to substantiate the conclusions.
A self-managing system that uses autonomy and autonomicity is provided with the self-* property of autopoiesis (self-creation). In the event of an agent in the system self-destructing, autopoiesis auto-generates a replacement. A self-esteem reward scheme is also provided and can be used for autonomic agents, based on their performance and trust. Art agent with greater self-esteem may clone at a greater rate compared to the rate of an agent with lower self-esteem. A self-managing system is provided for a high volume of distributed autonomic/self-managing mobile agents, and autonomic adhesion is used to attract similar agents together or to repel dissimilar agents from an event horizon. An apoptotic system is also provided that accords an "expiry date" to data and digital objects, for example, that are available on the internet, which finds usefulness not only in general but also for controlling the loaning and use of space scientific data.
The Electronic Regulator Project at TRW has constructed a breadboard system demonstrating pressure regulator operation over a wide range of flow conditions. The test hardware is composed of pneumatic-actuated, solenoid-controlled valves and associated tubing, gas volumes and transducers. System control, data recording and data reduction functions are controlled by a 486-class PC running Labview software. This breadboard system has demonstrated that a single system, using several control modes, is versatile enough to be used across the TRW spacecraft product line, from the very low Xenon flows associated with an electric propulsion system, to the high Helium flows required in a constant thrust situation such as in the AXAF spacecraft. A conceptual flight version, composed largely of off-the-shelf parts, has been designed and evaluated relative to requirements of various TRW spacecraft.
Research on flight control systems is reviewed. Investigations were conducted to evaluate and test a recent method of calculation of the performance index, application to various example designs, and programming for the discrete data control system case.
The development and evaluation of a model of the Earth Radiation Budget Satellite (ERBS) attitude dynamics for use in attitude determination is described. Attitude rates for the three axis stabilized spacecraft are derived analytically from Euler's equation using modeled disturbance torques and telemetered flight data from control system actuators for computation of the control torques. The modeled rates are used for attitude propagation in a batch least squares attitude estimation algorithm originally designed to use onboard gyro measurements of attitude motion. The capabilities for near fine attitude determination without gyros is important for operational attitude ground support of the aging ERBS spacecraft as two out of three gyro axes have failed. The accuracy of this method of dynamic propagation is strongly dependent on the quality and availability of control system telemetry data. Detailed disturbance torque models are also necessary to accurately model attitude rates. Even with the relatively coarse ERBS control system telemetry, accuracy to within 0.2 deg of gyro propagated reference attitudes is achieved.
The Cal Poly Space Project is an effort on the part of several highly motivated students to deploy a space canister which will examine the effects of microgravity on electroplating and immiscible metals. The experiments will be controlled and monitored by a specialized triple redundancy system developed to defer the possible electronic errors due to uncontrollable factors such as photons from the Sun. With the finalization of the payload design and the near completion of the data control system, the integration phase of the project is anticipated to be completed and the project ready for launching by early 1987. It is hoped that the experiments will lead to new insights in space research and also prove profitable to industry.
Propagating attitude by integrating Euler's equation for rigid body motion has long been suggested for the Earth Radiation Budget Satellite (ERBS) but until now has not been implemented. Because of limited Sun visibility, propagation is necessary for yaw determination. With the deterioration of the gyros, dynamic propagation has become more attractive. Angular rates are derived from integrating Euler's equation with a stepsize of 1 second, using torques computed from telemetered control system data. The environmental torque model was quite basic. It included gravity gradient and unshadowed aerodynamic torques. Knowledge of control torques is critical to the accuracy of dynamic modeling. Due to their coarseness and sparsity, control actuator telemetry were smoothed before integration. The dynamic model was incorporated into existing ERBS attitude determination software. Modeled rates were then used for attitude propagation in the standard ERBS fine-attitude algorithm. In spite of the simplicity of the approach, the dynamically propagated attitude matched the attitude propagated with good gyros well for roll and yaw but diverged up to 3 degrees for pitch because of the very low resolution in pitch momentum wheel telemetry. When control anomalies significantly perturb the nominal attitude, the effect of telemetry granularity is reduced and the dynamically propagated attitudes are accurate on all three axes.
General theory for X transform and its use in sample data systems analysis