VTechWorks Repository :: Browsing by Author "Stilwell, Daniel J."

Browsing by Author "Stilwell, Daniel J."

Now showing 1 - 20 of 165

Adaptation of Delayed Position Feedback to the Reduction of Sway of Container Cranes
Nayfeh, Nader Ali (Virginia Tech, 2002-12-04)
Cranes are increasingly used in transportation and construction. increasing demand and faster requirements necessitate better and more efficient controllers to guarantee fast turn-around time and to meet safety requirements. Container cranes are used extensively in ship-to-port and port-to-ship transfer operations. In this work, we will extend the recently developed delayed position feedback controller to container cranes. In contrast with traditional work, which models a crane as a simple pendulum consisting of a hoisting cable and a lumped mass at its end, we have modeled the crane as a four-bar mechanism. The actual configuration of the hoisting mechanism is significantly different from a simple pendulum. It consists typically of a set of four hoisting cables attached to four different points on the trolley and to four points on a spreader bar. The spreader bar is used to lift the containers. Therefore, the dynamics of hoisting assemblies of large container cranes are different from that of a simple pendulum. We found that a controller which treats the system as a four-bar mechanism has an improved response. We developed a controller to meet the following requirements: traverse an 80-ton payload 50 m in 21.5 s, including raising the payload 15 m at the beginning and lowering the payload 15 m at the end of motion, while reducing the sway to 50 mm within 5.0 s at the end of the transfer maneuver. The performance of the controller has been demonstrated theoretically using numerical simulation. Moreover, the performance of the controller has been demonstrated experimentally using a 1/10th scale model. For the 1/10th scale model, the requirements translate into: traverse an 80 kg payload 5 m in 6.8 s, including raising 1.5 m at the beginning and lowering 1.5 m at the end of motion, while reducing the sway to 5 mm in under 1.6 s. The experiments validated the controller.
Adaptive Control of Nonaffine Systems with Applications to Flight Control
Young, Amanda (Virginia Tech, 2006-05-05)
Traditional flight control design is based on linearization of the equations of motion around a set of trim points and scheduling gains of linear (optimal) controllers around each of these points to meet performance specifications. For high angle of attack maneuvers and other aggressive flight regimes (required for fighter aircraft for example), the dynamic nonlinearities are dependent not only on the states of the system, but also on the control inputs. Hence, the conventional linearization-based logic cannot be straightforwardly extended to these flight regimes, and non-conventional approaches are required to extend the flight envelope beyond the one achievable by gain-scheduled controllers. Due to the nonlinear-in-control nature of the dynamical system in aggressive flight maneuvers, well-known dynamic inversion methods cannot be applied to determine the explicit form of the control law. Additionally, the aerodynamic uncertainties, typical for such regimes, are poorly modelled, and therefore there is a great need for adaptive control methods to compensate for dynamic instabilities. In this thesis, we present an adaptive control design method for both short-period and lateral/directional control of a fighter aircraft. The approach uses a specialized set of radial basis function approximators and Lyapunov-based adaptive laws to estimate the unknown nonlinearities. The adaptive controller is defined as a solution of fast dynamics, which verifies the assumptions of Tikhonov's theorem from singular perturbations theory. Simulations illustrate the theoretical findings.
Advanced System Monitoring with Phasor Measurements
Zhou, Ming (Virginia Tech, 2008-05-19)
Phasor Measurement Units (PMUs) are widely acknowledged as one of the most promising developments in the field of real-time monitoring of power systems. By aligning the time stamps of voltage and current phasor measurements that are consistent with Coordinated Universal Time (UTC), a coherent picture of the power system state can be achieved through either direct measurements or simple linear calculations. With the growing number of PMUs planned for installation in the near future, both utilities and research institutions are looking for the best solutions to the placement of units as well as to the applications that make the most of phasor measurements. This dissertation explores a method for optimal PMU placement as well as two applications of synchronized phasor measurements in state estimation. The pre-processing PMU placement method prepares the system data for placement optimization and reduces the size of the optimization problem. It is adaptive to most of the optimal placement methods and can save a large amount of computational effort. Depth of un-observability is one of the criteria to allow the most benefit out of a staged placement of the units. PMUs installed in the system provide synchronized phasor measurements that are highly beneficial to power system state estimations. Two related applications are proposed in the dissertation. First, a post-processing inclusion of phasor measurements in state estimators is introduced. This method avoids the revision of the existing estimators and is able to realize similar results as mixing phasor data with traditional SCADA with a linear afterwards step. The second application is a method to calibrate instrument transformers remotely using phasor measurements. Several scans of phasor measurements are used to accomplish estimating system states in conjunction with complex instrument transformer correction factors. Numerical simulation results are provided for evaluation of the calibration performance with respect to the number of scans and load conditions. Conducting theoretical and numerical analysis, the methods and algorithms developed in this dissertation are aimed to strategically place PMUs and to incorporate phasor measurements into state estimators effectively and extensively for better system state monitoring. Simulation results show that the proposed placement method facilitates approaching the exact optimal placement while keep the computational effort low. Simulation also shows that the use of phasor measurement with the proposed instrument transformer correction factors and proposed state estimation enhancement largely improves the quality of state estimations.
Advances in the Use of Finite-Set Statistics for Multitarget Tracking
Jimenez, Jorge Gabriel (Virginia Tech, 2021-10-27)
In this dissertation, we seek to improve and advance the use of the finite-set statistics (FISST) approach to multitarget tracking. We consider a subsea multitarget tracking application that poses several challenges due to factors, such as, clutter/environmental noise, joint target and sensor state dependent measurement uncertainty, target-measurement association ambiguity, and sub-optimal sensor placement. The specific application that we consider is that of an underwater mobile sensor that measures the relative angle (i.e., bearing angle) to sources of acoustic noise in order to track one or more ships (targets) in a noisy environment. However, our contributions are generalizable for a variety of multitarget tracking applications. We build upon existing algorithms and address the problem of improving tracking performance for multiple maneuvering targets by incorporation several target motion models into a FISST tracking algorithm known as the probability hypothesis density filter. Moreover, we develop a novel method for associating measurements to targets using the Bayes factor, which improves tracking performance for FISST methods as well as other approaches to multitarget tracking. Further, we derive a novel formulation of Bayes risk for use with set-valued random variables and develop a real-time planner for sensor motion that avoids local minima that arise in myopic approaches to sensor motion planning. The effectiveness of our contributions are evaluated through a mixture of real-world and simulated data.
Aerodynamic and Aeroacoustic Analysis of Low Reynolds Number Propellers Using Higher-Order RANS Transition Turbulence Modeling
Pisharoti, Naina (Virginia Tech, 2024-06-05)
The advent of advanced vehicle concepts involving Urban Air Mobility (UAM) and small Unmanned Aerial Systems (sUAS) has brought about a new class of rotorcraft technology which operate predominantly in low-Reynolds ($Re$) number regimes. In such regimes, the flow experiences complex boundary layer phenomena like laminar separation, flow transition and reattachment. These effects are known to greatly alter the flow at and near the rotor wall, influencing its aerodynamic performance as well as the noise generated. Capturing these effects in our computational models is necessary to further our understanding of rotor aerodynamics and acoustics. The current study has introduced a novel RANS transition turbulence model, SSG/LRR-$\omega$-$\gamma$, that is capable of modeling different modes of transition involving natural, bypass, separation-induced and crossflow transition. The model framework uses a Reynolds stress transport model, SSG/LRR-$\omega$, as the base turbulence formulation and is coupled with Menter's $\gamma$ transition model. It was validated using a number of canonical cases that exhibited different transition mechanisms and the model performed equivalently or better than existing state-of-the-art transition models. It is worthy to note that the proposed model was able to perform well in three-dimensional flows, demonstrated using the case of a prolate spheroid. This underscores the capability of Reynolds stress models to accurately capture complex flow curvatures, improving upon the capabilities of linear eddy viscosity models. The transition model, integrated into OpenFOAM, was then employed to analyze two different UAV propellers. The rotor flow was examined using a URANS simulation with an overset grid. The objective was twofold: firstly, to validate the predictions generated by the proposed model for low-Reynolds number (low-$Re$) rotors, and secondly, to evaluate its effectiveness across a range of operating conditions. Comparisons were drawn against established fully turbulent and transition models. The analysis showed that transition models in general tended to be consistent in their predictions and less sensitive to changing operating conditions when compared to fully turbulent models. They also demonstrated the ability to accurately predict the mechanisms leading to separation and transition. Further, the proposed transition model demonstrated superior capability in capturing detailed flow features, particularly in the wake, compared to other fully turbulent and transition models, which is attributed to its Galilean invariant framework. To leverage the boundary layer information obtained from the proposed model, a semi-empirical broadband noise prediction method was implemented. This method utilized boundary layer data predicted by URANS simulations to estimate blade self-noise. An evaluation of the fully turbulent $k$-$\omega$ SST model and the proposed transition model revealed that both exhibited reasonable accuracy at lower rotor advance ratios. However, the transition model performed better at higher advance ratios. It was also observed that CFD-based approaches provided superior prediction accuracy compared to lower-fidelity aerodynamic models in the context of blade self-noise prediction Finally, the proposed aerodynamic and acoustic computational framework was applied to a design case study of swept propellers to understand the advantages of blade sweep on rotor aerodynamics and noise. A qualitative analysis of the flow suggested that the swept rotor exhibited lower levels of blade wake interaction compared to the unswept geometry, in line with the experimental observations.
Algorithm to enable intelligent rail break detection
Bhaduri, Sreyoshi (Virginia Tech, 2013-12-11)
Wavelet intensity based algorithm developed previously at VirginiaTech has been furthered and paired with an SVM based classifier. The wavelet intensity algorithm acts as a feature extraction algorithm. The wavelet transform is an effective tool as it allows one to narrow down upon the transient, high frequency events and is able to tell their exact location in time. According to prior work done in the field of signal processing, the local regularities of a signal can be estimated using a Lipchitz exponent at each time step of the signal. The local Lipchitz exponent can then be used to generate the wavelet intensity factor values. For each vertical acceleration value, corresponding to a specific location on the track, we now have a corresponding intensity factor. The intensity factor corresponds to break-no break information and can now be used as a feature to classify the vertical acceleration as a fault or no fault. Support Vector Machines (SVM) is used for this binary classification task. SVM is chosen as it is a well-studied topic with efficient implementations available. SVM instead of hard threshold of the data is expected to do a better job of classification without increasing the complexity of the system appreciably.
Analysis and Design of a Novel E-Core Common-Pole Switched Reluctance Machine
Lee, Cheewoo (Virginia Tech, 2010-03-03)
In this dissertation, a novel two-phase switched reluctance machine (SRM) with a stator comprised of E-core structure having minimum stator core iron is presented for low-cost high-performance applications. In addition, three new magnetic structures for the E-core SRM comprising two segmented stator cores or a monolithic stator core are proposed for good manufacturability, mechanically robustness, ease of assembly, and electromagnetic performance improvement. Each E-core stator in the segmented structure has three poles with two small poles at the ends having windings and a large center pole containing no copper windings. The common stator pole at the centers in the segmented E-core is shared by both phases during operation. Other benefits of the common poles contributing to performance enhancement are short flux paths, mostly flux-reversal-free-stator, constant minimum reluctance around air gap, and wide pole arc equal to one rotor pole pitch. Therefore, two additional common poles in the monolithic E-core configuration are able to significantly improve efficiency due to more positive torque and less core loss by the unique design. Using a full MEC analysis, the effect of the common-pole structure on torque enhancement is analytically verified. Efficiency estimated from the dynamic simulation is higher by 7% and 12% at 2000 rpm and by 3% and 7 % at 3000 rpm for the segmented and single-body SRMs, respectively, compared to a conventional SRM with four stator poles and two rotor poles. The new E-core SRMs are suitable for low-cost high-performance applications which are strongly cost competitive since all the new E-core SRMs have 20% cost savings on copper and the segmented E-core SRMs have 20% steel savings as well. Strong correlation between simulated and experimentally measured results validates the feasibility of the E-core common-pole structure and its performance. A simple step-by-step analytical design procedure suited for iterative optimization with small computational effort is developed with the information of the monolithic E-core SRM, and the proposed design approach can be applied for other SRM configurations as well. For investigating thermal characteristics in the two-phase single-body E-core SRM, the machine is modeled by a simplified lumped-parameter thermal network in which there are nine major parts of the motor assembly.
Applied Nonlinear Control of Unmanned Vehicles with Uncertain Dynamics
Morel, Yannick (Virginia Tech, 2009-04-17)
The presented research concerns the control of unmanned vehicles. The results introduced in this dissertation provide a solid control framework for a wide class of nonlinear uncertain systems, with a special emphasis on issues related to implementation, such as control input amplitude and rate saturation, or partial state measurements availability. More specifically, an adaptive control framework, allowing to enforce amplitude and rate saturation of the command, is developed. The motion control component of this framework, which works in conjunction with a saturation algorithm, is then specialized to different types of vehicles. Vertical take-off and landing aerial vehicles and a general class of autonomous marine vehicles are considered. A nonlinear control algorithm addressing the tracking problem for a class of underactuated, non-minimum phase marine vehicles is then introduced. This motion controller is extended, using direct and indirect adaptive techniques, to handle parametric uncertainties in the system model. Numerical simulations are used to illustrate the efficacy of the algorithms. Next, the output feedback control problem is treated, for a large class of nonlinear and uncertain systems. The proposed solution relies on a novel nonlinear observer which uses output measurements and partial knowledge of the system's dynamics to reconstruct the entire state for a wide class of nonlinear systems. The observer is then extended to operate in conjunction with a full state feedback control law and solve both the output feedback control problem and the state observation problem simultaneously. The resulting output feedback control algorithm is then adjusted to provide a high level of robustness to both parametric and structural model uncertainties. Finally, in a natural extension of these results from motion control of a single system to collaborative control of a group of vehicles, a cooperative control framework addressing limited communication issues is introduced.
Autonomous Underwater Vehicle (AUV) Propulsion System Analysis and Optimization
Schultz, James Allen (Virginia Tech, 2009-05-04)
One of the largest design considerations for autonomous underwater vehicles (AUV's) that have specific mission scenarios is the propulsive efficiency. The propulsive efficiency affects the amount of power storage required to achieve a specific mission. As the efficiency increases the volume of energy being stored decreases. The decrease in volume allows for a smaller vehicle, which results in a vehicle that requires less thrust to attain a specific speed. The process of selecting an efficient propulsive system becomes an iterative process between motor, propeller, and battery storage. Optimized propulsion systems for mission specific AUV's require costly motor and propeller fabrication which may not be available to the designer. Recent advancements in commercially available electric motors and propellers allows for cost effective propulsion systems. The design space selection of motors and propellers has recently increased due to component demand of remote control airplane and boats. The issue with such systems is how to predict small propeller and small motor performance interactions since remote control motor and propeller designers usually don't provide enough information about the performance of their product. The mission statement is to design a propeller and motor combination that will allow an autonomous underwater vehicle to travel large distances while maintaining good efficiency. The vehicle will require 12 N of thrust with a forward velocity of 2 m/s. The propeller needs to be larger than 2.5â due to inflow velocity interaction and smaller than 4â due to loss of thrust when in surface transit due to suction.
Back to Back Active Power Filter for Multi-Generator Power Architecture with Reduced dc-link Capacitor
Kim, Jong Wan (Virginia Tech, 2020-01-30)
Multi-pulse converters have been widely used for a multi-megawatt scale power generating system to comply with harmonic regulations. Among all types of multi-pulse converters, a 12-pulse converter is the most widely used due to the simple structure, which consists of a delta-delta and a delta-wye phase-shift transformer pair and it effectively mitigates undesirable harmonics from the nonlinear load. In the early 2000s, a shunt type passive front-end for a shipboard power system was proposed. By shunting the two gensets with 30° phase angle difference, a single phase-shift transformer effectively eliminates 5th and 7th harmonics. It achieves a significant size and weight reduction compared to a 12-pulse converter while keep the comparable harmonic cancellation performance. Recently, a hybrid type front-end was proposed. On top of the passive front-end, 3 phase active power filter was added and an additional harmonic cancellation was achieved to further eliminate 11th and 13th harmonics. However, the performance of both the passive and hybrid type front-end are highly dependent on the size of the line reactor in ac mains. A back to back active power filter is proposed in this dissertation to replace the phase-shift transformer in the multi-generator power architecture. The proposed front-end does not include phase-shift transformer and the size and the weight of the overall front-end can be significantly reduced. Due to the active harmonic compensation, the back to back front-end achieves better power quality and the line reactor dependency is improved. The number of required dc-link capacitors is reduced by half by introducing a back to back configuration and the capacitor size is reduced by adjusting the phase angle difference of genset to cancel out the most significant voltage harmonics in the shared dc-link bus. The overview of the existing shunt type front-end is provided and the concept of back to back active power filter is validated by simulation and prototype hardware. The comparison between existing front-end and the proposed front-end is provided to highlight the superior performance of back to back active front-end. The dc-link bus current and voltage ripple analysis is provided to explain the dc-link ripple reduction mechanism.
Bearing Estimation for Underwater Acoustic Source Using Autonomous Underwater Vehicle
Murali, Rohit (Virginia Tech, 2022-07-07)
This thesis describes the challenges involved in detecting sources of acoustic noise using an autonomous underwater vehicle (AUV) in real world environments. The initial part of this thesis describes the developments made for redesigning an acoustic sensing system that can be used to estimate the relative bearing between a source of acoustic noise and an AUV. With an estimate of the relative bearing, the AUV can maneuver toward the source of noise. The class of algorithms that are used to estimate bearing angle are known as beamforming algorithms. A comparison of the performance of a variety of beamforming algorithms is presented. When estimating the bearing to a source of noise from a small AUV, the noise of the AUV, especially its propulsor, pose significant challenges. Toward the goal of active cancellation of AUV self-noise, we propose placing an additional hydrophone inside the AUV in order to estimate the AUV self-noise that appears on the exterior hydrophones that are used for bearing estimation.
Cardiac Signals: Remote Measurement and Applications
Sarkar, Abhijit (Virginia Tech, 2017-08-25)
The dissertation investigates the promises and challenges for application of cardiac signals in biometrics and affective computing, and noninvasive measurement of cardiac signals. We have mainly discussed two major cardiac signals: electrocardiogram (ECG), and photoplethysmogram (PPG). ECG and PPG signals hold strong potential for biometric authentications and identifications. We have shown that by mapping each cardiac beat from time domain to an angular domain using a limit cycle, intra-class variability can be significantly minimized. This is in contrary to conventional time domain analysis. Our experiments with both ECG and PPG signal shows that the proposed method eliminates the effect of instantaneous heart rate on the shape morphology and improves authentication accuracy. For noninvasive measurement of PPG beats, we have developed a systematic algorithm to extract pulse rate from face video in diverse situations using video magnification. We have extracted signals from skin patches and then used frequency domain correlation to filter out non-cardiac signals. We have developed a novel entropy based method to automatically select skin patches from face. We report beat-to-beat accuracy of remote PPG (rPPG) in comparison to conventional average heart rate. The beat-to-beat accuracy is required for applications related to heart rate variability (HRV) and affective computing. The algorithm has been tested on two datasets, one with static illumination condition and the other with unrestricted ambient illumination condition. Automatic skin detection is an intermediate step for rPPG. Existing methods always depend on color information to detect human skin. We have developed a novel standalone skin detection method to show that it is not necessary to have color cues for skin detection. We have used LBP lacunarity based micro-textures features and a region growing algorithm to find skin pixels in an image. Our experiment shows that the proposed method is applicable universally to any image including near infra-red images. This finding helps to extend the domain of many application including rPPG. To the best of our knowledge, this is first such method that is independent of color cues.
Centralized Control of Power System Stabilizers
Sanchez Ayala, Gerardo (Virginia Tech, 2014-10-09)
This study takes advantage of wide area measurements to propose a centralized nonlinear controller that acts on power system stabilizers, to cooperatively increase the damping of problematic small signal oscillations all over the system. The structure based on decision trees results in a simple, efficient, and dependable methodology that imposes much less computational burden than other nonlinear design approaches, making it a promising candidate for actual implementation by utilities and system operators. Details are given to utilize existing stabilizers while causing minimum changes to the equipment, and warranting improvement or at least no detriment of current system behavior. This enables power system stabilizers to overcome their inherent limitation to act only on the basis of local measurements to damp a single target frequency. This study demonstrates the implications of this new input on mathematical models, and the control functionality that is made available by its incorporation to conventional stabilizers. In preparation of the case of study, a heuristic dynamic reduction methodology is introduced that preserves a physical equivalent model, and that can be interpreted by any commercial software package. The steps of this method are general, versatile, and of easy adaptation to any particular power system model, with the aggregated value of producing a physical model as final result, that makes the approach appealing for industry. The accuracy of the resulting reduced network has been demonstrated with the model of the Central American System.
CFD-informed Lumped Parameter Models Result In High-Fidelity Maneuvering Predictions of AUVs
Miller, Lakshmi Madhavan (Virginia Tech, 2023-07-11)
Recent developments in autonomous underwater vehicles (AUV) have created the need for a low cost AUV that is comparable in class and payload capabilities to existing, commercially available, expensive and sub-optimal crafts. The Navy is active in research of autonomous, unmanned, highly efficient, high speed underwater craft. Small, low cost AUVs capable of swarm control are of special interest for military mine applications. No matter the nature of the application or class of craft, a common challenge is the accuracy of maneuvering predic- tions. Maneuvering predictions not only affect design, but also the real time understanding of mission capabilities and endurance. Thus the proliferation of AUVs in recent times for commercial and defense applications have led to the need of higher fidelity of physics based lumped parameter models. The sensor data, along with maneuvering model data can tie into a more accurate trajectory. Multiple such incremental advances in the literature for prediction of maneuvering shall lead to a more accuracy. This work hopes to bridge some important gaps that ensure the creation of such a non-linear LPM to predict the maneuver- ing characteristics of an AUV using non linear hydrodynamic derivatives obtained through static and dynamic CFD. This model shall be implemented for the craft designed for DIVE technologies, our industrial sponsor and an in-house craft, the 690. This model shall also be made generalized for most submerged craft with a torpedo or slender hull form, with cruciform or X configuration of fins. This dissertation looks to provide the framework to identify CFD informed high fidelity dynamic model for AUVs. The model thus created shall be spe- cialized to account for specific important effects such as flow interaction among appendages, effect of using steady and unsteady maneuvers as CFD information and kinematic charac- teristics of captive maneuvers. The specific, innovative contributions in this dissertation are listed below: 1. Definition of a new stability index to incorporate effects of gravity at low-moderate speeds 2. Novel method for identification of hydrodynamic derivatives 3. Systematic and comprehensive study on the parameters affecting VPMM
Collision Avoidance Using a Low-Cost Forward-Looking Sonar for Small AUVs
Morency, Christopher Charles (Virginia Tech, 2024-03-22)
In this dissertation, we seek to improve collision avoidance for autonomous underwater vehicles (AUVs). More specifically, we consider the case of a small AUV using a forward-looking sonar system with a limited number of beams. We describe a high-fidelity sonar model and simulation environment that was developed to aid in the design of the sonar system. The simulator achieves real-time visualization through ray tracing and approximation, and can be used to assess sonar design choices, such as beam pattern and beam location, and to evaluate obstacle detection algorithms. We analyze the benefit of using a few beams instead of a single beam for a low-cost obstacle avoidance sonar for small AUVs. Single-beam systems are small and low-cost, while multi-beam sonar systems are more expensive and complex, often incorporating hundreds of beams. We want to quantify the improvement in obstacle avoidance performance of adding a few beams to a single-beam system. Furthermore, we developed a collision avoidance strategy specifically designed for the novel sonar system. The collision avoidance strategy is based on posterior expected loss, and explicitly couples obstacle detection, collision avoidance, and planning. We demonstrate the strategy with field trials using the 690 AUV, built by the Center for Marine Autonomy and Robotics at Virginia Tech, with a prototype forward-looking sonar comprising of nine beams.
Communication-Aware, Scalable Gaussian Processes for Decentralized Exploration
Kontoudis, Georgios Pantelis (Virginia Tech, 2022-01-25)
In this dissertation, we propose decentralized and scalable algorithms for Gaussian process (GP) training and prediction in multi-agent systems. The first challenge is to compute a spatial field that represents underwater acoustic communication performance from a set of measurements. We compare kriging to cokriging with vehicle range as a secondary variable using a simple approximate linear-log model of the communication performance. Next, we propose a model-based learning methodology for the prediction of underwater acoustic performance using a realistic propagation model. The methodology consists of two steps: i) estimation of the covariance matrix by evaluating candidate functions with estimated parameters; and ii) prediction of communication performance. Covariance estimation is addressed with a multi-stage iterative training method that produces unbiased and robust results with nested models. The efficiency of the framework is validated with simulations and experimental data from field trials. The second challenge is to perform predictions at unvisited locations with a team of agents and limited inter-agent information exchange. To decentralize the implementation of GP training, we employ the alternating direction method of multipliers (ADMM). A closed-form solution of the decentralized proximal ADMM is provided for the case of GP hyper-parameter training with maximum likelihood estimation. Multiple aggregation techniques for GP prediction are decentralized with the use of iterative and consensus methods. In addition, we propose a covariance-based nearest neighbor selection strategy that enables a subset of agents to perform predictions. Empirical evaluations illustrate the efficiency of the proposed methods
Competitive Algorithms and System for Multi-Robot Exploration of Unknown Environments
Premkumar, Aravind Preshant (Virginia Tech, 2017-09-08)
We present an algorithm to explore an orthogonal polygon using a team of p robots. This algorithm combines ideas from information-theoretic exploration algorithms and computational geometry based exploration algorithms. The algorithm is based on a single-robot polygon exploration algorithm and a tree exploration algorithm. We show that the exploration time of our algorithm is competitive (as a function of p) with respect to the offline optimal exploration algorithm. We discuss how this strategy can be adapted to real-world settings to deal with noisy sensors. In addition to theoretical analysis, we investigate the performance of our algorithm through simulations for multiple robots and experiments with a single robot.
Computationally Driven Algorithms for Distributed Control of Complex Systems
Abou Jaoude, Dany (Virginia Tech, 2018-11-19)
This dissertation studies the model reduction and distributed control problems for interconnected systems, i.e., systems that consist of multiple interacting agents/subsystems. The study of the analysis and synthesis problems for interconnected systems is motivated by the multiple applications that can benefit from the design and implementation of distributed controllers. These applications include automated highway systems and formation flight of unmanned aircraft systems. The systems of interest are modeled using arbitrary directed graphs, where the subsystems correspond to the nodes, and the interconnections between the subsystems are described using the directed edges. In addition to the states of the subsystems, the adopted frameworks also model the interconnections between the subsystems as spatial states. Each agent/subsystem is assumed to have its own actuating and sensing capabilities. These capabilities are leveraged in order to design a controller subsystem for each plant subsystem. In the distributed control paradigm, the controller subsystems interact over the same interconnection structure as the plant subsystems. The models assumed for the subsystems are linear time-varying or linear parameter-varying. Linear time-varying models are useful for describing nonlinear equations that are linearized about prespecified trajectories, and linear parameter-varying models allow for capturing the nonlinearities of the agents, while still being amenable to control using linear techniques. It is clear from the above description that the size of the model for an interconnected system increases with the number of subsystems and the complexity of the interconnection structure. This motivates the development of model reduction techniques to rigorously reduce the size of the given model. In particular, this dissertation presents structure-preserving techniques for model reduction, i.e., techniques that guarantee that the interpretation of each state is retained in the reduced order system. Namely, the sought reduced order system is an interconnected system formed by reduced order subsystems that are interconnected over the same interconnection structure as that of the full order system. Model reduction is important for reducing the computational complexity of the system analysis and control synthesis problems. In this dissertation, interior point methods are extensively used for solving the semidefinite programming problems that arise in analysis and synthesis.
Continuous Low-Thrust Trajectory Optimization: Techniques and Applications
Kim, Mischa (Virginia Tech, 2005-04-18)
Trajectory optimization is a powerful technique to analyze mission feasibility during mission design. High-thrust trajectory optimization problems are typically formulated as discrete optimization problems and are numerically well-behaved. Low-thrust systems, on the other hand, operate for significant periods of the mission time. As a result, the solution approach requires continuous optimization; the associated optimal control problems are in general numerically ill-conditioned. In addition, case studies comparing the performance of low-thrust technologies for space travel have not received adequate attention in the literature and are in most instances incomplete. The objective of this dissertation is therefore to design an efficient optimal control algorithm and to apply it to the minimum-time transfer problem of low-thrust spacecraft. We devise a cascaded computational scheme based on numerical and analytical methods. Whereas other conventional optimization packages rely on numerical solution approaches, we employ analytical and semi-analytical techniques such as symmetry and homotopy methods to assist in the solution-finding process. The first objective is to obtain a single optimized trajectory that satisfies some given boundary conditions. The initialization phase for this first trajectory includes a global, stochastic search based on Adaptive Simulated Annealing; the fine tuning of optimization parameters — the local search — is accomplished by Quasi-Newton and Newton methods. Once an optimized trajectory has been obtained, we use system symmetry and homotopy techniques to generate additional optimal control solutions efficiently. We obtain optimal trajectories for several interrelated problem families that are described as Multi-Point Boundary Value Problems. We present and prove two theorems describing system symmetries for solar sail spacecraft and discuss symmetry properties and symmetry breaking for electric spacecraft systems models. We demonstrate how these symmetry properties can be used to significantly simplify the solution-finding process.
Control and Modeling of High-Frequency Voltage Regulator Modules for Microprocessor Application
Li, Virginia (Virginia Tech, 2021-06-11)
The future voltage regulator module (VRM) challenges of high bandwidth control with fast transient response, high current output, simple implementation, and efficient 48V solution are tackled in this dissertation. With the push for control bandwidth to meet design specifications for microprocessor VRM with larger and faster load transients, control can be saturated and lost for a significant period of time during transient. During this time, undesirable transient responses such as large undershoot and ringback occurs. Due to the loss of control, the existing tools to study the dynamic behavior of the system, such as small signal model, are insufficient to analyze the behavior of the system during this time. In order to have a better understanding of the system dynamic performance, the operation the VRM is analyzed in the state-plane for a clear visual understanding of the steady-state and transient behaviors. Using the state-plane, a simplified state-plane trajectory control is proposed for constant on-time (COT) control to achieve the best transient possible for applications with adaptive voltage positioning (AVP). When the COT control is lost during a load step-up transient, the state-plane trajectory control will extend on-time to provide the a near optimal transient response. By observing the COT control law in the state-plane, a simplified state-plane trajectory control with analog implementation is proposed to achieve the best transient possible with smooth transitions in and out of the steady-state COT control. The concept of the simplified state-plane trajectory control is then extended to multiphase COT. For multiphase operation, additional operating behavior, such as phase overlapping during transient and interleaving during steady-state, need to be taken into consideration to design the desired state-plane trajectory control. A simple state-plane trajectory control with improved Ton extension is proposed and verified using multiphase COT control. After tackling the state-plane trajectory control for current mode COT, the idea is then extended to V2 COT. V2 COT is a more advanced current mode control which requires a more advanced state-plane trajectory control to COT. By calculating the intersection of the extended on-stage trajectory during transient and the ideal off trajectory in the form of a current limiting wall, a near optimal transient response can be achieved. For V2 COT with state-plane trajectory control, implementations using inductor vs. capacitor current, effect of component tolerance, and effect of IC delay are studied. The proposed state-plane trajectory control is then extended to enhanced V2 COT. Aside from tackling existing VRM challenges, the future datacenter 48V VRM challenge of a high efficiency, high power density solution to meet the VRM specifications is studied. The sigma converter is proposed for the 48V VRM solution due to exhibition of high efficiency and high-power density from hardware evaluation. An accurate model for the sigma converter is derived using the new modeling approach of modularizing the small signal components. Using the proposed model, the sigma converter is shown to naturally have very low output impedance, making the sigma converter suitable for microprocessor applications. The sigma converter is designed and optimized to achieve AVP and very fast transient response using both voltage-mode and current-mode controls.

Browsing by Author "Stilwell, Daniel J."

Results Per Page

Sort Options