Browsing by Author "Paul Ro, Committee Member"

Now showing 1 - 10 of 10

BATMAV - A Bio-Inspired Micro-Aerial Vehicle for Flapping Flight.
(2010-11-04) Bunget, Gheorghe; Stefan Seelecke, Committee Chair; Edward Grant, Committee Member; Ashok Gopalarathnam, Committee Member; Paul Ro, Committee Member
Competitive Relative Performance and Fitness Selection for Evolutionary Robotics
(2003-05-21) Nelson, Andrew Lincoln; Edward Grant, Committee Chair; Mark White, Committee Member; Paul Ro, Committee Member; Wesley E Snyder, Committee Member; John Muth, Committee Member
Evolutionary Robotics (ER) is a field of research that applies evolutionary computing methods to the automated design and synthesis of behavioral robotics controllers. In the general case, reinforcement learning (RL) using high-level task performance feedback is applied to the evolution of controllers for autonomous mobile robots. This form of RL learning is required for the evolution of complex and non-trivial behaviors because a direct error-feedback signal is generally not available. Only the high-level behavior or task is known, not the complex sensor-motor signal mappings that will generate that behavior. Most work in the field has used evolutionary neural computing methods. Over the course of the preceding decade, ER research has been largely focused on proof-of-concept experiments. Such work has demonstrated both the evolvablility of neural network controllers and the feasibility of implementation of those evolved controllers on real robots. However, these proof-of-concept results leave important questions unanswered. In particular, no ER work to date has shown that it is possible to evolve complex controllers in the general case. The research described in this work addresses issues relevant to the extension of ER to generalized automated behavioral robotics controller synthesis. In particular, we focus on fitness selection function specification. The case is made that current methods of fitness selection represent the primary factor limiting the further development of ER. We formulate a fitness function that accommodates the Bootstrap Problem during early evolution, but that limits human bias in selection later in evolution. In addition, we apply ER methods to evolve networks that have far more inputs, and are of a much greater complexity than those used in other ER work. We focus on the evolution of robot controllers for the competitive team game Capture the Flag. Games are played in a variety of maze environments. The robots use processed video data requiring 150 or more neural network inputs for sensing of their environment. The evolvable artificial neural network (ANN) controllers are of a general variable-size architecture that allows for arbitrary connectivity. Resulting evolved ANN controllers contain on the order of 5000 weights. The evolved controllers are tested in competitions of 240 games against hand-coded knowledge-based controllers. Results show that evolved controllers are competitive with the knowledge-based controllers and can win a modest majority of games in a large tournament in a challenging world configuration.
A Computational Intelligence Approach to the Mars Precision Landing Problem
(2008-04-21) Birge, Brian Kent III; Mark White, Committee Member; Paul Ro, Committee Member; Larry Silverberg, Committee Member; Gerald Walberg, Committee Chair
Fault Detection/Isolation and Fault Tolerant Control for Hypersonic Vehicle.
(2010-05-14) Cai, Xuejing; Fen Wu, Committee Chair; Branda Nowell, Committee Member; Paul Ro, Committee Member; Gregory Buckner, Committee Member; Fuh-Gwo Yuan, Committee Member
H-infinity Control of Active Magnetic Bearings: An Intelligent Uncertainty Modeling Approach
(2004-12-01) Gibson, Nathan Scott; Gregory D. Buckner, Committee Chair; Fen Wu, Committee Member; Paul Ro, Committee Member; Jeff W. Eischen, Committee Member
Robust control techniques require a dynamic model of the plant and bounds on model uncertainty to formulate control laws with guaranteed stability. Although techniques for modeling and identifying dynamic systems are well established, very few procedures exist for estimating uncertainty bounds. In the case of H-infinity control synthesis, a conservative weighting function for model uncertainty is usually chosen to ensure closed-loop stability over the entire operating space. The primary drawback of this conservative, "hard computing" approach is reduced performance due to the number of plants the resulting controller can stabilize. This paper demonstrates a novel "soft computing" approach to estimate bounds of model uncertainty resulting from parameter variations, unmodeled dynamics, and non-deterministic processes in dynamic plants. This approach uses confidence interval networks (CINs), radial basis function neural networks trained using asymmetric bilinear error cost functions, to estimate confidence intervals on the uncertainty associated with nominal linear models for robust control synthesis. This research couples the hard computing features of H-infinity control with the soft computing characteristics of intelligent system identification, and realizes the combined advantages of both. Experimental demonstrations conducted on a multivariable, flexible-rotor active magnetic bearing system confirm these capabilities.
LPV H∞ Control for the Longitudinal Dynamics of a Flexible Air-Breathing Hypersonic Vehicle.
(2010-11-02) Hughes, Hunter; Fen Wu, Committee Chair; Lawrence Silverberg, Committee Member; Jules Silverman, Committee Member; Gregory Buckner, Committee Member; Paul Ro, Committee Member
A Multi-Channel Power Controller for Actuation and Control of Shape Memory Alloy Actuators.
(2010-07-29) Hangekar, Rohan; Stefan Seelecke, Committee Chair; Gregory Buckner, Committee Member; Paul Ro, Committee Member
Neural Networks for the Dynamics of Multi-Link Systems
(2005-03-28) Tracy, William Christopher; Larry Silverberg, Committee Chair; Greg Buckner, Committee Member; Paul Ro, Committee Member
This paper developed specialized artificial neural networks for dynamical systems. For single-degree-of-freedom systems, specialized networks were developed for a) first-order linear, b) second-order linear, complex, first-order, and c) second-order with sigmoid damping (a generalization of viscous damping and dry friction damping). Digitization errors were eliminated by a specialized sub-network that corrected amplitude and phase output. Next, a network was developed for normal-mode systems. Finally, networks were developed for multi-link systems. Trigonometric nonlinearities were handled by the activation functions and multiplicative nonlinearities were handled by a custom sub-network. The treatment of the trigonometric nonlinearities and the multiplicative nonlinearities are kernels that can be expanded into specialized networks for a broad class of multi-link systems. The paper closes with an illustrative example of a three-link system resembling an upper arm, forearm, and hand. The three-link system is optimized to throw a basketball in a hoop with minimum effort.
Process Control Modeling and Real-Time Error Compensation in Tube Hydroforming.
(2010-06-30) Kilonzo, Obadiah; Gracious Ngaile, Committee Chair; Yuan-Shin Lee, Committee Member; Paul Ro, Committee Member; Fen Wu, Committee Member
Rotary Clock based High-Frequency ASIC Design Methodology
(2007-11-01) Yu, Zhengtao; W. Rhett Davis, Committee Member; Alex Huang, Committee Member; Xun Liu, Committee Chair; Paul Ro, Committee Member
With the increase of operating frequencies and dimensions, modern VLSI chips consume substantial power dissipation. For synchronous circuit designs, the clock network is a major power consumer, often contributing more than 20\% of overall power consumption. Consequently, power-efficient clock distribution methods have been extensively researched. In particular, the rotary clock technique is one low-power clocking approach that utilizes the LC oscillating principle to reduce power consumption. However, there are several design challenges that prevent the application of rotary technique. In this dissertation, we have developed a software tool based on the method of partial element equivalent circuit (PEEC) that is capable of extracting the SPICE netlist from the layout specification of a rotary clock design. Using our tool, linked various design parameters of a rotary clock design to its oscillation frequency and power dissipation. We then propose a power minimization algorithm. Our algorithm derives a rotary clock structure that dissipates the minimal power while satisfying the clock dimension requirement and oscillating at the target frequency with the given clock load. We then developed a design methodology to implement the rotary clock based VLSI system. In particular, we present a circuit optimization scheme called skew spreading for rotary clock. Given an edge-triggered sequential circuit, skew spreading relocates the registers and derives the corresponding required clock arrival times, or skews, so that all skew values are distributed evenly in a preselected time window without changing the circuit functionality or the operating speed. We make the first attempt to design rotary based circuit by proposing a unified clock and circuit design methodology. Given a sequential circuit and a clock frequency, our scheme derives a rotary clock network and a functionally equivalent circuit so that they can be integrated to operate reliably at the target frequency. We have developed a physical design flow for rotary clock based design to address placement and timing issues. We have investigated the phase-locked-loop design under rotary clock technique. Our experiment results demonstrate that the charge-pump based PLL can be applied to to rotary clock network with 7% tuning range to cover the frequency deviation due to the process variations. Based on this design methodology, we have implemented a parallel transpose direct form 10-tap programmable FIR filter. Our experiment results show that in comparison with conventional clock tree based design we have achieved 12.8% total power savings and 34.6% clock tree network power savings without degrading the speed of circuits. In addition, the rotary clock based FIR filer has the peak current reduction of more than 40% with area degradation of less than 2% in comparison with clock tree based FIR filter design.