The sensory abilities of humanoid robots are progressing every year with the development of advanced and complex sensors. At the same time, humanoid robots are learning to coordinate sensory events in their environments with the actions that they produce. This dissertation presents results of mechanisms developed to allow a humanoid robot to discover important sensory events, both internal and external, and to determine which of those events requires attention. The mechanisms are implemented on the Sensory Ego-Sphere (SES), a spatio-temporal data structure used by the robot as an interface between sensing and cognition. The mechanisms presented are event binding and attention. Attention gives a robot the ability to focus on areas of its environment that are salient, that contain events related to the robots current tasks or events that are unexpected and sudden. It also enables the robot to disregard irrelevant areas, that contain events unrelated to current tasks or events that are habitual. Event binding selects all co-occurring events near the focus of attention that originated from the same source. The architecture that uses these capabilities and designs for attention and event binding mechanisms are presented. Salience of events is determined by incidence of the event on the SES, relevance of the event to current tasks and habituation of the event over time. Therefore, salience spikes around events that occur in the same area around the same time. The area with the highest salience is selected as the focus of attention. Event binding examines the spatial and temporal features of events in the focal area and tags those events as co-occurring. Experiments were performed on ISAC, Vanderbilt Universitys humanoid and Robonaut, the NASA/DARPA humanoid. Results from these experiments are presented and discussed. The results show that with the Sensory Ego-Sphere, a humanoid can detect the most salient areas of its environment and bind together those events that originated from a single source.

It is really challenging to develop large scale, real time embedded systems that are fault adaptive and tolerant. The interactions between different components of the system increase with the size of the system, and can be difficult to trace as the system grows in size. Well designed simulations can play a vital role in understanding such large scale systems. Any number of fault scenarios can be generated and tested in a simulation environment, safely without risking the actual system. This thesis describes the development of such simulations using Matlab Simulink and Stateflow. These simulations are developed using behavioral and structural models of components at the desired level of modeling detail. These models are specifically designed for the proposed system that will be used in BTeV project, a Fermi National Laboratorys high energy physics experiment.

The work in this thesis seeks to integrate the motion of a humanoid robot with its auditory and visual sensory information to achieve various reflex actions that mimic those of people. Such reflexes in the form of reach-grasp behaviors can enable the robot to learn through experience its own state and that of the world. A humanoid robot with auditory capabilities, stereo vision, and artificial pneumatic arms and hands was used to demonstrate tightly coupled sensory-motor behaviors in five different demonstrations. The complexity of succeeding demonstrations was increased to show that the reflexive sensory-motor behaviors combine to perform increasingly complex tasks. The humanoid robot executed these tasks effectively and established the ground-work for the further development of hardware and software systems, sensory-motor vector-space representations, and coupling with higher level cognition.

This dissertation presents a method to extract features salient to a mobile robot navigation task in a specific environment. The extraction process is bootstrapped by a human operators tele-operation and is based on the sensory-motor coordination principle. Salient feature extraction consists of three steps: tele-operation, offline association, and evaluation. First, the mobile robot is tele-operated in an environment along a path several times. All sensory data and motor drive commands are recorded. Then these recorded sensory-motor sequences are partitioned into episodes according to the changes in the motor commands. Salient features are then extracted by using two statistical criteria: consistency and correlation with the motor commands within an interval around the episode boundaries. Finally, these features are used to drive the robot in the learned environment. Two sets of experiments, in both indoor and outdoor environments, were performed. The results endorsed this methodology.

The continuous demand for higher packing density and faster operating speeds in modern digital CMOS circuits has driven the scaling of the metal-oxide-silicon field effect transistors. The use of scaling schemes which do not keep the internal electric fields constant results in highly energetic carriers in the scaled devices. This work presents a study of the degradation of MOSFET parameters, which effect circuit operation, due to processes initiated by injection of high energy carriers into the gate oxide. The characterization techniques used to measure and attribute these parameter shifts to the underlying processes, and the use of empirical, semi-empirical, and physical models to predict the time dependence of the parameter degradation during circuit operation have been investigated. The results of this work clearly highlight the deficiencies of the conventional degradation models and provide a basis for an approach to degradation modeling based on fundamental physical mechanisms.

Oxygen vacancies have long been known to be the dominant intrinsic defect in amorphous SiO2 . They exist, in concentrations dependent on processing conditions, as neutral defects in thermal oxides without usually causing any significant deleterious effects, with some spatial and energy distribution. During irradiation they can capture holes and become positively charged E´-centers, contributing to device degradation. Over the years, a considerable database has been amassed on the dynamics of E´-centers in bulk SiO2 films, and near the interface under different irradiation and annealing conditions. Theoretical calculations so far have revealed the basic properties of prototype oxygen vacancies, primarily as they behave in either a crystalline quartz environment, or in small clusters that serve as a substitute for a real amorphous structure. To date at least three categories of E´ centers, existing at or above room temperature, have been observed in SiO2 . The unifying feature is an unpaired electron on a threefold coordinated silicon atom, having the form O3 ? Si*. Feigl et al. identified the E´1 center in crystalline a-quartz as a trapped hole on an oxygen vacancy, which causes an asymmetrical relaxation, resulting in a paramagnetic center. The unpaired electron in the E´1 center is localized on the three-fold coordinated Si atoms, while the hole is localized on the other Si atom. Results from an ab initio statistical simulation examination of the behaviors of oxygen vacancies, within amorphous structures, identify a new form of the E´-center, the E´g5, and help in the understanding of the underlying physical mechanisms involved in switched-bias annealing, and electron paramagnetic resonance (EPR) studies. The results also suggest a common border trap, induced by trapped holes in SiO2, is a hole trapped at an O vacancy defect, which can be compensated by an electron, as originally proposed by Lelis and co-workers at Harry Diamond Laboratories. This dissertation provides new insights into the basic mechanisms of a-SiO2 defects, and provides a link between basic mechanisms and Electronic Design Automation (EDA) tools, providing an enhanced design flow for radiation-resistant electronics.

Chemical vapor deposited (CVD) diamond or related carbon materials are excellent materials for electron field emitters. However, the key factor to successful exploitation of CVD diamond for vacuum microelectronic applications is closely related to further understanding and control of the physics, materials, and microfabrication technology. To achieve this goal, this research has been focused on two main issues: 1) Design, fabricate, characterize, and investigate the physics of diamond field emissions utilizing shape-designed diamond microtip structures. 2) Develop fabrication processes for achieving monolithic diamond vacuum diodes and triodes, evaluate and model their field emission performance. In the first part of this research, a uniquely engineered mold transfer process has been developed for the fabrication of micro-patterned pyramidal diamond emitters with uniform microtip structures. In addition, methods to improve the diamond field emission behavior have been systematically studied. These include the incorporation of sp2 into diamond tips, vacuum-thermal-electric (VTE) treatment, p-type doping, and tip sharpening. Measurements from field emission of the fabricated diamond microtips have achieved low turn-on field of ~1 V/mm. Hypotheses to explain the resulting diamond field emission enhancement have been proposed and modeled. The second part of this research focused on the development of diamond field emission devices: monolithic diamond vacuum diodes and triodes. For monolithic diamond vacuum diodes, fabrication methods, namely, (1) electrostatic bonding, (2) self-align volcano anode, and (3) self-align-anode molding (utilizing standard, epitaxial, and SOI wafers) have been developed. Monolithic diamond vacuum diodes have low turn-on voltage achieving the lowest turn-on voltage of 0.7 V and high emission current. Also, the emission current was observed to be temperature insensitive up to 200 °C. For the monolithic diamond vacuum triode, two fabrication methods were developed, integrated anode by electrostatic bonding and integrated anode utilizing SOI substrate. Triode emission characteristics were studied as a function of the device parameters of anode-cathode spacing, diamond doping, and array sizes. The anode emission current of the diamond vacuum triode was modeled based on the modified Fowler-Nordhiem (F-N) triode equation and a new empirical model for the emission transport factor. Diamond vacuum triodes with good emission performance, low gate turn-on voltage (10 V), high dc voltage gain (800), and high transconductance (100 mS) have been achieved. Moreover, a diamond triode amplifier with characteristics of high ac voltage gain (~65) and large ac output voltage (~90 Vpeak-peak) was demonstrated. The performance of these devices demonstrates their potential use for vacuum microelectronic applications.

We have studied 1/f noise and total-dose response associated with Al2O3/SiOxNy/Si(100) gate dielectrics. Both the radiation-induced threshold-voltage shifts and the low-frequency noise are significantly larger than are typically observed for high-quality thermal SiO2 thin films of comparable thicknesses. Charge trapping effects and low frequency noise are studied in the buried oxides of fully depleted nMOS SOI transistors. Silicon implantation in the buried oxide creates a higher density of oxygen vacancy-related defects that reduce the net oxide-trap charge, but increase the back-channel 1/f noise. The 1/f noise of MOSFETs fabricated on silicon-implanted SOI buried oxides shows little change after 1 Mrad(SiO2) irradiation. The temperature and frequency dependences of the 1/f noise of back channel SOI nMOS transistors shows thermally-activated charge exchange between the Si channel and defects in the buried oxide. Silicon implantation also creates shallow electron traps in the buried oxide, leading to large bias instabilities. Whether these traps are filled or empty does not significantly affect the 1/f noise. A detailed study of the 1/f noise, temperature dependence of charge trapping, and radiation response of these SOI nMOSFET transistors shows that charge exchange with shallow electron traps in the buried oxide occurs mostly via tunneling. Low frequency noise in the double-gate mode of device operation is also investigated, and found to help mitigate the 1/f noise in fully depleted SOI MOSFETs.

Hot-carrier-induced degradation is a significant reliability concern in aggressively scaled metal-oxide-semiconductor~(MOS) transistors. The physical mechanisms responsible for hot-carrier degradation have been studied over the past several decades in order to devise methods to mitigate their detrimental effects. Several empirical and semi-empirical models have been popularly used in the industry to estimate the hot-carrier lifetimes of devices produced using specific semiconductor technologies. However, such methods use several simplifying assumptions that are not applicable to ultra-small geometry MOS devices. Furthermore, such models are not capable of predicting hot-carrier reliability variation with technology modifications. Such capabilities are essential for optimizing a technology to meet performance as well as reliability goals. In this dissertation, we have developed a hot-carrier modeling approach that avoids the assumptions made by traditional hot-carrier modeling techniques. Physical models for mechanisms that are significant in aggressively scaled device geometries have been included. In particular, models for carrier transport and heating in the semiconductor and injection into the insulator layers have been carefully selected for application to short-channel devices. A comprehensive model for the transport of injected carriers in the oxide and their interactions with defects in the oxide and at the oxide-semiconductor interface has been developed. The effects of long-term carrier trapping and interface trap generation due to hot-carrier injection have been simulated using this approach in a set of commercial and non-commercial technologies. The framework presented here represents an essential component of the technological design process for assuring hot-carrier reliability in current and future technologies.

The radiation response and long term reliability of alternative gate dielectrics will play a critical role in determining the viability of these materials for use in future space applications. The total dose radiation responses of several near and long term alternative gate dielectrics to silicon dioxide are discussed. The midgap voltage shift increases monotonically with dose and depends strongly on both dielectric thickness and processing. The thinnest dielectrics, of most interest to industry, are extremely hard to ionizing irradiation, exhibiting only a few millivolts of shift at a total doses of 1 Mrad(SiO2) or more. Oxygen anneals are found to significantly improve the total dose radiation response and induce a small amount of capacitance-voltage hysteresis. The standard radiation-induced-trapping efficiency equation is adapted for calculating effective trapping efficiencies in alternative gate dielectrics and used to compare the radiation responses of several materials. The alternative gate dielectrics discussed here are shown to have effective trapping efficiencies which are up to 15 to 20 times larger than thermal oxide of comparable electrical thickness. The effects of common reliability screens such as time dependent dielectric breakdown tests, ``burn-in' tests, and bias-stress tests are also discussed. Constant-voltage accelerated life testing has shown these devices have failure distributions with a large population of early extrinsic failures. Baking can degrade some devices by reducing the oxide capacitance and inducing hysteresis. Additionally, large applied voltages inject excess charge into alternative dielectrics with low conduction band offset energies, which can lead to a overestimation of the radiation hardness of these materials. Alternative gate dielectrics have shown encouraging radiation hardness, but there are still several engineering problems that must be addressed before they can be reliably incorporated in future space electronics.