Cochlear implantation is a surgical procedure in which an electrode array is permanently implanted in the cochlea to stimulate the auditory nerve and allow deaf people to hear. Traditional techniques require wide excavation of the temporal bone to ensure that the surgeon does not damage sensitive structures. Recently, a new minimally invasive approach was proposed in which a single hole is drilled on a straight trajectory from skull surface to the cochlea and allows the trajectory to be chosen pre-operatively in a CT. A major challenge with this approach is to determine, in the CT, a safe and effective drilling trajectory, i.e., a trajectory that with high probability avoids vital structures and effectively reaches the cochlea. These features lie within a few millimeters, the drill is one millimeter in diameter, and drill positioning errors are approximately a half a millimeter root-mean square. Thus, trajectory selection is both difficult and critical to the success of the surgery. In this thesis, a method is presented for finding optimal drilling trajectories with an approach designed to maximize safety under conditions of drill positioning error. Results are compared with trajectories chosen manually by an experienced surgeon. In tests on thirteen cases, the technique is shown to find approximately twice as many acceptable trajectories as those found manually.

We have studied the effects of moisture exposure and total dose irradiation on MOS low frequency noise. We compare the effects of moisture on the 1/f noise of nMOS and pMOS devices and how the noise changes for each with irradiation. We also investigate the gate-voltage and frequency dependences of the 1/f noise of these devices, and their changes with total dose irradiation. Large increases in low frequency noise were observed in pMOS devices exposed to moisture, before and after irradiation, while the noise for moisture-exposed nMOS devices was comparable to that of non-exposed nMOS devices. It has been suggested that these differences may be related to the inhibited diffusion of moisture to the gate oxides in nMOS devices due to phosphorus incorporation in the field oxide regions that are adjacent to and/or overlie the sources and drains. Phosphorus inhibits moisture diffusion; in contrast, boron can enhance moisture diffusion in the gate oxides of pMOS devices. Significant changes in the gate-voltage dependence of the 1/f noise were also observed for the pMOS devices after irradiation, while for the nMOS devices the gate-voltage dependence changed relatively little. These changes suggest that irradiation has altered the defect energy distribution in our pMOS devices, resulting in a more uniform distribution after irradiation. We conclude that these results can be explained by a simple trapping model, and that frequency and gate-voltage dependences of the noise can be a valuable means in revealing trap energy distributions in MOSFET devices before and after irradiation.

Friction stir welding (FSW) is a relatively new (1991) solid-state welding technique where the material is joined via mechanical stirring. In this dissertation, methods for the automatic monitoring and control of FSW are presented. Monitoring techniques include gap fit-up detection, tool misalignment detection and estimation, and tool wear detection. Additionally, a technology which accomplishes automatic joint-line tracking for FSW (through-the-tool tracking) is presented. Joint-tracking has shown to be very useful in other welding technologies and will likely improve the flexibility and applicability of FSW.

Hydrogenous species play a key role in radiation induced charge buildup in metal oxide semiconductor field effect transistors (MOSFETs). The effects of water on defect formation in MOSFETs before and after radiation exposure have been studied. Transistors built in Sandia National Laboratories' 4/3-ìm technology were exposed to water at 130 °C for times up to three weeks. The n-channel transistors did not show as much sensitivity to water as the p-channel transistors. Irradiation of the n-channel transistors exposed to moisture, followed by a long-term biased anneal, resulted in a small increase in interface-trap and oxide-trap charge densities in the gate oxides. Greatly enhanced post-irradiation defect generation was observed in the gate oxides of p-channel MOS transistors that were exposed to water. Low frequency (1/f) noise measurements also showed enhanced noise power spectral densities in the moisture-exposed p-channel transistors consistent with the enhanced post-irradiation increase in defect density. Phosphorus and boron dopant atoms are present in the field oxides of the n-channel and p-channel transistors because of source and drain implant steps. Boron accelerates water penetration and phosphorus suppresses water diffusion in SiO2. This can lead to enhanced water-induced defect formation in the gate oxides of p-channel transistors compared to n-channel transistors before and after irradiation. These results are significant for the performance of MOS technologies in non-hermetic environments where water can be present; in particular, the degradation of devices and circuits may be larger in these cases than expected from reliability and radiation tests that do not account for the additional degradation that can occur because of water and its reactions in SiO2.

Sub-100 nm technologies are more vulnerable than older technologies to single event effects (SEE) due to Moore's Law scaling trend. The increased SEE vulnerability has been attributed to the decrease in nodal charge for information storage, reduced nodal separation, and increased switching frequency. The effect of the reduced nodal separation is the increased probability of simultaneous charge collection at several nodes from a single ion-strike (called charge sharing). Charge sharing is a significant SEE issue because it can render circuit-level hardening techniques ineffective. Conventional SEE radiation-hardened by design (RHBD) approaches provide excellent protection against single event upsets (SEU) resulting from charge collection occurs on a single node. However, for sub-100 nm technologies, the probability of multiple node charge collection is significant, thwarting RHBD protection. As CMOS processes continue to scale, there is a continued decrease in nodal pitch, but virtually no change in the charge generation radius of the heavy-ion strike. Hence, charge sharing is a troubling reliability roadblock for advanced technologies. This dissertation introduces and details the charge sharing effect. It examines through finite element simulations, focused laser testing, and broadbeam heavy ion experiments the effects of charge sharing at the 130 nm and 90 nm CMOS technology nodes. Results include quantification of the all-important angle of incidence on device and circuit response. Further, this dissertation examines the effectiveness of several charge sharing mitigation techniques. The work presented in this dissertation directly impacts the SEE qualification techniques used by the radiation community for sub-100 nm technologies. The mitigation techniques proposed and verified are useful for improving the radiation hardness of advanced technologies, and provide designers with design guidelines applicable to space-deployed applications.

Data processing in scientific and workflow-oriented computing is carried out as analysis campaigns, which consist of an input dataset and a set of interdependent jobs. Traditionally, these massively parallel computations required the services of supercomputers. However, recent trends show that the share of scientific computing carried out on clusters of commodity computers is on the rise. Commodity computers yield the highest performance per dollar but exhibit intermittent faults, which can result in systemic failures when operated over long continuous periods for executing analysis campaigns. Diagnosing job problems and failures in this complex environment is difficult, especially when the success of a campaign can be affected by even a single job failure. Manual administration, though essential, is slow to respond to the intermittent faults. Therefore, an autonomic approach is required that can ensure that the resources of the cluster are used to the best possible extent and improve the reliability of jobs, even in the presence of hardware/software failures. <p> Model-based design is a formal system design methodology that has gained momentum in recent years as a sound methodology of applying computer-based modeling and synthesis methods to a variety of problem domains, including distributed systems. A benefit of using formal models is that they can be queried or transformed to produce a variety of domain specific artifacts, which are critical to deployment and execution of the system, but are tedious and error-prone to produce manually. <p> This dissertation presents the design and discusses applicability of a model-based cluster management framework called Scientific Computing Autonomic Reliability Framework (SCARF). Basic components of this framework are distributed monitoring units, fault-mitigation units and a workflow-management system for dealing with workflow-specific concerns in case of failures. <p> Model-based techniques are used to capture workflow specifications, along with pre, post conditions and invariants for checking the validity of system state during execution. Formal data models are used to provide provenance and execution tracking of workflow jobs. Health monitoring is provided by synchronized, light-weight, distributed sensors that are augmented with a real-time fault-mitigation framework. This framework consists of hierarchical fault management entities called reflex engines, which use a timed automaton based abstraction for capturing failure management strategies. These engines track the state of components under their management zone and initiate reflexive mitigation actions upon occurrence of certain events or timeouts. This mitigation framework is verified against properties written in timed computation tree logic (TCTL).

In this thesis a system is developed whereby ISAC, a soft-arm control humanoid robot, can observe, track, and imitate hand motions made by a human being. This is accomplished by making use of the OpenCV libraries for Haar object detection and pre-trained Haar classifiers to detect the humans face and hand, applying stereo vision geometry to identify the relative locations of the face and hand and to map those coordinates onto the workspace of ISAC, and by transmitting those coordinates via UDP to the arm controller, which interpolates and activates the corresponding arm motions. Thus, ISAC can imitate motions in real-time. These motions are also stored in a database on the arm control computer for later use.

Real-time search techniques have been used extensively in the areas of task planning and decision making. In order to be effective, however, these techniques require task-specific domain knowledge in the form of heuristic or utility functions. These functions can either be embedded by the programmer, or learned by the system over time. Unfortunately, many of the reinforcement learning techniques that might be used to acquire this knowledge generally demand static feature vector representations defined a priori. Current neurobiological research offers key insights into how the cognitive processing of experience may be used to alleviate dependence on pre-programmed heuristic functions as well as on static feature representations. Research also suggests that emotion-based appraisals are influenced by such processing and that these appraisals integrate with the cognitive decision-making process, providing a range of useful and adaptive control signals that focus, inform, and mediate deliberation. While the integration of emotion and cognition may limit an agents ability to find the most optimal solution, it is argued here that many real-world tasks only require adequate solutions, so long as those solutions can be identified quickly. This dissertation investigates how experience, stored within episodic memory, may be processed to develop a set of emotion-based appraisals that can then be used as a guide for future deliberation. These appraisals include techniques for identifying relevant information, estimating utility, predicting and adjusting for urgency, and checking fit. When derived from experience, each appraisal should contribute uniquely to deliberation and enable robotic systems to quickly determine acceptable solutions for complex tasks.

Recent goals in space missions require innovative technologies to develop and to build infrastructure for space exploration. NASA plans to have robots construct modular systems and habitats and prepare them for life support prior to the arrival of astronauts. Modular truss structures are the hardware of choice for such tasks, requiring robots to grasp, manipulate, and assemble these trusses through cooperative work. Remote teleoperation for the purposes of assembly is difficult, tiring, and prone to errors. More so when the task involves numerous operators and robots. Additionally, the communication latencies encountered in space missions preclude the instantaneous teleoperation of the kind that would be required to control such a team of robots. A viable solution is possible through the use of a variable autonomy architecture. In this architecture, teleoperators could guide a robot to an optimal location where grasp or assembly can take place. The system would then slide to autonomous mode and allow the robot to perform the lower-level task autonomously. This thesis proposes a control framework that enables short term autonomy and cooperative assembly by two robots of highly differing morphology. The work advances the capabilities of heterogeneous robots to cooperate on some of the low-level tasks necessary for autonomous assembly. The proposed control strategy allows independent robots in loosely structured environments to execute insertion tasks. The approach is to modularize and encapsulate the control problem by recasting it in terms of locally robust controllers. The controllers do not require explicit planning, rather through sensory stimuli they drive the system to optimal state configurations. A number of novel basis controllers are presented in this work and used by a pneumatically actuated and highly compliant humanoid robot and a rigid industrial manipulator. Active-static and active-active robotic roles are tested in a number of different experiments to study the performance and efficacy of the approach in assembly tasks. Results show that the proposed framework effectively executed the assembly tasks in a variety of cooperative schemes, experiencing the fastest times in scenarios where both robots actively drove the insertion.

A multimodal attentional system for a robot operating in a dynamic environment is described. The system at once enables a robot (a) to maintain a Focus of Attention (FOA) on the region of space most important to its current task, (b) to shift that focus to attend to new stimuli, and (c) to ignore repetitive stimuli of no importance nor of any danger. That is to say, the system exhibits both sensitization and habituation. The system is both task- and data-driven. It is a collection of abstract data types that form a locally-connected network of processing nodes and an FOA object that is connected to all. Each connection permits two-way data flow. The connectivity of the nodes forms a geodesic tessellation of a virtual sphere centered on the base frame of the robot. Parallel independent sensory processing modules (SPM) write data to the nodes as a function of the direction from which the sensory stimuli originated. Those data include abstracted sensory information, a measure of data's importance, the direction from which it was received, and the time of reception. Each node: (1) can receive data from any of the sensory processors and of any modality, (2) has a two-stage memory for each modality, (3) computes, modulates, and maintains an activation value for each modality as a function of the data and its importance, and (4) communicates with its neighbors to share sensory data and activation values. The FOA maintains a prioritized list of tasks and an identifiers for the node with the largest active and the node with the largest current change in activation. The aggregate of simple local processes at each node together with a global task specification cause an FOA to emerge with the properties (a)-(c) above. Background on the network, on attention, and on habituation are presented. Experiments are described that demonstrate the strengths and weaknesses of the approach.