View All Items
(1 - 3 of 3)
- Title
- IMITATING INDIVIDUALIZED FACIAL EXPRESSIONS IN A HUMAN-LIKE AVATAR THROUGH A HYBRID PARTICLE SWARM OPTIMIZATION - TABU SEARCH ALGORITHM.
- Creator
-
Husk, Evan, Gonzalez, Avelino, University of Central Florida
- Abstract / Description
-
This thesis describes a machine learning method for automatically imitating a particular person's facial expressions in a human-like avatar through a hybrid Particle Swarm Optimization - Tabu Search algorithm. The muscular structures of the facial expressions are measured by Ekman and Friesen's Facial Action Coding System (FACS). Using a neutral face as a reference, the minute movements of the Action Units, used in FACS, are automatically tracked and mapped onto the avatar using a hybrid...
Show moreThis thesis describes a machine learning method for automatically imitating a particular person's facial expressions in a human-like avatar through a hybrid Particle Swarm Optimization - Tabu Search algorithm. The muscular structures of the facial expressions are measured by Ekman and Friesen's Facial Action Coding System (FACS). Using a neutral face as a reference, the minute movements of the Action Units, used in FACS, are automatically tracked and mapped onto the avatar using a hybrid method. The hybrid algorithm is composed of Kennedy and Eberhart's Particle Swarm Optimization algorithm (PSO) and Glover's Tabu Search (TS). Distinguishable features portrayed on the avatar ensure a personalized, realistic imitation of the facial expressions. To evaluate the feasibility of using PSO-TS in this approach, a fundamental proof-of-concept test is employed on the system using the OGRE avatar. This method is analyzed in-depth to ensure its proper functionality and evaluate its performance compared to previous work.
Show less - Date Issued
- 2012
- Identifier
- CFH0004286, ucf:44949
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFH0004286
- Title
- High Performance Low Voltage Power MOSFET for High-Frequency Synchronous Buck Converters.
- Creator
-
Yang, Boyi, Shen, Zheng, Yuan, Jiann-Shiun, Sundaram, Kalpathy, Wu, Xinzhang, Xu, Shuming, University of Central Florida
- Abstract / Description
-
Power management solutions such as voltage regulator (VR) mandate DC-DC converters with high power density, high switching frequency and high efficiency to meet the needs of future computers and telecom equipment. The trend towards DC-DC converters with higher switching frequency presents significant challenges to power MOSFET technology. Optimization of the MOSFETs plays an important role in improving low-voltage DC-DC converter performance. This dissertation focuses on developing and...
Show morePower management solutions such as voltage regulator (VR) mandate DC-DC converters with high power density, high switching frequency and high efficiency to meet the needs of future computers and telecom equipment. The trend towards DC-DC converters with higher switching frequency presents significant challenges to power MOSFET technology. Optimization of the MOSFETs plays an important role in improving low-voltage DC-DC converter performance. This dissertation focuses on developing and optimizing high performance low voltage power MOSFETs for high frequency applications.With an inherently large gate charge, the trench MOSFET suffers significant switching power losses and cannot continue to provide sufficient performance in high frequency applications. Moreover, the influence of parasitic impedance introduced by device packaging and PCB assembly in board level power supply designs becomes more pronounced as the output voltage continues to decrease and the nominal current continues to increase. This eventually raises the need for highly integrated solutions such as power supply in package (PSiP) or on chip (PSoC). However, it is often more desirable in some PSiP architectures to reverse the source/drain electrodes from electrical and/or thermal point of view. In this dissertation, a stacked-die Power Block PSiP architecture is first introduced to enable DC-DC buck converters with a current rating up to 40 A and a switching frequency in the MHz range. New high- and low-side NexFETs are specially designed and optimized for the new PSiP architecture to maximize its efficiency and power density. In particular, a new NexFET structure with its source electrode on the bottom side of the die (source-down) is designed to enable the innovative stacked-die PSiP technology with significantly reduced parasitic inductance and package footprint.It is also observed that in synchronous buck converter very fast switching of power MOSFETs sometimes leads to high voltage oscillations at the phase node of the buck converter, which may introduce additional power loss and cause EMI related problems and undesirable electrical stress to the power MOSFET. At the same time, the synchronous MOSFET plays an important role in determining the performance of the synchronous buck converter. The reverse recovery of its body diode and the Cdv/dt induced false trigger-on are two major mechanisms that impact the performance of the SyncFET. This dissertation introduces a new approach to effectively overcome the aforementioned challenges associated with the state-of-art technology. The threshold voltage of the low-side NexFET is intentionally reduced to minimize the conduction and body diode related power losses. Meanwhile, a monolithically integrated gate voltage pull-down circuitry is proposed to overcome the possible Cdv/dt induced turn-on issue inadvertently induced by the low VTH SynFET.Through extensive modeling and simulation, all these innovative concepts are integrated together in a power module and fabricated with a 0.35(&)#181;m process. With all these novel device technology improvements, the new power module delivers a significant improvement in efficiency and offers an excellent solution for future high frequency, high current density DC-DC converters. Megahertz operation of a Power Block incorporating these new device techniques is demonstrated with an excellent efficiency observed.
Show less - Date Issued
- 2012
- Identifier
- CFE0004642, ucf:49885
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0004642
- Title
- FALCONET: FORCE-FEEDBACK APPROACH FOR LEARNING FROM COACHING AND OBSERVATION USING NATURAL AND EXPERIENTIAL TRAINING.
- Creator
-
Stein, Gary, Gonzalez, Avelino, University of Central Florida
- Abstract / Description
-
Building an intelligent agent model from scratch is a difficult task. Thus, it would be preferable to have an automated process perform this task. There have been many manual and automatic techniques, however, each of these has various issues with obtaining, organizing, or making use of the data. Additionally, it can be difficult to get perfect data or, once the data is obtained, impractical to get a human subject to explain why some action was performed. Because of these problems, machine...
Show moreBuilding an intelligent agent model from scratch is a difficult task. Thus, it would be preferable to have an automated process perform this task. There have been many manual and automatic techniques, however, each of these has various issues with obtaining, organizing, or making use of the data. Additionally, it can be difficult to get perfect data or, once the data is obtained, impractical to get a human subject to explain why some action was performed. Because of these problems, machine learning from observation emerged to produce agent models based on observational data. Learning from observation uses unobtrusive and purely observable information to construct an agent that behaves similarly to the observed human. Typically, an observational system builds an agent only based on prerecorded observations. This type of system works well with respect to agent creation, but lacks the ability to be trained and updated on-line. To overcome these deficiencies, the proposed system works by adding an augmented force-feedback system of training that senses the agents intentions haptically. Furthermore, because not all possible situations can be observed or directly trained, a third stage of learning from practice is added for the agent to gain additional knowledge for a particular mission. These stages of learning mimic the natural way a human might learn a task by first watching the task being performed, then being coached to improve, and finally practicing to self improve. The hypothesis is that a system that is initially trained using human recorded data (Observational), then tuned and adjusted using force-feedback (Instructional), and then allowed to perform the task in different situations (Experiential) will be better than any individual step or combination of steps.
Show less - Date Issued
- 2009
- Identifier
- CFE0002746, ucf:48157
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0002746
