Wearable agents

My group at Georgia Tech is dedicated to creating wearable agents. By necessity, our work combines on-body perception, user modeling, and human-computer interfaces. The Web site, www.innovations.gatech.edu, demonstrates our first attempt at such an agent. The prototypes on the site predict the user's next location, perceive the user's availability, and help schedule an appointment based on a conversation with another researcher. The system is an effort to demonstrate how to make aspects of cell phone use more socially graceful through a wearable agent. As of this writing, an agent system that integrates all the demonstrated functionality is too imprecise and requires too many resources to be practical, but each perception subsystem described is operational in the laboratory.     

Functional MRI evidence for an abstract, not perceptual, word-form area.

Previous studies have found an area in left ventral visual cortex that responds more to words and pseudowords than to consonant strings. Does this area respond to the perceptual form of wordlike stimuli, or is it responding to some more abstract, linguistic property, such as orthographic regularity (i.e., conformity with the spelling rules of the language)? During a functional magnetic resonance imaging experiment, 90 right-handed male participants read alternating-case words and pseudowords, which are orthographically regular but are perceptually unfamiliar. These stimuli activated the same area that was activated by pure-case words and pseudowords. These results suggest that the response of the so-called word-form area is not based on perceptual familiarity but rather on some more abstract feature such as orthographic regularity. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Camelina Meal Increases Egg n-3 Fatty Acid Content Without Altering Quality or Production in Laying Hens

Camelina sativa is an oilseed plant rich in n-3 and n-6 fatty acids and extruding the seeds results in high protein meal (~40%) containing high levels of n-3 fatty acids. In this study, we examined the effects of feeding extruded defatted camelina meal to commercial laying hens, measuring egg production, quality, and fatty acid composition. Lohmann White Leghorn hens (29 weeks old) were randomly allocated to three dietary treatment groups (n = 25 per group) and data was collected over a 12 week production period. All the treatment groups were fed a corn soy based experimental diet containing 0% (control), 5, or 10% extruded camelina meal. We found no significant differences in percent hen-day egg production and feed consumed per dozen eggs. Egg shell strength was significantly higher in both camelina groups compared to the controls. Egg total n-3 fatty acid content increased 1.9- and 2.7-fold in 5 and 10% camelina groups respectively relative to the control. A similar increase in DHA content also occurred. Further camelina meal did not alter glucosinolate levels and no detectable glucosinolates or metabolic product isothiocyanates were found in the eggs from either the 5 or 10% camelina groups. These results indicate that camelina meal is a viable dietary source of n-3 fatty acids for poultry and its dietary inclusion results in eggs enriched with n-3 fatty acids.     

Activity recognition of assembly tasks using body-worn microphones and accelerometers

In order to provide relevant information to mobile users, such as workers engaging in the manual tasks of maintenance and assembly, a wearable computer requires information about the user's specific activities. This work focuses on the recognition of activities that are characterized by a hand motion and an accompanying sound. Suitable activities can be found in assembly and maintenance work. Here, we provide an initial exploration into the problem domain of continuous activity recognition using on-body sensing. We use a mock "wood workshop” assembly task to ground our investigation. We describe a method for the continuous recognition of activities (sawing, hammering, filing, drilling, grinding, sanding, opening a drawer, tightening a vise, and turning a screwdriver) using microphones and three-axis accelerometers mounted at two positions on the user's arms. Potentially "interesting” activities are segmented from continuous streams of data using an analysis of the sound intensity detected at the two different locations. Activity classification is then performed on these detected segments using linear discriminant analysis (LDA) on the sound channel and hidden Markov models (HMMs) on the acceleration data. Four different methods at classifier fusion are compared for improving these classifications. Using user-dependent training, we obtain continuous average recall and precision rates (for positive activities) of 78 percent and 74 percent, respectively. Using user-independent training (leave-one-out across five users), we obtain recall rates of 66 percent and precision rates of 63 percent. In isolation, these activities were recognized with accuracies of 98 percent, 87 percent, and 95 percent for the user-dependent, user-independent, and user-adapted cases, respectively.     

The challenges of wearable computing: Part 2

For pt. 1 see ibid. Wearable computing pursues an interface ideal of a continuously worn, intelligent assistant that augments memory, intellect, creativity, communication, and physical senses and abilities. Many challenges await wearable designers. Part 2 begins with the challenges of network resources and privacy concerns. This survey describes the possibilities offered by wearable systems and, in doing so, demonstrates attributes unique to this class of computing     

Automated optics inspection analysis for NIF

The National Ignition Facility (NIF) is a high-energy laser facility comprised of 192 beamlines that house thousands of optics. These optics guide, amplify and tightly focus light onto a tiny target for fusion ignition research and high energy density physics experiments. The condition of these optics is key to the economic, efficient and maximally energetic performance of the laser. Our goal, and novel achievement, is to find on the optics any imperfections while they are tens of microns in size, track them through time to see if they grow and if so, remove the optic and repair the single site so the entire optic can then be re-installed for further use on the laser. This paper gives an overview of the image analysis used for detecting, measuring, and tracking sites of interest on an optic while it is installed on the beamline via in situ inspection and after it has been removed for maintenance. In this way, the condition of each optic is monitored throughout the optic's lifetime.This overview paper will summarize key algorithms and technical developments for custom image analysis and processing and highlight recent improvements. (Associated papers will include more details on these issues.) We will also discuss the use of OI Analysis for daily operation of the NIF laser and its extension to inspection of NIF targets.     

Use of sediment elemental and isotopic compositions to record the eutrophication of a polymictic reservoir in central Texas, USA

Paleolimnological studies are rarely performed on reservoirs because of concern that sediments might not accurately chronicle reservoir history. Eutrophication indicators might behave differently in polymictic reservoirs and stratified natural lakes because of system and/or mixing regime differences. Particulate organic carbon (POC), particulate organic nitrogen (PON), and total phosphorus (TP) concentrations, carbon:nitrogen (C:N) and nitrogen:phosphorus (N:P) ratios, and carbon (δ¹³C) and nitrogen (δ¹⁵N) stable isotopes from a sediment core were measured to demonstrate that sufficient information can be derived from sediments to permit a historical reconstruction. The scattered POC data were likely biased by seasonal/annual variability in allochthonous organic matter (OM) loading. The upwardly increasing PON in the sediment core supported historic primary productivity (PP) data, suggesting PON could be a better PP indicator than POC. The upwardly increasing TP documented historic P enrichment. The upwardly decreasing C:N ratio identified an OM source shift from allochthonous to increasingly autochthonous sources with reservoir age. The upwardly increasing N:P ratio implied that N‐fixation rates have increased with reservoir age, to compensate for increasing N limitation as the P loading increased. The δ¹³C decreased as the PP increased with reservoir age producing an atypical relationship compared to stratified natural lakes. The OM source shifts likely biased the δ¹³C–PP relationship, and might weaken δ¹³C‐inferred PP reconstructions in similar reservoirs. The δ¹⁵N increased with reservoir age, likely resulting from dissolved inorganic N (DIN) source changes, rather than nitrate utilization. Watershed urban growth and dairy operation intensification potentially contributed greater loads of isotopically heavy DIN. This study demonstrated that paleolimnology has great potential to assist eutrophication assessment and management efforts in reservoirs.