Gisele Bennett

Reality testing: HCI challenges in non-traditional environments

Non-traditional environments often change rapidly without forewarning, are difficult or impossible to control, and have other environmental and operational constraints that cannot easily be modeled in a laboratory, partly because the necessary level of ecological validity is almost impossible to achieve in the artificial lab environment. Current in situ field study evaluation techniques are insufficient in these environments. Furthermore, it is often difficult or impossible to ascertain which behavioral data are needed to answer questions about user requirements, interface design, and user acceptance. In this workshop, we will use case studies to create and explore frameworks for future non-traditional field study evaluations.

Tag testing methodology for RFID enabled temperature tracking and shelf life estimation

Recent advances in sensory devices using radio frequency identification (RFID) led to applications such as monitoring the temperature during the transportation of heat sensitive products where recorded data can be used to detect refrigeration equipment failure along the supply chain or estimate remaining shelf life of the product. For the project discussed in this paper, a handheld based portable RFID system is used to track the storage and transportation temperatures of perishable products using battery assisted passive temperature tags. The information from the tags is used in shelf life prediction models to estimate the remaining shelf life based on the recorded temperature data to provide a dynamic expiration date. Instead of the full application development effort, this paper focuses on the unique project requirements and challenges which led to the introduction of three novel concepts related to RFID enabled temperature tracking systems. First, due to absence of a common standard for testing RFID temperature tags, we develop a requirement driven, comprehensive testing protocol combining statistical tools and common industry standards with the help of a uniquely designed test setup to realistically simulate and evaluate the real life performances of different temperature tags. Next, a novel context based accuracy metric is derived for objective and application (such as shelf life prediction) specific comparison of different technologies. Finally, a pallet temperature estimation algorithm is developed to overcome some of the physical difficulties encountered in reading ultra-high frequency tags near the presence of metals and liquids.

Multiple Objective Vector-Based Genetic Programming Using Human-Derived Primitives

Traditional genetic programming only supports the use of arithmetic and logical operators on scalar features. The GTMOEP (Georgia Tech Multiple Objective Evolutionary Programming) framework builds upon this by also handling feature vectors, allowing the use of signal processing and machine learning functions as primitives, in addition to the more conventional operators. GTMOEP is a novel method for automated, data-driven algorithm creation, capable of outperforming human derived solutions. As an example, GTMOEP was applied to the problem of predicting how long an emergency responder can remain in a hazmat suit before the effects of heat stress cause the user to become unsafe. An existing third-party physics model was leveraged for predicting core temperature from various situational parameters. However, a sustained high heart rate also means that a user is unsafe. To improve performance, GTMOEP was evaluated to predict an expected pull time, computed from both thresholds during human trials. GTMOEP produced dominant solutions in multiple objective space to the performance of predictions made by the physics model alone, resulting in a safer algorithm for emergency responders to determine operating times in harsh environments. The program generated by GTMOEP will be deployed to a mobile application for their use.

RFID testing and evaluation for an RF-harsh environment

Radio Frequency Identification (RFID) has been proposed as the solution to the real-time asset visibility problem in numerous supply chain, health-care, and manufacturing applications. Some of those applications occur in RF-harsh environments which degrade the performance of an RFID system. Such environments can create a mismatch between the anticipated performance and the actual performance of RFID systems. In this paper, aspects of the geometry and RF phenomena commonly found in manufacturing plants are simulated in a laboratory to evaluate the robustness of UHF RFID technology for a manufacturing factory environment. Preliminary results show that multipath effects and inconsistencies in performance across tag and reader models continue to impede the successful operation of RFID technology in manufacturing environments. Furthermore, these obstacles can be navigated only with very careful equipment selection and environment characterization. The systems-based RFID testing and evaluation methods in this paper serve to emulate some of the facets of an RF-harsh environment in order to pinpoint what recommendations can be made to improve the effectiveness of future RFID implementations in the manufacturing world.

Performance testing of RFID systems with RF-harsh materials

Radio Frequency Identification (RFID) has been adopted to track items in supply chain, healthcare, and manufacturing applications. Hospitals and factories, however, are difficult environments for radiowave propagation. Cinder block walls with steel rebar, metal obstructions, and RF noise present significant obstacles to RFID system performance. Tagging lossy materials in these environments, such as metals and liquids, can also degrade the performance of RFID systems. In a previous paper [1] we simulated the RF-harsh conditions prevalent in these environments to evaluate UHF RFID system performance. In this paper, we utilize the same laboratory environment to measure RFID system performance when RF-harsh materials are tagged. These tests serve to examine the effect of water and plastic car parts on RFID system performance in an RF harsh environment. We show that the problems posed when tagging RF-harsh materials can be mitigated with either the strategic placement of tags on the item, or the careful choice of tags. While UHF RFID systems can be used in the presence of RF-harsh circumstances, the system architecture must be carefully tested in order to minimize the effects of performance-hindering RF obstacles.

Imaging systems and signal recovery: introduction to feature issue

Imaging systems and signal recovery algorithms have a wide range of applications, from industrial to military, and from consumer electronics to medical instrumentations. Advancement in this field involves both basic science and applied multidisciplinary engineering capabilities. This issue brings together academic researchers and practitioners to show how different imaging hardwares and signal processing techniques combine to determine the overall system performance; together, they underscore an important trend in a closer integration between algorithms and systems design.

Improved night vision demonstrator program status

Although existing night vision equipment provides a significant improvement in target detection in low light conditions, there are several limitations that limit their effectiveness. Focus is a significant problem for night vision equipment due to the low f-number optics required to obtain sufficient sensitivity as well as the dynamic nature of night vision applications, which requires frequent focus adjustments. The Georgia Tech Research Institute has developed a prototype next-generation night vision device called the Improved Night Vision Demonstrator (INVD) in order to address these shortfalls. This paper will describe the design of the INVD system as well as an analysis of its performance.

The Future of Manufacturing Utilizing Tagging Technology

Lean, agile, just-in-time, future, green, energy efficient and robust are just some of the terms used to describe the next generation manufacturing. The focus for manufacturing is dependent on the year, the challenges, the feature technologies, the consumer demands, and business pressures. This paper discusses the potential of the future of manufacturing, which utilizes technologies that allow for real time visibility of location and condition of assets for use in decision making to anticipate the manufacturing process and meet demands.

Imaging systems based on the encoding of optical coherence functions

An imaging scheme is described that is based on the transmission of image-forming information encoded within optical coherence functions. The scheme makes use of dynamic random-valued encoding-decoding masks placed in the input-output planes of any linear optical system. The mask transmittance functions are complex conjugates of each other, as opposed to a similar coherence encoding scheme proposed earlier by two of this papers authors that used identical masks. [Rhodes and Welch, in Euro-American Workshop on Optoelectronic Information Processing, SPIE Critical Review Series (SPIE, 1999), Vol. CR74, p. 1]. General analyses of the two coherence encoding schemes are performed by using the more general mutual coherence function as opposed to the mutual intensity function used in the earlier scheme. The capabilities and limitations of both encoding schemes are discussed by using simple examples that combine the encoding-decoding masks with free-space propagation, passage through a four-f system, and a single-lens imaging system.

Comparison of Measurement Techniques Used to Determine Atmospheric Structure Parameter

Cn2 measurements were taken at the recent CABLE/TRAX test at the NASA Shuttle Landing Facility along a 1.5 km horizontal path by three atmospheric turbulence measurement devices (DELTA, PROPS, IACS) and two commercial-grade scinitillometers (BLS-900 and K&Z LAS). The fundamental operating principles of these devices and their measurement data are analyzed for consistency in measurement outcomes.