Wearable eye tracking devices have broad uses in medicine, psychology, augmented & virtual reality and consumer market research. Most mobile eye trackers available today utilize infrared imaging of the pupil and corneal reflections with video cameras. This tracking method requires sophisticated real-time processing of video signals consuming substantial electrical power. This method is thus unsuitable for light weight wearables such as adaptive smart eyeglasses for correction of presbyopia. In this paper we present a low-profile, low-power (7.7 mJ/sample) digital eye tracker oculometer based on infrared sclera tracking. The system is implemented using eight, 24-bit infrared proximity sensors and synchronous infrared LEDs. The pupil location is determined from 32 reflected pulsed light measurements independent of ambient illumination. The digital oculometer is 3.1 mm thick and weighs ~3 g. The tracker mounts adjacent to the tunable lenses in the smart eyeglasses frame. The eye tracker showed a pointing error of 1.3 degrees rms over a vertical and horizontal range of 30 degrees when tested by an observer.

We are developing a personal micronavigation system that uses high-resolution gait-corrected inertial measurement units. The goal of this project is to develop a navigation system that use secondary inertial variables, such as velocity, to enable long-term precise navigation in the absence of Global Positioning System (GPS) and beacon signals. In this scheme, measured zero velocity durations from the ground reaction sensors are used to reset the accumulated integration errors from the accelerometers and gyroscopes in position calculation. We achieved an average position error of 4 meters at the end of half-hour walks.

Microfluidic devices are increasingly important in biomedical assays. In this paper, we review recent progress on these devices for chemical amplification, hybridization, separation, and detection of nucleic acids. Recent developments in microfluidic technologies for detection of protein binding events in immunoassay applications are also reviewed.

This paper presents a review of microfabricated devices for genetic diagnostics. Genetic diagnostics are powerful technology drivers and excellent candidate applications for miniaturization technologies because the demand for inexpensive genetic information is essentially unlimited, and the cost and time for the diagnostic decreases with sample volume. Genetic information is stored in long DNA molecules in solution. This information is processed and extracted using a series of enzymatic and other chemical reactions well known in molecular biology. Processing of DNA molecules in the microscale hence requires the implementation of microfluidic devices capable of handling, mixing, thermal cycling, separating, and detecting nano- and picoliter liquid samples. This paper discusses some of the fundamental macroscale protocols used for genetic analyses and how these processes scale down to microscopic volumes. The construction and performance of microfluidic devices of DNA amplification, separation, hybridization, and detection are discussed, showing that so far, no fundamental impediments exist for genetic diagnostics based on microelectromechanical systems. Some of the unresolved storage and packaging issues and future challenges for the practical implementation of these devices are also presented.

This paper describes a systematic method for the automatic generation of fabrication processes of thin film devices. The method uses a partially ordered set (poset) representation of device topology describing the order between its various components in the form of a directed acyclic graph. The sequence in which these components are fabricated is determined from the poset linear extensions, and the component sequence is expanded into a corresponding process flow. The graph-theoretic synthesis method is powerful enough to establish existence and multiplicity of flows thus creating a design space D suitable for optimization. The cardinality /spl par/D/spl par/ for a device with N components is large with a worst case /spl par/D/spl par//spl les/(N-1)! yielding in general a combinatorial explosion of solutions. The number of solutions is controlled through a priori estimates of /spl par/D/spl par/ and condensation of the device graph. The method has been implemented in the computer program MISTIC (Michigan Synthesis Tools for Integrated Circuits) which calculates specific process parameters using an internal database of process modules and materials. Currently, MISTIC includes process modules for deposition, lithography, etching, ion implantation, coupled simultaneous diffusions, and reactive growth. The compilation procedure was applied to several device structures. For a double metal twin-well BiCMOS structure, the compiler generated 168 complete process flows.