Luis G. Sison

Characterization and adaptive filtering of motion artifacts in pulse oximetry using accelerometers

Noise, in the form of motion artifact, often leads to false information and acts as a limiting factor in the analysis of pulse oximetric signals. We propose an adaptive filter for filtering motion artifacts from pulse oximetry signals, with accelerometer signals as noise references. We study two adaptive filtering schemes: (1) single-axis and (2) dual-axes stress tests; and apply both the LMS and RLS algorithms to each scheme to compare their effectiveness. Results show that a single-axis adaptive filter employing the RLS algorithm (N=32 and /spl lambda/=0.9999) is adequate to minimize motion artifact.

Adaptive noise cancelling of motion artifact in stress ECG signals using accelerometer

Electrocardiographic (ECG) signals obtained from stress examinations are diagnostically significant in detecting a number of heart diseases, which may not be apparent when the patient is at rest. However, the noise produced by the environment and by the patient often distorts the ECG data. Motion artifact, the most prevalent and difficult type of noise to filter in exercise ECG, corrupts the intelligibility of the desired signal thus reducing the reliability of the stress test. In this paper, the researchers aim to demonstrate a new adaptive filtering method for stress ECG signals. This noise cancellation scheme uses an accelerometer as a source of noise reference. Experiments involving single-axis and dual-axis motion sensors are conducted to evaluate the efficiency of this technique. The acquired real ECG and accelerometer data are simultaneously processed and analyzed using the two most widely used adaptive filtering algorithms, least mean squares (LMS) and recursive least squares (RLS). The results show that the proposed method can be adapted to effectively reduce motion artifact in stress ECG by just using a single-axis noise reference.

MotesArt: Wireless Sensor Network for Monitoring Relative Humidity and Temperature in an Art Gallery

Artworks in museums and art galleries are significantly affected by temperature and relative humidity (RH) variations. Significant variations in temperature can double the deterioration rate of paper and other canvases. Fluctuating levels of RH cause materials to corrode, shrink, swell, or warp. Monitoring and control of indoor climate conditions is thus crucial in ensuring the preservation of art collections. This study addresses the need for monitoring by designing and implementing a wireless sensor network composed of several specialized sensor nodes dumping temperature and RH readings to a base station computer. Each sensor node is equipped with an ASK transmitter and no ASK receiver. This design allows for lower power consumption and highly simplifies communication between the sensor node and base station. The MAC protocol used for the wireless sensor network is ALOHA without retransmissions. The wireless sensor network is evaluated in terms of cost, range, power consumption and packet error rate.

Fuzzy modeling by induction and pruning of decision trees

Presents a Fuzzy ID3 algorithm that generates a fuzzy rule base from a set of input-output data. The rules are induced by the ID3 algorithm of Quinlan (1986). Pruning conditions are presented that assist in the elimination of irrelevant attributes and simplification of the rule base. The algorithm is applied to the modeling of an ARMA process with both known and unknown model order.<<ETX>>

No-reset iterative learning control

In this paper, we study the repetitive control of discrete-time linear periodically time-varying (LPTV) plants using a no-reset iterative learning control (NRILC) system, which is an iterative learning control (ILC) system where the plant is not reset at the beginning of each iteration. We apply this scheme to discrete-time LTI SISO plants; the approach can be extended to LPTV and MIMO plants. Using results from output feedback theory, we show that the closed-loop eigenvalues of the system can almost always be placed with the selection of an appropriate finite learning gain.

Energy-saving Flyback Converter for EDM Applications

Electrical discharge machining (EDM) uses controlled electric sparks to erode the metal in a workpiece, and this method is now a well established machining process for high power applications. EDM power supplies utilizing different topologies for high power applications (e.g. full bridge configuration) are widely available. Through the years, EDM processes have increasingly been used in high precision machining and in manufacturing micromechanical components, with new low power topologies being investigated for the latter. We implemented a current mode flyback converter for use in low power EDM applications. This topology has an energy conservation feature and removes the need for output bulk capacitors. Energy used in the erosion process will come from the energy stored in the flyback transformer and will be transferred to the small output filter capacitors and the effective capacitance of the tool, workpiece, and dielectric combination. A servomechanism that controls the arc distance will determine the output voltage. We have developed an EDM power supply prototype utilizing the flyback topology for low power applications with minimal components count and inherent protection under short circuit conditions. Our work includes the design, fabrication, and characterization of the whole system. We compared the efficiency and quality of work of our proposed topology with that of a linear converter.

ALESSA: MDS - based localization algorithm for Wireless Sensor Networks

Self - localization in Wireless Sensor Networks (WSN) should be precise and reliable. Alternative Least-Square Scaling Algorithm (ALESSA) is a recently proposed centralized Multidimensional Scaling (MDS)-based localization algorithm, which uses an iterative approach to solve for the coordinates of discrete points. While ALESSA converges most of the time, like most iterative algorithm, it can be trapped in local minima causing large errors in the location estimates. In this paper, we propose the reseeding of the initial random estimates to improve the convergence of the algorithm. Performance of the proposed algorithm is evaluated under different network topologies with limited connectivity. We also analyzed the effects of low Signal to Noise Ratio (SNR) and the population of the nodes deployed in the network to the algorithms localization precision. Simulation results show that at 26 dB SNR reseeding always results in convergence with the estimation errors within 5 % of the reference communication range. Analysis and test runs also verified that our algorithm provides accurate and consistent localization estimates under range-based localization with limited network connectivity, even with a low SNR. The algorithm also performs well with limited number of nodes.

Switching performance characterization of SiC Schottky diodes in switch-mode DC-DC converters

The SiC Schottky diode, with its high blocking voltage capability, absence of reverse recovery, and negligible temperature dependence on switching behavior, is suitable for most power electronics applications especially in circuits that demand high efficiency and high switching frequencies. Recent studies dealing with the comparison of SiC Schottky with ultra-fast diodes are application-specific and they do not quantify the contribution of the superior switching performance of the SiC Schottky to the improvement of the switch-mode converters efficiency. To address this deficiency, we compare the performances of the D12S60 SiC Schottky diode to leading ultra-fast diodes of comparable rating, such as the STTA1206D, in hard-switching buck and buck-boost switching converters. Our study includes the measurement of the diode and FET switching and conduction losses as well as converter efficiency while varying the input voltage, load current, and switching frequency.

Regenerative clamp as reset winding in flyback converters for EDM applications

Switch mode power supplies in electrical discharge machining (EDM) applications remove the need for bulk output capacitors. The switch mode converter topology investigated in this paper is the flyback converter. The operation of the flyback topology without bulk output capacitors requires resetting the flyback transformer. In this paper, the absence of the reset mechanism and two approaches to reset the flyback transformer are investigated. The traditional RC clamp will only be an effective reset mechanism if the resistor is reduced to a very small value which will lead to higher power losses and will not be suitable for this application. The application of a reset winding similar to forward converters may be utilized; however, the need for a turn off snubber is not eliminated. A regenerative clamp used in flyback converters will act as a reset winding and as a turn off snubber. The use of this configuration is proposed for EDM applications.

Comparison of classical and fuzzy control in active mass damping of a flexible structure using acceleration feedback

Most research on fuzzy control claim that fuzzy controllers outperformed classical controllers. In this study, classical and fuzzy controllers are compared in the control of a flexible structure employing an active mass damper system using acceleration feedback. Both controllers are optimized and evaluated using Matlab and implemented in real-time using real-time workshop. The parameters of both controllers are tuned to achieve minimum sum-of-squares of the top floor acceleration. The controllers are evaluated based on the rate of convergence and performance criterion.