Sensors2014, 14(12), 24523-24542; doi:10.3390/s141224523 - published 19 December 2014 Show/Hide Abstract
Abstract: A new algorithm called Huber-based iterated divided difference filtering (HIDDF) is derived and applied to cooperative localization of autonomous underwater vehicles (AUVs) supported by a single surface leader. The position states are estimated using acoustic range measurements relative to the leader, in which some disadvantages such as weak observability, large initial error and contaminated measurements with outliers are inherent. By integrating both merits of iterated divided difference filtering (IDDF) and Huber’s M-estimation methodology, the new filtering method could not only achieve more accurate estimation and faster convergence contrast to standard divided difference filtering (DDF) in conditions of weak observability and large initial error, but also exhibit robustness with respect to outlier measurements, for which the standard IDDF would exhibit severe degradation in estimation accuracy. The correctness as well as validity of the algorithm is demonstrated through experiment results.
Sensors2014, 14(12), 24502-24522; doi:10.3390/s141224502 - published 19 December 2014 Show/Hide Abstract
Abstract: A new concept of a high-frequency amplitude detector and demodulator for Giant-Magneto-Impedance (GMI) sensors is presented. This concept combines a half wave rectifier, with outstanding capabilities and high speed, and a feedback approach that ensures the amplitude detection with easily adjustable gain. The developed detector is capable of measuring high-frequency and very low amplitude signals without the use of diode-based active rectifiers or analog multipliers. The performances of this detector are addressed throughout the paper. The full circuitry of the design is given, together with a comprehensive theoretical study of the concept and experimental validation. The detector has been used for the amplitude measurement of both single frequency and pulsed signals and for the demodulation of amplitude-modulated signals. It has also been successfully integrated in a GMI sensor prototype. Magnetic field and electrical current measurements in open- and closed-loop of this sensor have also been conducted.
Sensors2014, 14(12), 24483-24501; doi:10.3390/s141224483 - published 19 December 2014 Show/Hide Abstract
Abstract: The implementation of signal filters in a real-time form requires a tradeoff between computation resources and the system performance. Therefore, taking advantage of low lag response and the reduced consumption of resources, in this article, the Recursive Least Square (RLS) algorithm is used to filter a signal acquired from a fiber-optics-based sensor. In particular, a Long-Period Fiber Grating (LPFG) sensor is used to measure the bending movement of a finger. After that, the Gaussian Mixture Model (GMM) technique allows us to classify the corresponding finger position along the motion range. For these measures to help in the development of an autonomous robotic hand, the proposed technique can be straightforwardly implemented on real time platforms such as Field Programmable Gate Array (FPGA) or Digital Signal Processors (DSP). Different angle measurements of the finger’s motion are carried out by the prototype and a detailed analysis of the system performance is presented.
Sensors2014, 14(12), 24472-24482; doi:10.3390/s141224472 - published 19 December 2014 Show/Hide Abstract
Abstract: This work describes the fabrication, characterization, and application of a gold microband array electrode (MAE) for the determination of phosphate in fresh water samples. The working principle of this MAE is based on the reduction of a molybdophosphate complex using the linear sweep voltammetric (LSV) method. The calibration of this microsensor was performed with standard phosphate solutions prepared with KH2PO4 and pH adjusted to 1.0. The microsensor consists of a platinum counter electrode, a gold MAE as working electrode, and an Ag/AgCl electrode as reference electrode. The microelectrode chips were fabricated by the Micro Electro-Mechanical System (MEMS) technique. To improve the sensitivity, gold nanoparticles (AuNPs) were electrodeposited on the working electrode. With a linear range from 0.02 to 0.50 mg P/L, the sensitivity of the unmodified microsensor is 2.40 µA per (mg P/L) (R2 = 0.99) and that of the AuNPs-modified microsensor is 7.66 µA per (mg P/L) (R2 = 0.99). The experimental results showed that AuNPs-modified microelectrode had better sensitivity and a larger current response than the unmodified microelectrode.
Sensors2014, 14(12), 24462-24471; doi:10.3390/s141224462 - published 19 December 2014 Show/Hide Abstract
Abstract: Doppler sonographic measurement of flow velocity in the basal cerebral arteries through the intact skull was developed using a pulsed Doppler technique and 2 MHz emitting frequency. Relaxor-based ferroelectric single crystals Pb(Mg1/3Nb2/3)O3-PbTiO3 (PMN-PT) were chosen to be the piezoelectric transducer material due to their ultrahigh piezoelectric coefficients, high electromechanical coupling coefficients and low dielectric loss. The pulse-echo response of the transducer was measured using the conventional pulse-echo method in a water bath at room temperature. The −6 dB bandwidth of the transducer is 68.4% and the sensitivity is −17.4 dB. In order to get a good match between transducer and detection system, different transmission powers have been regulated by changing the impedance of the transmitting electric circuit. In the middle cerebral artery (MCA) measurement photograph results, as the transmission power is increasing, the detection results become clearer and clearer. A comparison at the same transmission power for different transducers shows that the detection photograph obtained by the crystal transducer was clearer than that obtained with a commercial transducer, which should make it easier for doctors to find the cerebral arteries.
Sensors2014, 14(12), 24441-24461; doi:10.3390/s141224441 - published 18 December 2014 Show/Hide Abstract
Abstract: Recent advances in wireless networking technology and the proliferation of industrial wireless sensors have led to an increasing interest in using wireless networks for closed loop control. The main advantages of Wireless Networked Control Systems (WNCSs) are the reconfigurability, easy commissioning and the possibility of installation in places where cabling is impossible. Despite these advantages, there are two main problems which must be considered for practical implementations of WNCSs. One problem is the sampling period constraint of industrial wireless sensors. This problem is related to the energy cost of the wireless transmission, since the power supply is limited, which precludes the use of these sensors in several closed-loop controls. The other technological concern in WNCS is the energy efficiency of the devices. As the sensors are powered by batteries, the lowest possible consumption is required to extend battery lifetime. As a result, there is a compromise between the sensor sampling period, the sensor battery lifetime and the required control performance for the WNCS. This paper develops a model-based soft sensor to overcome these problems and enable practical implementations of WNCSs. The goal of the soft sensor is generating virtual data allowing an actuation on the process faster than the maximum sampling period available for the wireless sensor. Experimental results have shown the soft sensor is a solution to the sampling period constraint problem of wireless sensors in control applications, enabling the application of industrial wireless sensors in WNCSs. Additionally, our results demonstrated the soft sensor potential for implementing energy efficient WNCS through the battery saving of industrial wireless sensors.