Augusto Garcia-Valenzuela

Sensing means and sensor shells: a new method of comparative study of piezoelectric, piezoresistive, electrostatic, magnetic, and optical sensors

Abstract The immense computational power of todays processors cannot be fully utilized unless appropriate sensors and actuators are developed to link them with the outside world. A method is presented here to compare different sensors with each other and to evaluate their relative merits. On the basis of physical considerations, sensors are decomposed into two essential building blocks: a sensor shell and a sensing mean. A figure of merit is defined and 11 different types of sensors, having the same sensing shells, are analyzed and compared with each other.

Comparative study of piezoelectric, piezoresistive, electrostatic, magnetic, and optical sensors

Immense computational power of todays processors cannot be fully utilized unless appropriate sensors and actuators are developed to link them with the outside world. A method is presented here to compare different sensors with each other and to evaluate their relative merits. Sensors are decomposed into two essential building blocks: a sensor shell and a sensing mean. Eleven different types of sensors, having the same sensing shells, are analyzed and compared to each other.

Fiber-optic force and displacement sensor based on speckle detection with 0.1 nN and 0.1 Å resolution

Abstract A new mode of operation for a high-multimode optical-fiber sensor based on the detection of modal interference is described. The main mechanism on which the sensor is based is coupling between different modes due to curvature induced upon bending the fiber. In contrast to previous work, the variation of the length and the average longitudinal strain of the fiber under such curvatures may be neglected in this work. A sensor of this kind has been implemented and shows a resolution of 10 −11 m in displacement and 10 −10 N in force. Such a high resolution in this type of sensor has not been reported previously. The ‘spring’ constant of the sensor is relatively small (5.6 N/m) compared to that of similar sensors. A theoretical analysis of the transduction mechanism is given and some guidelines to understanding its performance are discussed. The potential for this sensor to perform as the sensing element of an atomic force microscope is discussed.

Mode conversion and large-angle transmission in symmetric multimode Y-junction couplers

A multimode symmetric Y-junction waveguide is investigated and the optical power transmission and conversion efficiency between the lower and higher modes was calculated. Overlap integrals of the fields of the incident and transmitted guided-wave modes at the Y- junction were used and it is shown that there is a possibility of transmitting up to 29 percent of the incident optical power into the output branching waveguides at very large branching angles. Results show that an appropriate branching angle of the Y-junction can be used to selectively transmit the waveguide modes into the output branches and this can be used advantageously in integrated optic devices. A very simple model using ray optics is developed to explain the results on physical basis. With this simple model we developed a simple framework to interpret the results of the overlap integrals.

Large-angle transmission and mode conversion characteristics of symmetric multimode Y-junction couplers

We investigated the optical power transmission and conversion efficiency between the lower and higher modes in a multimode symmetric Y-junction waveguide. Using the overlap integrals of the fields of the incident and transmitted guided-wave modes at the Y junction, we obtained a first-order approximation of the transmission characteristics and found that there is a possibility of transmitting a useful amount of the incident optical power into the output branching waveguides at very large branching angles. Results showed that an appropriate branching angle of the Y junction can be used to selectively transmit the waveguide modes into the output branches. We developed a very simple model using ray optics to explain the results on a more intuitive basis.

Improvement of the theoretical minimum detectable angle of the optical beam deflection method

We study the possibility of improving the theoretical limit to the resolution of the optical beam deflection method (OBDM) by reflecting the beam from a curved surface. We suggest a new detection scheme by measuring the average intensity over a cross section of the beam after it is reflected from a cylindrical mirror. We show that there is a possibility of decreasing the theoretical minimum detectable angle (MDA) as compared to the usual OBDM when the beam grazes the cylindrical mirror.

Sensing Means and Sensor Shells

Given a sensing task, in general there are a variety of methods that can be used to accomplish it. These methods may incorporate charge sensing, current/voltage dc/ac measurements, magnetization, or optical measurements to achieve the desired sensing task. In addition to these schemes or sensing means, one should also decide on the structure of the sensor. The structure of the sensor, will clearly depend on the identity of the parameter that one wishes to sense (this is usually called measurand) and also, it will depend on the sensing mean. Thus, we separate the sensing shell, which is the structure of the sensor, from the sensing mean which is the physical scheme that is used to read the sensor shell. Clearly, sensor shell and sensing means are interdependent. Using this division of sensors to different parts, we are able to compare merits of different sensing means and sensor shells to each other.

Smart integrated-optics displacement/force sensor based on speckle pattern detection using neural net with 0.1Åresolution

Abstract The use of a multimode optical fiber in sensing displacements of the order of 0.1Ahas been reported. It has been shown that for very small movements induced at the center of the free-standing portion of the fiber, the behavior of the sensor is linear. The movement is caused by the vibration of an actuator that is in contact with the center of the free-standing portion of the fiber. The authors, however, caution that the expected nonlinearity of the sensor upon large vibrations restricts the general applicability in displacement sensing. In this paper, we use the same sensor assembly and show how a related nonlinear behavior can be taken care of by a neural network, pointing a way to obviate this limitation. Thus, the use of a neural network in an integrated-optics sensor for the first time has been a very useful start, showing that a sensor to measure displacements of 0.1Aresolution is now feasible. The speckle configuration depends not only on the amplitude of vibration of the actuator, but also on the mean position of the fiber. It is found that the output of a photodetector, placed so as to measure a small region of the speckle configuration, follows a nonlinear function as the mean position of the fiber is varied. We have trained a neural network to learn this nonlinear function. The neural network trained for displacement values within a 1.5 μm range is found to generalize with less than ± 0.015 μm error in the input range. It is found that a functional-link net with 11 functional links and 4 neurons can be trained for a convergence criterion of 5 × 10 −6 in less than 1000 iterations. The error in the individual targeted output during training is less than ±0.01 μm.

Smart integrated-optics displacement/force sensor based on speckle pattern detection using neural-net with 0.1-A resolution

This paper describes the usage of neural networks in the application of speckle patterns, as seen at the output of an optical waveguide, for sensing displacement/force. The neural network trained for displacement values within a 1.5 micrometers range is found to generalize with less than +/- 0.02 micrometers error in the input range. It is found that a functional-link net with 11 functional links and 3 neurons can be trained for a convergence criterion of 1e-05 in less than 3000 iterations. The error in the individual targeted output during training was less than +/- 0.01 micrometers . Thus an integrated-optics sensor is now feasible.