Structure learning of graphical models for task-oriented robot grasping

In the collective imaginaries a robot is a human like machine as any androids in science fiction. However the type of robots that you will encounter most frequently are machinery that do work that is too dangerous, boring or onerous. Most of the robots in the world are of this type. They can be found in auto, medical, manufacturing and space industries. Therefore a robot is a system that contains sensors, control systems, manipulators, power supplies and software all working together to perform a task. The development and use of such a system is an active area of research and one of the main problems is the development of interaction skills with the surrounding environment, which include the ability to grasp objects. To perform this task the robot needs to sense the environment and acquire the object informations, physical attributes that may influence a grasp. Humans can solve this grasping problem easily due to their past experiences, that is why many researchers are approaching it from a machine learning perspective finding grasp of an object using information of already known objects. But humans can select the best grasp amongst a vast repertoire not only considering the physical attributes of the object to grasp but even to obtain a certain effect. This is why in our case the study in the area of robot manipulation is focused on grasping and integrating symbolic tasks with data gained through sensors. The learning model is based on Bayesian Network to encode the statistical dependencies between the data collected by the sensors and the symbolic task. This data representation has several advantages. It allows to take into account the uncertainty of the real world, allowing to deal with sensor noise, encodes notion of causality and provides an unified network for learning. Since the network is actually implemented and based on the human expert knowledge, it is very interesting to implement an automated method to learn the structure as in the future more tasks and object features can be introduced and a complex network design based only on human expert knowledge can become unreliable. Since structure learning algorithms presents some weaknesses, the goal of this thesis is to analyze real data used in the network modeled by the human expert, implement a feasible structure learning approach and compare the results with the network designed by the expert in order to possibly enhance it.

Algorithms for spacecraft formation flying navigation based on wireless positioning system measurements

Spacecraft formation flying navigation continues to receive a great deal of interest. The research presented in this dissertation focuses on developing methods for estimating spacecraft absolute and relative positions, assuming measurements of only relative positions using wireless sensors. The implementation of the extended Kalman filter to the spacecraft formation navigation problem results in high estimation errors and instabilities in state estimation at times. This is due tp the high nonlinearities in the system dynamic model. Several approaches are attempted in this dissertation aiming at increasing the estimation stability and improving the estimation accuracy. A differential geometric filter is implemented for spacecraft positions estimation. The differential geometric filter avoids the linearization step (which is always carried out in the extended Kalman filter) through a mathematical transformation that converts the nonlinear system into a linear system. A linear estimator is designed in the linear domain, and then transformed back to the physical domain. This approach demonstrated better estimation stability for spacecraft formation positions estimation, as detailed in this dissertation. The constrained Kalman filter is also implemented for spacecraft formation flying absolute positions estimation. The orbital motion of a spacecraft is characterized by two range extrema (perigee and apogee). At the extremum, the rate of change of a spacecraft’s range vanishes. This motion constraint can be used to improve the position estimation accuracy. The application of the constrained Kalman filter at only two points in the orbit causes filter instability. Two variables are introduced into the constrained Kalman filter to maintain the stability and improve the estimation accuracy. An extended Kalman filter is implemented as a benchmark for comparison with the constrained Kalman filter. Simulation results show that the constrained Kalman filter provides better estimation accuracy as compared with the extended Kalman filter. A Weighted Measurement Fusion Kalman Filter (WMFKF) is proposed in this dissertation. In wireless localizing sensors, a measurement error is proportional to the distance of the signal travels and sensor noise. In this proposed Weighted Measurement Fusion Kalman Filter, the signal traveling time delay is not modeled; however, each measurement is weighted based on the measured signal travel distance. The obtained estimation performance is compared to the standard Kalman filter in two scenarios. The first scenario assumes using a wireless local positioning system in a GPS denied environment. The second scenario assumes the availability of both the wireless local positioning system and GPS measurements. The simulation results show that the WMFKF has similar accuracy performance as the standard Kalman Filter (KF) in the GPS denied environment. However, the WMFKF maintains the position estimation error within its expected error boundary when the WLPS detection range limit is above 30km. In addition, the WMFKF has a better accuracy and stability performance when GPS is available. Also, the computational cost analysis shows that the WMFKF has less computational cost than the standard KF, and the WMFKF has higher ellipsoid error probable percentage than the standard Measurement Fusion method. A method to determine the relative attitudes between three spacecraft is developed. The method requires four direction measurements between the three spacecraft. The simulation results and covariance analysis show that the method’s error falls within a three sigma boundary without exhibiting any singularity issues. A study of the accuracy of the proposed method with respect to the shape of the spacecraft formation is also presented.

Exploiting sensor redundancy for the calculation of fractional derivatives in the presence of noise

In this paper a new method for the calculation of the fractional expressions in the presence of sensor redundancy and noise, is presented. An algorithm, taking advantage of the signal characteristics and the sensor redundancy, is tuned and optimized through genetic algorithms. The results demonstrate the good performance for different types of expressions and distinct levels of noise.

Highly Sensitive Sensor Based On Intracavity Laser Absorption Spectroscopy By Means of Relative Intensity Noise

This thesis concerns with the main aspects of medical trace molecules detection by means of intracavity laser absorption spectroscopy (ICLAS), namely with the equirements for highly sensitive, highly selective, low price, and compact size sensor. A novel two modes semiconductor laser sensor is demonstrated. Its operation principle is based on the competition between these two modes. The sensor sensitivity is improved when the sample is placed inside the two modes laser cavity, and the competition between the two modes exists. The effects of the mode competition in ICLAS are discussed theoretically and experimentally. The sensor selectivity is enhanced using external cavity diode laser (ECDL) configuration, where the tuning range only depends on the external cavity configuration. In order to considerably reduce the sensor cost, relative intensity noise (RIN) is chosen for monitoring the intensity ratio of the two modes. RIN is found to be an excellent indicator for the two modes intensity ratio variations which strongly supports the sensor methodology. On the other hand, it has been found that, wavelength tuning has no effect on the RIN spectrum which is very beneficial for the proposed detection principle. In order to use the sensor for medical applications, the absorption line of an anesthetic sample, propofol, is measured. Propofol has been dissolved in various solvents. RIN has been chosen to monitor the sensor response. From the measured spectra, the sensor sensitivity enhancement factor is found to be of the order of 10^(3) times of the conventional laser spectroscopy.

Application of noise analysis for sensor surveillance in Trillo NPP

It is a known fact that noise analysis is a suitable method for sensor performance surveillance. In particular, controlling the response time of a sensor is an efficient way to anticipate failures and to have the opportunity to prevent them. In this work the response times of several sensors of Trillo NPP are estimated by means of noise analysis. The procedure applied consists of modeling each sensor with autoregressive methods and getting the searched parameter by analyzing the response of the model when a ramp is simulated as the input signal. Core exit thermocouples and in core self-powered neutron detectors are the main sensors analyzed but other plant sensors are studied as well. Since several measurement campaigns have been carried out, it has been also possible to analyze the evolution of the estimated parameters during more than one fuel cycle. Some sensitivity studies for the sample frequency of the signals and its influence on the response time are also included. Calculations and analysis have been done in the frame of a collaboration agreement between Trillo NPP operator (CNAT) and the School of Mines of Madrid.

A scatterometer neural network sensor model with input noise

The ERS-1 satellite carries a scatterometer which measures the amount of radiation scattered back toward the satellite by the ocean's surface. These measurements can be used to infer wind vectors. The implementation of a neural network based forward model which maps wind vectors to radar backscatter is addressed. Input noise cannot be neglected. To account for this noise, a Bayesian framework is adopted. However, Markov Chain Monte Carlo sampling is too computationally expensive. Instead, gradient information is used with a non-linear optimisation algorithm to find the maximum em a posteriori probability values of the unknown variables. The resulting models are shown to compare well with the current operational model when visualised in the target space.

A scatterometer neural network sensor model with input noise

The ERS-1 satellite carries a scatterometer which measures the amount of radiation scattered back toward the satellite by the ocean's surface. These measurements can be used to infer wind vectors. The implementation of a neural network based forward model which maps wind vectors to radar backscatter is addressed. Input noise cannot be neglected. To account for this noise, a Bayesian framework is adopted. However, Markov Chain Monte Carlo sampling is too computationally expensive. Instead, gradient information is used with a non-linear optimisation algorithm to find the maximum em a posteriori probability values of the unknown variables. The resulting models are shown to compare well with the current operational model when visualised in the target space.

Noise resilient discrete-time cross-relation sensor characterisation

In-situ characterisation of thermocouple sensors is a challenging problem. Recently the authors presented a blind characterisation technique based on the cross-relation method of blind identification. The method allows in-situ identification of two thermocouple probes, each with a different dynamic response, using only sampled sensor measurement data. While the technique offers certain advantages over alternative methods, including low estimation variance and the ability to compensate for noise induced bias, the robustness of the method is limited by the multimodal nature of the cost function. In this paper, a normalisation term is proposed which improves the convexity of
the cost function. Further, a normalisation and bias compensation hybrid approach is presented that exploits the advantages of both normalisation and bias compensation. It is found that the optimum of the hybrid cost function is less biased and more stable than when only normalisation is applied. All results were verified by simulation.

Quantum noise in optical fibers. I. Stochastic equations

We analyze the quantum dynamics of radiation propagating in a single-mode optical fiber with dispersion, nonlinearity, and Raman coupling to thermal phonons. We start from a fundamental Hamiltonian that includes the principal known nonlinear effects and quantum-noise sources, including linear gain and loss. Both Markovian and frequency-dependent, non-Markovian reservoirs are treated. This treatment allows quantum Langevin equations, which have a classical form except for additional quantum-noise terms, to be calculated. In practical calculations, it is more useful to transform to Wigner or 1P quasi-probability operator representations. These transformations result in stochastic equations that can be analyzed by use of perturbation theory or exact numerical techniques. The results have applications to fiber-optics communications, networking, and sensor technology.

Sensor Integration for Smart Cities Using Multi-Hop Networks

Smart Cities are designed to be living systems and turn urban dwellers life more comfortable and interactive by keeping them aware of what surrounds them, while leaving a greener footprint. The Future Cities Project [1] aims to create infrastructures for research in smart cities including a vehicular network, the BusNet, and an environmental sensor platform, the Urban Sense. Vehicles within the BusNet are equipped with On Board Units (OBUs) that offer free Wi-Fi to passengers and devices near the street. The Urban Sense platform is composed by a set of Data Collection Units (DCUs) that include a set of sensors measuring environmental parameters such as air pollution, meteorology and noise. The Urban Sense platform is expanding and receptive to add new sensors to the platform. The parnership with companies like TNL were made and the need to monitor garbage street containers emerged as air pollution prevention. If refuse collection companies know prior to the refuse collection which route is the best to collect the maximum amount of garbage with the shortest path, they can reduce costs and pollution levels are lower, leaving behind a greener footprint. This dissertation work arises in the need to monitor the garbage street containers and integrate these sensors into an Urban Sense DCU. Due to the remote locations of the garbage street containers, a network extension to the vehicular network had to be created. This dissertation work also focus on the Multi-hop network designed to extend the vehicular network coverage area to the remote garbage street containers. In locations where garbage street containers have access to the vehicular network, Roadside Units (RSUs) or Access Points (APs), the Multi-hop network serves has a redundant path to send the data collected from DCUs to the Urban Sense cloud database. To plan this highly dynamic network, the Wi-Fi Planner Tool was developed. This tool allowed taking measurements on the field that led to an optimized location of the Multi-hop network nodes with the use of radio propagation models. This tool also allowed rendering a temperature-map style overlay for Google Earth [2] application. For the DCU for garbage street containers the parner company provided the access to a HUB (device that communicates with the sensor inside the garbage containers). The Future Cities use the Raspberry pi as a platform for the DCUs. To collect the data from the HUB a RS485 to RS232 converter was used at the physical level and the Modbus protocol at the application level. To determine the location and status of the vehicles whinin the vehicular network a TCP Server was developed. This application was developed for the OBUs providing the vehicle Global Positioning System (GPS) location as well as information of when the vehicle is stopped, moving, on idle or even its slope. To implement the Multi-hop network on the field some scripts were developed such as pingLED and “shark”. These scripts helped upon node deployment on the field as well as to perform all the tests on the network. Two setups were implemented on the field, an urban setup was implemented for a Multi-hop network coverage survey and a sub-urban setup was implemented to test the Multi-hop network routing protocols, Optimized Link State Routing Protocol (OLSR) and Babel.

Transimpedance amplifier for integrated SpO2 optic sensor

Dissertação para obtenção do Grau de Mestre em Engenharia Electrotécnica e de Computadores

Design of a low-voltage CMOS RF receiver for energy harvesting sensor node

In this thesis a CMOS low-power and low-voltage RF receiver front-end is presented. The main objective is to design this RF receiver so that it can be powered by a piezoelectric energy harvesting power source, included in a Wireless Sensor Node application. For this type of applications the major requirements are: the low-power and low-voltage operation, the reduced area and cost and the simplicity of the architecture. The system key blocks are the LNA and the mixer, which are studied and optimized with greater detail, achieving a good linearity, a wideband operation and a reduced introduction of noise. A wideband balun LNA with noise and distortion cancelling is designed to work at a 0.6 V supply voltage, in conjunction with a double-balanced passive mixer and subsequent TIA block. The passive mixer operates in current mode, allowing a minimal introduction of voltage noise and a good linearity. The receiver analog front-end has a total voltage conversion gain of 31.5 dB, a 0.1 - 4.3 GHz bandwidth, an IIP3 value of -1.35 dBm, and a noise figure lower than 9 dB. The total power consumption is 1.9 mW and the die area is 305x134.5 m2, using a standard 130 nm CMOS technology.

Low noise pixel detectors based on gated geiger mode avalanche photodiodes

The gated operation is proposed as an effective method to reduce the noise in pixel detectors based on Geiger mode avalanche photodiodes. A prototype with the sensor and the front-end electronics monolithically integrated has been fabricated with a conventional HV-CMOS process. Experimental results demonstrate the increase of the dynamic range of the sensor by applying this technique.

Avoiding sensor blindness in Geiger mode avalanche photodiode arrays fabricated in a conventional CMOS process

The need to move forward in the knowledge of the subatomic world has stimulated the development of new particle colliders. However, the objectives of the next generation of colliders sets unprecedented challenges to the detector performance. The purpose of this contribution is to present a bidimensional array based on avalanche photodiodes operated in the Geiger mode to track high energy particles in future linear colliders. The bidimensional array can function in a gated mode to reduce the probability to detect noise counts interfering with real events. Low reverse overvoltages are used to lessen the dark count rate. Experimental results demonstrate that the prototype fabricated with a standard HV-CMOS process presents an increased efficiency and avoids sensor blindness by applying the proposed techniques.

Gated Geiger mode avalanche photodiode pixels with integrated readout electronics for low noise photon detection

Avalanche photodiodes operated in the Geiger mode offer a high intrinsic gain as well as an excellent timing accuracy. These qualities make the sensor specially suitable for those applications where detectors with high sensitivity and low timing uncertainty are required. Moreover, they are compatible with standard CMOS technologies, allowing sensor and front-end electronics integration within the pixel cell. However, the sensor suffers from high levels of intrinsic noise, which may lead to erroneous results and limit the range of detectable signals. They also increase the amount of data that has to be stored. In this work, we present a pixel based on a Geiger-mode avalanche photodiode operated in the gated mode to reduce the probability to detect noise counts interfering with photon arrival events. The readout circuit is based on a two grounds scheme to enable low reverse bias overvoltages and consequently lessen the dark count rate. Experimental characterization of the fabricated pixel with the HV-AMS 0.35µm standard technology is also presented in this article.