No description available.

This study investigates if a smartphones orientation in the pocket affects the result of a decision tree model trained with data from personal falls, and also how a low-pass filter affects these results. A comparison is made between the results gathered from this study, compared to previous studies and products within the field. The data was gathered using a smartphone application and was later split up to get datasets for all the different orientations of the smartphone. Before training the models, the data was processed through a low pass filter. Results showed that low pass filtered signals generally performed better and that two of the trained models, could outscore at least one other algorithm cited in this thesis in at least one category. However, existing products on the market that were investigated do not disclose their statistics and a comparison to these products could not be made. The best two orientations for the phone to be placed in the pocket was when the face of the phone was pointing out from the leg, and top of the phone was pointing up and also when the face of the phone was pointing out from the leg, and the top of the phone was pointing down.

Underwater communication within underwater sensor network is crucial for surveillance of coast and ocean areas. The aim of this report was to examine whether it is realistic to use the communication signal which is sent from one node to another as a sonar pulse, and in such case at what distances. To examine the problem, a system consisting of two nodes and one approaching target was simulated in Matlab. At first, the system tries to detect the target by using a generalized likelihood ratio test, which calculates the probability of a present target from the surrounding sounds. When a target is detected by a node, it estimates the bearing to the target by using beamforming and sends out a communication signal to the other node. The communication signal spreads out in the water, and bounces on the target before it is received by the second node. To calculate the distance, the second node decodes the signal to get the time difference, from which the distance is calculated. In the end, the target's position was estimated with a weighted least square estimator with measurements of the bearing and distance. The result shows that the distance to the target could be estimated with high precision in the given scenario, and that the width of the Cramér-Rao lower bound depends mainly on the variance of the beamforming algorithm. The maximum distance reached up to two kilometers but was mainly restricted by the detection algorithm. In conclusion, the result shows that the communication pulse can be used as a sonar pulse at the tested distances. However, the simulated scenario is a simplified version of the real world so more testing should be performed before a final conclusion can be made. / För övervakning av kust- och havsområden, vid exempelvis militära operationer eller för oceanografska observationer, används ofta ett undervattenssystem som är uppbyggt av flera noder som finns utplacerade på botten. Noderna lyssnar efter mål såsom ubåtar, fartyg etc, med syftet att kunna detektera och lokalisera dessa. Om en nod lyckas detektera ett mål så skickar den ut en akustisk kommunikationssignal till övriga noder i systemet. Målet med detta examensarbete var att undersöka om den kommunikationssignal som skickas mellan noderna också kan användas som en sonarpuls för att bestämma avståndet till målet, och därmed förbättra lokaliseringen av målets position. Under antagandet att kommunikationssignalen kan användas som sonarpuls, så undersöktes dessutom vid vilka avstånd mellan noden och målet som det var möjligt att använda signalen som sonarpuls. Resultatet visar att det är möjligt att använda kommunikationssignalen som en sonarpuls. Bäst funkar det på nära avstånd, då är den estimerade positionen i stort sett lika med det riktiga positionen. I takt med att avståndet till målet ökar så ökar även osäkerheten i vilken rikting målet befinner sig, estimeringen av avståndet höll sig däremot väldigt nära den faktiska distansen i alla simuleringar som gjordes. Simuleringen som gjordes var dock en förenkling av verkligheten, och flera av de störningsmoment som finns ute i naturen har inte tagits med i beräkningarna. För att undersöka detta så simulerades ett sensorsystem bestående av två noder tillsammans med ett mål som närmade sig noderna. Noderna försöker detektera målet genom att lyssna efter ljud som tillhör målet. Genom att mäta energinivåer i de ljudsignaler som noderna hör, så kan man utifrån sannolikhetslära bestämma hur troligt det är att det finns ett mål i närheten. När sannolikheten är tillräckligt hög säger man att ett mål detekterats. För att bestämma positionen så uppskattades målets riktning och avstånd i förhållande till noderna, som i sin tur användes för att beräkna målets position.

A Wireless Sensor Network (WSN) is a dense network of autonomous devices (or motes) with sensors that cooperatively monitor some physical or environmental conditions. These devices are resource constrained -limited memory, power and computational resources. Thus, any algorithm developed for WSN should be deigned such that the algorithm consumes the resources as minimal as possible. The problem addressed in this thesis is developing a low-complexity algorithm for intrusion detection in the presence of clutter arising from moving vegetation, using Passive Infra-Red (PIR) sensors. The algorithm is based on a combination of Haar Transform (HT) and Support-Vector-Machine (SVM) based training. The spectral signature of the waveforms is used to separate between the intruder and clutter waveforms. The spectral signature is computed using HT and this is fed to SVM which returns an optimal hyperplane that separates the intruder and clutter signatures. This hyperplane obtained by offline training is used online in the mote for surveillance. The algorithm is field-tested in the Indian Institute of Science campus. Based on experimental observations about the PIR sensor and the lens system, an analytical model for the waveform generated by an intruder moving along a straight line with uniform velocity in the vicinity of the sensor is developed. Analysis on how this model can be exploited to track the intruder path by optimally positioning multiple sensor nodes is provided. Algorithm for tracking the intruder path using features of the waveform from three sensors mounted on a single mote is also developed.

This thesis investigates the problem of detecting an intruder in the presence of clutter in a Passive Infra-Red (PIR) based Wireless Sensor Network (WSN). As one of the major objectives in a WSN is to maximize battery life, data transmission and local computations must be kept to a minimum as they are expensive in terms of energy. But, as intrusion being a rare event and cannot be missed, local computations expend more energy than data transmission. Hence, the need for a low-complexity algorithm for intrusion detection is inevitable. A low-complexity algorithm for intrusion detection in the presence of clutter arising from wind-blown vegetation, using PIR sensors is presented. The algorithm is based on a combination of Haar Transform (HT) and Support Vector Machine (SVM) based training. The amplitude and frequency of the intruder signature is used to differentiate it from the clutter signal. The HT was preferred to Discrete Fourier Transform (DFT) in computing the spectral signature because of its computational simplicity -just additions and subtractions suffice (scaling coefficients taken care appropriately). Intruder data collected in a laboratory and clutter data collected from various types of vegetation were fed into SVM for training. The optimal decision rule returned by SVM was then used to separate intruder from clutter. Simulation results along with some representative samples in which intrusions were detected and the clutter being rejected by the algorithm is presented. The implementation of the proposed intruder-detection algorithm in a network setting comprising of 20 sensing nodes is discussed. The field testing performance of the algorithm is then discussed. The limitations of the algorithm is also discussed. A closed-form analytical expression for the signature generated by a human moving along a straight line in the vicinity of the PIR sensor at constant velocity is provided. It is shown to be a good approximation by showing a close match with the real intruder waveforms. It is then shown how this expression can be exploited to track the intruder from the signatures of three well-positioned sensing nodes.

In this report I present a feasibility study of using open astronomical data to make Initial Orbit Determination (IOD) for Resident Space Objects (RSO) appearing as streaks in telescope images. The purpose is to contribute to Space Surveillance and Tracking (SST) for maintaining Space Situation Awareness (SSA). Data from different wide-field survey telescopes were considered but due to availability constraints only mask images from Zwicky Transient Facility (ZTF) survey were chosen for the analysis. An algorithm was developed to detect streaks in the mask images and match them to RSO known to be within the Field of View (FoV) at the observation time. Further, the IOD was made with angles-only Laplace’s method and the state vectors calculated for the streaks from the IOD were compared to those from the TLE for the matching RSO. The algorithm was tested with 6 different image fields acquired between the 14th to the 16th December 2019, of which 4 are characterised as non-crowded and 2 as crowded. The streak finding algorithm has a better precision and sensitivity for the non-crowded field, with an F1-score of 0.65, but is worse for the crowded fields with an F1-score of 0.035. In the non-crowded fields 95% of all streak and object matches are true matches to unique RSO, while for the crowded field only 10% are true matches. It was found that the 1''/pixel resolution in the images is too low for doing an IOD with Laplace’s method, despite how well the streak finding algorithm performs. However, with some improvements, the method is suitable as a cost effective way to verify known RSO in catalogues. / I den här rapporten presenterar jag en studie om att använda öppen astronomiska data för att göra initial banbestämning för artificiella rymdobjekt avbildade som streck i teleskopbilder. Syftet är att tillhandahålla information för att upprätthålla en god rymdlägesbild. Data från olika kartläggnings teleskop övervägdes men på grund av begränsningar i tillgänglighet valdes endast mask-bilderna från Zwicky Transient Facility för analysen. En algoritm utvecklades för att upptäcka streck i mask-bilderna och matcha dem med kända objekt i bildens synfält vid observationstillfället. Vidare gjordes den initiala banbestämningen med Laplaces metod, som använder vinkelkoordinaterna för streckens position vid observationen. Tillståndsvektorerna för strecken och de matchade objekten jämfördes, de beräknades från den initiala banbestämningen respektive objektets TLE. Algoritmen testades med 6 olika bildfält från observationsdatum mellan den 14:e till den 16:e december 2019, av dessa karakteriseras 4 som glesa och 2 som fyllda. Algoritmen för streck detektering har bättre precision och känslighet för de glesa fälten, med ett F1-värde på 0.65, men sämre för de fulla fälten med ett F1-värde på 0.035. I de glesa fälten är 95% av alla streck- och objektmatchningar korrekta matchningar med unika objekt, medan för det fulla fälten är endast 10% korrekta matchningar. Det visar sig att upplösningen på 1''/pixel i bilderna är för låg för att göra en initial banbestämning med Laplaces metod, oavsett hur bra algoritmen för streck detektering presterar. Genom att göra vissa förbättringar i algoritmen är metoden lämplig för att, på ett kostnadseffektivt sätt, verifiera kända objekt i kataloger.

Sea level rise concerns millions of people in coastal areas across the globe. One of the largest uncertainties to project future sea level rise is the frontal ablation (accounting for calving and submarine melt) at marine ice margins, around the Greenland and Antarctic Ice Sheet. High rates of frontal ablation have been observed to imply, through loss of the buttressing effect but not limited to it, increased mass loss from marine terminating glaciers and hence, associated sea level rise. This study focuses on calving processes at a freshwater lake in northern Sweden, which represents a simpler environment to study calving processes than the marine one, because impacts of tides, salinity, and circulation (all known to be relevant at marine ice-ocean boundaries) can be neglected. A multi-method approach to quantify and characterize calving events is presented here, exploring and analysing the underwater acoustic soundscape at a calving glacier front, in connection with optical, image-based methods such as time- lapse photography, and photogrammetry based on footage acquired by an uncrewed aerial vehicle (UAV). An acoustic detector is developed, tested and applied to data set acquired during 2020, and results indicate that the acoustic detector can be an important complement in the range of tools used to observe, and quantify, calving. Applied in remote locations, where continuous monitoring is difficult and where optical methods are of limited use, collecting acoustic data and monitoring calving by means of its acoustic signature could render insights previously not available (because of lacking data and methodology).

A problem that is prevalent for testers at Ericsson is that performance test results are continuously generated but not analyzed. The time between occurrence of problems and information about the occurrence is long and variable. This is due to the manual analysis of log files that is time consuming and tedious. The requested solution is automation with an algorithm that analyzes the performance and notifies when problems occur. A binary classifier algorithm, based on statistical methods, was developed and evaluated as a solution to the stated problem. The algorithm was evaluated with simulated data and produced an accuracy of 97.54 %, to detect trend breaks. Furthermore, correlation analysis was carried out between performance and hardware to gain insights in how hardware configurations affect test runs. / Ett allmänt förekommande problem för testare på Ericsson är att resultat från flera prestandatester genereras kontinuerligt men inte analyseras. Tiden mellan förekommande fel och informationen av dessa är hög och varierande. Detta på grund av manuell analys av loggfiler som är tidsödande och ledsamt. Den efterfrågade lösningen är automatisering med en algoritm, baserad på statistisk metodik, som analyserar data om prestanda och meddelar när problem förekommer. En algoritm för binär klassifikation utvecklades och utvärderades som lösning till det fastställda problemet. Algoritmen utvärderades med simulerad data och alstrade en noggrannhet på 97,54%, för att detektera trendbrott. Dessutom utfördes korrelationsanalys mellan prestandan och hårdvaran för att få insikt i hur hårdvarukonfigurationen påverkar testkörningar.

Recently ABB AB/Metallurgy has developed a novel temperature measuring system in the mold copper plate for continuous casting, named OptiMold Monitor, with the purpose of dynamically monitoring mold conditions and to control the FC Mold.The OptiMold Monitor temperature signals can be further analyzed for the information of the shape of the meniscus of the molten steel together with fluid flow symmetry and speed. Also, it can be analyzed for extracting information about how the steel has started to solidify in the mold and to detect solidification deficiencies such as cracks or risks of shell break-outs. Algorithms and Matlab codes developed by ABB for the thermal data analysis has perceived good insight into the results. The OptiMold Monitor system is currently being tested in Tata Steel IJmuiden steelworks.Algorithms for local cold and hot spot detection have been suggested for robust performance and to address the issue of false alarms. And the nail bed tests for meniscus profile and speed have been analyzed. / ABB AB / Metallurgi har utvecklat ett nytt temperaturmätningssystem i gjutkopparplattan för kontinuerlig gjutning, benämnd OptiMold Monitor, med syfte att dynamiskt övervaka aktuell gjutningsstatus och för att styra FC Mold.Temperatursignalerna från OptiMold Monitor- kan analyseras vidare för ge informationen om formen av det smälta stålet tillsammans med fluidflödessymmetri och hastighet. Det kan också analyseras för att extrahera information om hur stålet har börjat stelna i formen och för att upptäcka stelningsdefekter som sprickor eller risker för utbrytning av skal. Algoritmer och Matlab-koder som utvecklats av ABB för termisk dataanalys har givit god inblick i resultaten. OptiMold Monitor-systemet testas för närvarande i stålverket Tata Steel IJmuiden. Algoritmer för lokal kylning och detektering av ”hot spots” har föreslagits för att ge robusta prestanda och att för att hantera risken för falska larm. Även resultat från nagelbäddstesterna för gjutprofil och hastighet har analyserats.

Music analysis and source separation have become important and allied areas of research over the last decade. Towards this, analyzing a music signal for important events such as onsets, offsets and transients are important problems. These tasks help in music source separation and transcription. Approaches in source separation too have been making great strides, but most of these techniques are aimed at Western music and fail to perform well for Indian music. The fluid style of instrumentation in Indian music requires a slightly modified approach to analysis and source separation. We propose an onset detection algorithm that is motivated by the human auditory system. This algorithm has the advantage of having a unified framework for the detection of both onsets and offsets in music signals. This onset detection algorithm is further extended to detect percussive transients. Percussive transients have sharp onsets followed closely by sharp offsets. This characteristic is exploited in the percussive transients detection algorithm. This detection does not lend itself well to the extraction of transients and hence we propose an iterative algorithm to extract all types of transients from a polyphonic music signal. The proposed iterative algorithm is both fast and accurate to extract transients of various strengths. This problem of transient extraction can be extended to the problem of harmonic/percussion sound separation(HPSS), where a music signal is separated into two streams consisting of components mainly from percussion and harmonic instruments. Many algorithms that have been proposed till date deal with HPSS for Western music. But with Indian classical/film music, a different style of instrumentation or singing is seen, including high degree of vibratos or glissando content. This requires new approaches to HPSS. We propose extensions to two existing HPSS techniques, adapting them for Indian music. In both the extensions, we retain the original framework of the algorithm, showing that it is easy to incorporate the changes needed to handle Indian music. We also propose a new HPSS algorithm that is inspired by our transient extraction technique. This algorithm can be considered a generalized extension to our transient extraction algorithm and showcases our view that HPSS can be considered as an extension to transient analysis. Even the best HPSS techniques have leakages of harmonic components into percussion and this can lead to poor performances in tasks like rhythm analysis. In order to reduce this leakage, we propose a post processing technique on the percussion stream of the HPSS algorithm. The proposed method utilizes signal stitching by exploiting a commonly used model for percussive envelopes. We also developed a vocals extraction algorithm from the harmonic stream of the HPSS algorithm. The vocals extraction follows the popular paradigm of extracting the predominant pitch followed by generation of the vocals signal corresponding to the pitch. We show that HPSS as a pre-processing technique gives an advantage in reducing the interference from percussive sources in the extraction stage. It is also shown that the performance of vocal extraction algorithms improve with the knowledge about locations of the vocal segments. This is shown with the help of an oracle to locate the vocal segments. The use of the oracle greatly reduces the interferences from other dominating sources in the extracted vocals signal.