This thesis presents the development of a localization and mapping algorithm for an autonomous underwater vehicle (AUV). It is based on probabilistic scan matching of raw sonar scans within a pose-based simultaneous localization and mapping (SLAM) framework.<\/p>\n
<\/p>\n
To address the motion-induced distortions affecting the generation of full sector scans, an extended Kalman filter (EKF) is used to estimate the robot motion during that scan. The filter uses a constant velocity model with acceleration noise for motion prediction.Velocities from Doppler velocity log (DVL) and heading measurements from attitude and heading reference system (AHRS) are fed asynchronously and update the state. The scan is undistorted by compounding the relative robot position in the scan, with the range and bearing measurements of the beams gathered by the sonar. Assuming Gaussian noise, the algorithm is able to estimate the uncertainty of the sonar measurements with respect to a frame located at the center of the scan.<\/p>\n
<\/p>\n
For estimating the global trajectory of the vehicle, a second filter, an augmented state EKF (ASEKF), stores the pose of the vehicle where each full scan was completed. Each new full scan is cross registered with all the previous scans that are in a certain range and a modified probabilistic iterative correspondence (pIC) algorithm is applied. This technique has a twofold effect: first, we obtain a better estimate of the vehicle’s displacement that is then used to update the ASEKF, and second, loop-closing events are updated automatically and simultaneously.In addition, we present a closed form method for estimating the uncertainty of the scan matching result.<\/p>\n
<\/p>\n
The proposed method is well suited for confined environments including but not limited to: geological folds, boulder areas, caves, man-made walls and structures, where horizontal scanning sonar can constantly detect and distinguish its surroundings over most of the vehicle’s trajectory.<\/p>\n
<\/p>\n
The method was tested with three real-world datasets: one obtained in an abandoned marina during an engineering test mission, and two additional ones in the natural environment of underwater cavern systems. In the marina dataset, the results show the quality of our algorithm by comparing it to the ground truth from a global positioning system (GPS) receiver and to other previously published algorithms. For the cavern datasets, the results are compared against fixed ground truth points that the vehicle visits twice along the trajectory it travels. In all the experiments the trajectory correction is notable and the unoptimized algorithm execution time is much faster than the experiment time, indicating the potential of the algorithm for real-time on-board AUV operation.<\/p>\n
<\/p>\n
________________________<\/p>\n
<\/p>\n
Aquesta tesis presenta el desenvolupament d\u2019un algoritme de localitzaci\u00f3 i mapeig per un vehicle submar\u00ed aut\u00f2nom (AUV).<\/p>\n
<\/p>\n
L\u2019algoritme es basa en t\u00e8cniques probabil\u00edstiques de scan matching<\/i> utilitzant scans<\/i> de sonar dins un sistema de localitzaci\u00f3 i construcci\u00f3 simult\u00e0nia de mapes (SLAM) basat en posici\u00f3.<\/p>\n
<\/p>\n
S\u2019utilitza un filtre de Kalman ext\u00e8s (EKF) per estimar el moviment del robot durant l\u2019adquisici\u00f3 de l\u2019scan<\/i> i aix\u00ed solucionar les distorsions causades pel moviment al llarg de l\u2019adquisici\u00f3 d\u2019un scan<\/i> complet.<\/p>\n
<\/p>\n
Aquest filtre utilitza un model de velocitat constant amb l\u2019acceleraci\u00f3 com a soroll per predir el moviment.<\/p>\n
<\/p>\n
Velocitats provinents d\u2019un Doppler velocity log<\/i> (DVL) i mesures d’una unitat d’orientaci\u00f3 s\u2019introdueixen al filtre as\u00edncronament per tal d\u2019actualitzar l\u2019estat. La distorsi\u00f3 de l\u2019scan<\/i> s\u2019elimina gr\u00e0cies a la composici\u00f3 de la posici\u00f3 relativa del robot dins l\u2019scan<\/i> amb les mesures de rang i la orientaci\u00f3 dels feixos del s\u00f3nar. Assumint un soroll Gaussi\u00e0, l\u2019algoritme \u00e9s capa\u00e7 d\u2019estimar la incertesa de les mesures del s\u00f3nar respecte a un sistema de coordenades situat al centre de l\u2019scan<\/i>.<\/p>\n
<\/p>\n
Per estimar la traject\u00f2ria global del vehicle, un segon filtre, en aquest cas un EKF d\u2019estat augmentat (ASEKF), guarda la posici\u00f3 del vehicle on es finalitza l\u2019adquisici\u00f3 de cada scan<\/i> complet. Cada nou scan<\/i> es registra amb tots els scans<\/i> previs localitzats dins un rang determinat i s\u2019aplica una modificaci\u00f3 de l\u2019algoritme probabilistic iterative correspondence <\/i>(pIC).<\/p>\n
<\/p>\n
Aquesta t\u00e8cnica t\u00e9 dos prop\u00f2sits: primer, obtenir una millor estimaci\u00f3 del despla\u00e7ament del vehicle que posteriorment s\u2019utilitzar\u00e0 per actualitzar l\u2019estat del ASEKF, i segon, actualitzar autom\u00e0ticament i simult\u00e0niament events de tancament de bucles. A m\u00e9s a m\u00e9s, es presenta un m\u00e8tode de formulaci\u00f3 tancada per estimar la incertesa del resultat de l\u2019scan matching<\/i>.<\/p>\n
<\/p>\n
El m\u00e8tode proposat \u00e9s adequat per entorns confinats, com per exemple:\u00a0 plecs geol\u00f2gics, zones rocoses, coves, parets artificials i estructures on un sonar d\u2019escaneig horitzontal pugui constantment detectar i distingir el seu voltant al llarg de la major part de la traject\u00f2ria del vehicle.<\/p>\n
<\/p>\n
El m\u00e8tode s\u2019ha provat en tres datasets<\/i> adquirits en entorns reals: un obtingut en una marina abonadonada i dos m\u00e9s en l\u2019entorn natural d\u2019un sistema de coves submar\u00ed. Els resultats presentats pel dataset<\/i> de la marina demostren la qualitat de l\u2019algoritme comparant-lo amb mesures d\u2019un sistema de posicionament global (GPS) i amb dos algoritmes pr\u00e8viament publicats. Pel qu\u00e8 fa els datasets<\/i> de la cova, els resultats es comparen amb punts de refer\u00e8ncia coneguts i fixes que el vehicle revisita al llarg de la traject\u00f2ria. En tots els experiments la correcci\u00f3 de la traject\u00f2ria \u00e9s notable i el temps d\u2019execuci\u00f3 de l\u2019algoritme \u00e9s molt m\u00e9s r\u00e0pid que el temps de l\u2019experiment, indicant aix\u00ed el potencial de l\u2019algoritme per operar en temps real des d\u2019un AUV.<\/p>\n","protected":false},"excerpt":{"rendered":"