No

The objective of this research effort is to provide a methodology for assessing the effectiveness of sampling techniques used to gather different types of geo-physical information by a robotic agent. We focus on assessing how well unique real-time sampling strategies acquire information that is, otherwise, too dangerous or costly to collect by human scientists. Traditional sampling strategies and informed search tech- niques provide the underlying structure for a navigating robotic surveyor whose goal is to collect samples that yield an accurate representation of the measured phenomena under realistic constraints. These sampling strategies are alternative improvements that provide greater information gain than current sampling technology allows. The contributions of this work include the following: 1) A method for estimating spa- tially distributed phenomena, using a partial sample set of information, that shows improvement over that of a more traditional estimation method. 2) A method for sampling this phenomena in the form of a navigation scheme for a mobile robotic survey system. 3) A method of ranking and comparing different navigation algorithms relative to one another based on performance (reconstruction error) and resource (distance) constraints. We introduce a specific class of navigation algorithms as example sampling strategies to demonstrate how our methodology allows different robot navigation options to be contrasted and the most practical strategy selected.

A constructed wetland (0.08 ha) receiving drainage water from a small agricultural catchment (22 ha) with clay soil, was investigated with respect to phosphorus dynamics and retention. The aim was to evaluate the function of the wetland with respect to phosphorus retention, and relate that to gross sedimentation as measured with sediment traps. Hydraulic load and phosphorus retention were estimated for 2003-2010 based on monitoring data. Furthermore, water quality dynamics was studied during three intensive sampling periods of 3-5 days during 2010. For each period, phosphorus retention was calculated and the relationship between flow and phosphorus concentrations analysed. Additionally, the gross sedimentation rate was estimated using sediment traps, and the phosphorus, carbon and nitrogen content analysed. The results suggested that there was no net retention of phosphorus during 2003-2010, except for 04/05. During the intensive sampling periods, release of phosphorus from the wetland mainly occurred during high flow. Sediment analyses showed that settling of inflow particles mostly occurred in the inlet pond, while the sediment found in a shallow vegetated area and outlet pond likely originated from internal processes rather than from the catchment. In fact, the gross sedimentation of phosphorus during April-July and July-August, respectively, exceeded the measured phosphorus inflow. The results showed that short periods with rapid flow increases were crucial for the wetlands function and thus high frequency sampling must be done during these periods. Furthermore, it seems that the particles lost from the catchment during high flows are too small to settle in the wetland.

Convolutional Neural Networks (CNNs) have been proved to deliver great results in the area of computer vision, however, one fundamental bottleneck with CNNs is the fact that it is heavily dependant on the ground truth, that is, labeled training data. A labeled dataset is a group of samples that have been tagged with one or more labels. In this degree project, we mitigate the data greedy behavior of CNNs by applying deep active learning with various kinds of sampling strategies. The main focus will be on the sampling strategies random sampling, least confidence sampling, margin sampling, entropy sampling, and K- means sampling. We choose to study the random sampling strategy since it will work as a baseline to the other sampling strategies. Moreover, the least confidence sampling, margin sampling, and entropy sampling strategies are uncertainty based sampling strategies, hence, it is interesting to study how they perform in comparison with the geometrical based K- means sampling strategy. These sampling strategies will help to find the most informative/representative samples amongst all unlabeled samples, thus, allowing us to label fewer samples. Furthermore, the benchmark datasets MNIST and CIFAR10 will be used to verify the performance of the various sampling strategies. The performance will be measured in terms of accuracy and less data needed. Lastly, we concluded that by using least confidence sampling and margin sampling we reduced the number of labeled samples by 79.25% in comparison with the random sampling strategy for the MNIST dataset. Moreover, by using entropy sampling we reduced the number of labeled samples by 67.92% for the CIFAR10 dataset. / Faltningsnätverk har visat sig leverera bra resultat inom området datorseende, men en fundamental flaskhals med Faltningsnätverk är det faktum att den är starkt beroende av klassificerade datapunkter. I det här examensarbetet hanterar vi Faltningsnätverkens giriga beteende av klassificerade datapunkter genom att använda deep active learning med olika typer av urvalsstrategier. Huvudfokus kommer ligga på urvalsstrategierna slumpmässigt urval, minst tillförlitlig urval, marginal baserad urval, entropi baserad urval och K- means urval. Vi väljer att studera den slumpmässiga urvalsstrategin eftersom att den kommer användas för att mäta prestandan hos de andra urvalsstrategierna. Dessutom valde vi urvalsstrategierna minst tillförlitlig urval, marginal baserad urval, entropi baserad urval eftersom att dessa är osäkerhetsbaserade strategier som är intressanta att jämföra med den geometribaserade strategin K- means. Dessa urvalsstrategier hjälper till att hitta de mest informativa/representativa datapunkter bland alla oklassificerade datapunkter, vilket gör att vi behöver klassificera färre datapunkter. Vidare kommer standard dastaseten MNIST och CIFAR10 att användas för att verifiera prestandan för de olika urvalsstrategierna. Slutligen drog vi slutsatsen att genom att använda minst tillförlitlig urval och marginal baserad urval minskade vi mängden klassificerade datapunkter med 79, 25%, i jämförelse med den slumpmässiga urvalsstrategin, för MNIST- datasetet. Dessutom minskade vi mängden klassificerade datapunkter med 67, 92% med hjälp av entropi baserad urval för CIFAR10datasetet.

No description available.