a. r. chehreghan

With the increasing use of spatial data in daily life, the production of this data from diverse information sources with different precision and scales has grown widely. Generating new data requires a great deal of time and money. Therefore, one solution is to reduce costs is to update the old data at different scales using new data (produced on a similar scale). One approach to updating data is to use the updated large-scale dataset as reference data to update small-scale datasets; in this way, the modified features are identified in the two datasets and then updated by modifying the changes to fit the small-scale dataset. In terms of the type of the updated feature, map updating issues in the vector dataset are divided into three categories: pointwise, linear, and polygon. One of the most important features of the class of polygonal features in urban environments are the buildings that are vital in urban maps and their updating process in urban applications has a high priority. In this research, an attempt has been made to study the issue of updating polygonal features from different perspectives by a careful and comprehensive examination. These perspectives include spatial clustering methods, pattern extraction, and updating methods. 
Spatial clustering methods are classified into five

natural principles, partition-based, graph-based, Region Merging, and density-based approaches. Each clustering method have been used in different studies according to Gestalt criteria (proximity, similarity, and continuity). In pattern extraction, different types of patterns have been studied in various studies while the linear pattern extraction as a sample has been comprehensively examined in this study.  In the updating methods, three updating approaches including Propagating, Local, and Constraint-based are examined. In the Propagation updating approach, only large-scale data are updated, then these updates can be propagated into small-scale data. This update is especially used for the MRDB spatial database by propagating updates at various scales. Local updating consists of three steps: 1) Change detection between the recently updated large-scale dataset and the old small-scale dataset. 2) Integrating the discovered changes into the small-scale dataset (by quantifying and formulating these changes). 3) Ensuring that the consistency is maintained. In Constraint-based updating approach which demonstrates the necessities of this research, first, by grouping the buildings, useful information such as the area of ​​the buildings, the average and the standard deviation of the separation distance between the buildings is obtained. Then, with the appropriate operators, Constraint-based generalization is performed to update the maps.  There are various criteria for evaluating the updating methods, while the Precision and Recall criteria have been used to evaluate the effectiveness of these problems. Precision criterion examines the ratio of the number of correctly matched features to the total number of features that that have been matched correctly or incorrectly. Recall criterion evaluates the ratio of correctly matched features to the total features that have been matched correctly or not matched. In other words, in the Precision criterion, the number of features that are incorrectly matched and in the Recall criterion, the number of features that are not matched, are effective in comparing the two criteria. The categorization mentioned in each perspective and the advantages and disadvantages of different methods are also presented in this research.

The expansion of location-based services (LBS) and their applications has led to a growing interest in localization, which can be done on the smartphone platform. Various positioning techniques can be used for indoor or outdoor positioning. Indoor positioning systems have been one of the most challenging technologies in location-based services over the past decade. Considering the increase of people activities inside buildings such as offices, hospitals, and large stores, determining the position and guidance of people inside these buildings is one of the most urgent and important issues to be discussed and challenged in the area of Location-based Services (LBS). There are various ways to determine the position inside a building. The method(s) used to determine the position in an indoor environment depends on several factors such as cost, accuracy, independence of, or dependence on the infrastructure, security, and system scalability. This study focuses on the infrastructure requirements necessary to determine the position of individuals thorough a comprehensive study of previous studies. Moreover, focusing on the Pedestrian Dead Reckoning positioning method using smartphones as an infrastructure-free method, several effective aspects of the accuracy and positioning process are examined. The effective measures examined include the use of a variety of noise filtering, combined filters (Particle filter, Kalman filter), the criterion of the of sensor data classification algorithm, the criterion of the initial point determination, the use of landmarks as checkpoints and plot maps for setting the estimated position, the detection criteria and estimation of the length of the step, and the user direction estimation criteria. The particle filter has good accuracy in small-scale areas, but in large-scale areas, it is out of date and has problems due to the limited source of the smartphone. In studies, Kalman filter has been used to integrate the information of different sensors, some of which have reached the desired accuracy according to the state model and the measurement model. Given that the generalized Kalman filter has a simple formula for nonlinear estimation, the linearization of the positioning problem causes an error in the Jacobi Matrix model and reduces the accuracy of the estimate, which negatively affects the cost of calculations and system timeliness. Step length varies from person to person. In fact, there should be a variable associated with pedestrians in the step estimation model. Also, a person's walking rate during a walk is not constant. Accordingly, assuming a constant value of step length for users causes an error during positioning and a large drift at the end of the path. Determining the heading is one of the most challenging parts of the PDR system because the heading error leads to a quick increase in the positioning error. It is difficult to determine the reliable heading in the environments with high magnetic disturbances. Another problem is that the heading of the smartphone may vary with the heading of the pedestrian movement. Therefore, two main tasks must be performed before implementing indoor positioning. One of them is to determine the heading of the smartphone. Another is to infer the offset heading between the smartphone and the pedestrian movement. Therefore, determining the state of the smartphone is necessary for specifying the heading of the pedestrian movement.  Finally, the advantages and disadvantages of each of the infrastructure-based and infrastructure-free methods are compared and evaluated. Also, the research uses algorithms such as Naive Baye, MLP, SVM, DT and KNN to classify the type of user movement and phone holding mode.

Spatial data, which is one of the main needs of human societies from business organizations to the general users today, cannot meet the needs of a wide range of users without changing the structure of conventional methods of data registration and updating on a metropolitan scale. Open Street Map, as one of the most successful implementations of the crowdsourcing approach to spatial data with the participation of the public users as the millions of scattered sensors in the environment, has provided a solution to achieve several goals such as improvement in rapid data collection and coverage of areas. Apart from the numerous benefits of volunteered geographic information, the reliance of this information solely on the participation and activity of users challenges the use of this data in various applications. However, this problem is exacerbated by the lack of the necessary frameworks for entering incorrect data, different levels of user knowledge, and different methods of data entry. Among the quality parameters, the completeness parameter, which provides the presence of one data set compared to another data set and shows the degree of importance of the features from the users' point of view, was discussed in the present study. It is also important to address this aspect of the quality of VGI in terms of its impact on the other quality parameters. In this study, after introducing the common approaches in evaluating this element of data quality and comparing the results of methods, the strengths and limitations of each method are also explained. In the general classification, the completeness parameter is examined by two region-based and the object-based approaches. After comparing results, the necessary pre-processing, and the analysis time of each method, the findings of this study demonstrate the high speed of evaluation of the region-based approach and the higher accuracy of object-based methods. Moreover, the mentioned advantage of region-based approach is achieved if there is no topological error in the VGI and there are correspondence relationships between the uses of both data sets. On the contrary, the object-based methods despite the high calculation due to the matching process; If they use the appropriate matching algorithm and guarantee the positional accuracy, they will provide more accurate results for the completeness parameter. Moreover, the object-based methods, with the possibility of calculating the omission and commission of the VGI data set compared to the official data set, provide a more accurate understanding of the completeness parameter. In the region-based approach, it is not possible to evaluate the completeness parameter separately due to the combination of these two sub-sectors. In the region-based approach, evaluation of the completeness parameter was implemented by parameters such as the ratio of the length of the two data sets, the ratio of the number of constituent points of features, and the ratio of the number of features. From the above three criteria, the length ratio of the two data sets was selected as a suitable evaluation criterion by less pre-processing analysis such as creating the graph structure in the number of features criteria and simplifying the lines in the constituent points of features. However, the length criterion in topologically erroneous datasets strongly affects the evaluation results. At the end, after introducing and using the vector matching algorithm based on determining the geometric similarity of the features, the completeness parameter was calculated as 66.2. High rate of two quantities of omission and commission information of the VGI data set compared to the official data set of 33.8 and 47.8 in the current study due to the weakness of the matching algorithm in the two data sets, it is expected that with the improvement of this algorithm, the amount of identified omission and commission information will be reduced.

The purpose of this paper is to evaluate and improve the accuracy of indoor positioning using smartphone sensors based on Pedestrian Dead Reckoning (PDR) method. In some specific situations, such as fires or power outages that disable infrastructure-based positioning techniques, using PDR method based on smartphone sensors that perform positioning continuously is a good solution.This paper focuses on determination of the user’s movement type to evaluate effective components of indoor positioning method. First, movement samples are evaluated with the feature-vectors of data from sensors and three classification algorithms (Decision Trees (DT), Support Vector Machine (SVM), and K-Nearest Neighbor (K-NN)). From the perspective of feature-vectors, the proposed features significantly improve the performance of three classification algorithms compared to previous research features. From the perspective of classification algorithm also Support Vector Machine had best performance with %99.3 accuracy, while spending the most time.  In the second phase, step detection is performed for norm acceleration values based on the definition of the upper and lower threshold and the time threshold. The directional component is also obtained by combining accelerometers, magnetometer and gyroscope sensors. Localization tests were performed while the user holding the phone in front of him with two states (normal walking, running) in three paths of different geometry (squares, circles and rectangles). The final accuracy obtained from normal walking test for three paths of square, circular and rectangular shapes was %4.8, %3.6, and %2, respectively. The final accuracy of the running mode was also obtained for three paths of square, circular and rectangular shapes equal to %8.4, %5.7, and %4, respectively.