碩士 / 國立虎尾科技大學 / 電機工程系碩士班 / 106 / In order to face the weakness of GPS used in indoor environments, the ZigBee systems have widely discussed to be as a possible solution in recent years. In the previous studies, the proposed methods were often considered to shape a geometric topology (e.g., rectangle) to fit the whole area of the experiment environment, and then to arrange suitable number of reference nodes on the area of the geometric topology. The anchor node receives the RSSI values of its around reference nodes to estimate the distance between the anchor node and each reference node, and then applies the Trilateration method to evaluate its location. In practice, the new applied positioning area could be much larger than the former designed one. In this case, the result will be worse than the original design because the accuracy of estimated distance by using RSSI value is decayed due to the increase of distances between the anchor node and each reference node. In addition, when the original designed geometric topology is unsuitable to fit the contour of new applied positioning area, it needs to redesign the system. In other words, it means that the method needs to be considered for adapting different cases and then it increases the complexity to apply for various environments.
In our previous study, we presented a hexagonal topology to arrange the reference nodes to fit all different contours and sizes of applied positioning areas. We first setup a fixed size of regular hexagonal area (called a cell) with reference nodes put on six hexagonal vertices and one on its center. Then, we use our previous proposed dynamic hybrid trilateral positioning method to do the positioning task. As the results, a better positioning accuracy is obtained than it is with rectangular topology. Then, we arrange the cells (regular hexagons) to cover the whole positioning area, called cellular structure. According to the 7 most strength of RSSI values received from all reference nodes, we proposed a rule to judge the correct cell of the anchor node located. When the anchor is located near the border between two adjacent cells (called blurred border) it is possible to make an error judgment to the incorrect cell since the anchor received the almost same level RSSI values from the reference nodes of the two adjacent cells. Therefore, in order to overcome the problem of blurred border, we will propose a new cellular structure called Dual-Cell topology. We first are like the former cellular structure to arrange cells to cover whole positioning area as the first layer. Meanwhile, we arrange cells of the second layer to cover the whole area with each cell is shifted 1/3 hexagonal distance in horizontal and vertical direction to each one of the first layer cells. After the Dual-Cell topology is constructed, the anchor node receives the RSSI values of all around reference nodes. According to the strength arrangement of RSSI values, we will propose a dual-cell decision strategy to select a suitable cell of the anchor node located from the two layers where the anchor can avoid located in the blurred border of the selected cell. When we use the dynamic hybrid trilateral positioning method to do the positioning task, the experimental results show that the proposed strategy can correct all of the wrong cells selected with the former cellular structure.
In addition, we apply our ZigBee dual-cell positioning system to the indoor navigation of a mobile robot to confirm the practicality. In order to avoid the positioning error caused from the hardware of the mobile robot, we will first propose a calibration experiment to correct the errors of the four wheels driving system. After minimizing the positioning error from the hardware, the proposed positioning system is performed by the mobile robot in our Lab and then is analyzed. The analysis results show the high positioning resolution and good performing stability. As a result, the positioning system is not only to be a valuable indoor positioning scheme, but also can be developed in the future for the applications of mobile robots used in any indoor area size.
| Identifer | oai:union.ndltd.org:TW/106NYPI0441003 |
| Date | January 2018 |
| Creators | ZHENG, GUANG-JIE, 鄭光杰 |
| Contributors | HUANG, KUO-TING, 黃國鼎 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 67 |
Page generated in 0.0016 seconds