Fusion of local and global information for the localization of an intelligent vehicle | |
| Author | Chirdpong Deelertpaiboon |
| Call Number | AIT Diss. no.ISE-09-02 |
| Subject(s) | Global Positioning System Electronics in nagivation Software localization |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Dissertation ; no. ISE-09-02 |
| Abstract | Mapping serves as an important feature for autonomous vehicle guidance system. Nowadays, there are three common approaches to localize the vehicle as well as to build a map. The first one is based on magnetic sour ces embedded along the desired trajectory. The second approach requires a GPS to provide the global location of some specified points and then connecting these points by either straight or curve lines to draw a global map. The last one becomes widely accepted among robotic researchers. It is called Simultaneous Localization and Mapping (SLAM). This method allows the vehicle to move itself in the unknown environment and build a global map based on environment sensed by various sensors attached to the vehicle. However, even though the magnetic based approach has some advantages. For example, it is a passive device. There is no physical contact. It is independent of weather condition. However, the important disadvantage is that magnetic sources must be installed al ong the trajectory. For the second approach, GPS errors are the main concern as well as the signal blockage that might cause the loss of satellite signal. Meanwhile, SLAM approach illustrates very autonomous method. However, the calculation time depends largel y on the number of detected landmarks. This thesis establishes a novel strategy to eas ily obtain waypoints which will be positioned right at the middle of road lane by simply choosing waypoints from a map image. This map image can be obtained by scanning a pape r map or capturing satellite image from available software such as Google Earth. Then , straight lines are connected between two consecutive waypoints to draw a designated path. However, this kind of map image certainly has errors associated with it. This dissertation will address three kinds of errors namely scaling, orientation, and translational errors. Our method can automatically correct them by fusing the information from a GPS, a compass, and a camera. In the real world, even the most expensive sensor still has some errors. As a result, uncertainties must be considered. In order to preserve and propagate the sensor errors until the end results are achieved, fuzzy based alpha-c ut operation is applied. Hence, each sensor output is modelled as fuzzy set with triangu lar shape membership function. Alpha-cut is the closed interval that can fully and un iquely define fuzzy set. Consequently, mathematical operations and functions can be performed based on closed interval operation and the extension principle. Furthermor e, this dissertation w ill explain in detail of the algorithm to combine the information fr om two referencing types of sensors. In other words, one is the global referencing sensors which are a GPS and a compass. The other one is the local referencing sensor which is a camera. As a whole, this work compromises between computational time and inflexibility. To prove this statement, an experiment was c onducted inside Asian In stitute of Technology (Thailand) campus. The result shows a prom ising output as the computed waypoints relocated themselves correctly along the designed path. |
| Year | 2009 |
| Corresponding Series Added Entry | Asian Institute of Technology. Dissertation ; no. ISE-09-02 |
| Type | Dissertation |
| School | School of Engineering and Technology (SET) |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Industrial Systems Engineering (ISE) |
| Chairperson(s) | Poompat Saengudomlert; |
| Examination Committee(s) | Afzulpurkar, Nitin V. ;Manukid Parnichkun; |
| Scholarship Donor(s) | Ministry of Science and Technology Scholarship; |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2009 |