This thesis presents a solution to the problem of part localization in fixtureless assembly. Fixtureless assembly is a new technology in which complex and expensive dedicated fixtures are replaced by robots. However, the elimination of the fixtures implies that the robots must now be able to achieve the same locational accuracy of parts as precision fixtures provide. The implementation of an Automatic Part Localization System (APLS) allows the robots to determine the accurate position and orientation of the parts that are placed in arbitrary locations. In this thesis, the problem of part localization is studied in the context of an application to automotive sheet metal part assembly, specifically car body fenders. The development of the APLS is summarized. The APLS uses laser beam sensors to estimate, in real time, the spatial position and orientation of a sheet metal part held by a robot. It comprises two groups of sensors:​ proximity sensors that measure the location of the part’s surface, and edge sensors that measure the location of the part’s edges. From these surface and edge feature data, the APLS is able to localize the part for all possible coordinates. In other words, it is able to solve 6 DOF localization problems. The implementation of an existing localization method is described. The localization algorithm presented uses minimization of least-square to generate the best-fit mapping between the part’s geometry and the corresponding sensor information. Experiments are conducted to investigate the feasibility of the proposed part localization concept. The experimental results show that the presented method meets the target performance of this investigation - to localize the part within the industrial tolerance of ±0.5mm in selected positional coordinates. The experiments are carried out using a 6 DOF Fanuc commercial robot.