Hazardous liquid storage tanks are subjected to damages and failures. These tanks need to be inspected for risks of damage and collapse. Many robotic systems have been developed for the technical inspection of these tanks replacing manual inspection based on the classical non destructive techniques. The use of robots in the field still an open research problem. Several related issues are not well studied - for example, the navigation of the robot autonomously within the tank which requires the determination of the robot position and the mapping of the environment. This dissertation addresses the localization problem of the robot using a systematic approach. First, an overview of the sensors used for the localization of Autonomous Underwater Vehicle is done. This overview leads us to choose the appropriate sensors for the robot in oil tank. This dissertation also presents an error analysis study to choose the appropriate number of sensors. Finally, it addresses the implementation issues of these sensors through several experimental studies.