It will be proposed different approaches for each of the described subsystems: the Dynamic Mission Planner, the Guidance System, the Control System and the Navigation System. They will be evaluated and verified on the experimental prototype.
For the DMP, it is expected to add new rules for describing more complex scenarios and cases required by the users as, for instance, multiple pipelines tracking, recognition of different pattern shapes like fishing nets, mines, anchors, and wastes nearby to the pipeline. To achieve this it is proposed a devoted sensor that uses the sonar information for filtering it by means of an artificial neural network and operates as a special input to the knowledge base. For the Guidance and Control Systems, it is expected to continue searching which is the best approach for the problem to be resolved. Specially, is desirable that the system that finally controls the AUV is of the adaptive type in order to minimize possible errors that can appear due to on-board variable payload. For the Navigation System, after studies and experimental tests it will be selected a method to implement a sensor fusion system to determine the proper AUV position and attitude, with a bounded accuracy. Based on this information it will be possible to obtain the target positions and recognized patterns in the seabed. Among the possible approaches to be analyzed, the most promising seems to be Kalman's filters and those who combine sensor inputs using fuzzy logic techniques. It is also expected that a side-scan sonar and a magnetic sensor for target tracking will be added in future stages of the project to enhance the possibilities of target and obstacle perception of the AUV.
Was this article helpful?