Preliminary field trials of autonomous path following

As part of the ongoing Responsive AUV Localization and Mapping project (REALM) Memorial University has been developing an autonomous path localization and following system. This Qualitative Navigation System (QNS) localizes an AUV to a predefined path and generates control inputs to maintain the AUV along that path without the need for an absolute position estimate. QNS processes sonar data into the two-dimensional image space and performs feature extraction and matching. The results of this matching are input into a filter which allows localization of the AUV on the path and determination of either a waypoint or heading to maintain the AUVs traversal along the path. This ability to autonomously follow a path will be of great use for long term environmental monitoring. In May of 2013 QNS was deployed on Memorial's Explorer AUV and field tested in Holyrood, Newfoundland. International Submarine Engineering, manufacturer of the Explorer AUV, provided an interface which allowed the QNS software to request control of the AUV, provide command inputs, and relinquish control. A test path consisting of two connected 755m and 470m line sections was defined and used for preliminary tests. Although testing time was constrained, a series of successful tests were completed in which the AUV autonomously detected the path, localized itself and traversed the path to completion. The results of these tests validate the concept of QNS and the AUV control interface and will drive ongoing development and testing.