Micro Unmanned Surface Vehicle for Shallow Littoral Data Sampling

This paper describes the creation of an autonomous air boat that can be carried by one person, called a micro unmanned surface vehicle (USV), for sensor sampling in shallow littoral areas such as inlets and creeks. A USV offers advantages over other types of unmanned marine vehicles. Unlike an autonomous underwater vehicle, the Challenge 1.0 air boat can operate in shallow water of less than 15 cm depth and maintain network connectivity for control and data sampling. A USV does not require a tether, like a remotely operated marine vehicle (ROV), which would limit the distance and mobility. However, a USV operating in shallow littoral areas poses several challenges. Navigation is a challenge since rivers and bays may have semi-submerged obstacles and there may be no depth maps; the approach taken in the Challenge 1.0 project is to let the operator specify a safe area of the water by visual inspection and then the USV autonomously creates a path to optimally sample the collision free area. Navigation is also a challenge because of platform dynamics–the USV we describe is a non-holonomic vehicle; this paper explores spiral paths rather than boustrophedon paths. Another challenge is the quality of sensing. Water-based sensing is noisy and thus a reading at a single point may not reflect the overall value. In practice, areas are sampled rather than a single point, but the noise in the point values within the sampled area produce a survey with widely varying numbers and are difficult for humans to interpret. This paper implements an inverse distance weighting interpolation algorithm to produce a visual “heatmap” that reliably portrays the smoothed data.