An adaptive scheme of sharing compressed flow information among networked underwater gliders

Underwater glider fleets are often used cooperatively for large-scale ocean observation. A key requirement for underwater glider operation is to reduce the amount of information that needs to be shared among gliders over communication links. This paper presents a method to estimate the quality of communication link between underwater gliders sharing their information. The representation of high flow region that gliders should notice is compressed using the support vector data description (SVDD) method. The ratio of compression of SVDD is optimized to match the link quality and SVDD error estimation. The compressed data is then communicated among gliders for a distributed path planning algorithm. We show that such an adaptive compression method can provide desired accuracy.