Partial System Identification and Sensor Fusion with the Jaiabot Micro Autonomous Underwater Vehicle

Abstract

This paper reports on the first approach to dynamical system modeling, identification, control, and environmental quantity estimation with a new, commercially available, micro-AUV named Jaiabot. The Jaiabot is a small and relatively inexpensive marine robotic sensor platform that dives vertically to collect measurements of related to water quality, such as salinity and temperature. The paper presents (i) the first dynamical model for Jaiabot's dive motion estimated based on field-data, (ii) a controller redesign given mission specifications associated with vertical environmental quantity profiling, and (iii) a first underwater implementation of a cooperative Kalman filter that fuses measurements from several Jaiabots to yield a confident estimate of temperature or salinity along a path within the convex closure of the vehicle formation.