Hybrid parameter identification of a multi-modal underwater soft robot

F Giorgio-Serchi,A Arienti,F Corucci,M Giorelli,C Laschi
DOI: https://doi.org/10.1088/1748-3190/aa5ccc
2017-02-28
Abstract:We introduce an octopus-inspired, underwater, soft-bodied robot capable of performing waterborne pulsed-jet propulsion and benthic legged-locomotion. This vehicle consists for as much as 80% of its volume of rubber-like materials so that structural flexibility is exploited as a key element during both modes of locomotion. The high bodily softness, the unconventional morphology and the non-stationary nature of its propulsion mechanisms require dynamic characterization of this robot to be dealt with by ad hoc techniques. We perform parameter identification by resorting to a hybrid optimization approach where the characterization of the dual ambulatory strategies of the robot is performed in a segregated fashion. A least squares-based method coupled with a genetic algorithm-based method is employed for the swimming and the crawling phases, respectively. The outcomes bring evidence that compartmentalized parameter identification represents a viable protocol for multi-modal vehicles characterization. However, the use of static thrust recordings as the input signal in the dynamic determination of shape-changing self-propelled vehicles is responsible for the critical underestimation of the quadratic drag coefficient.
robotics,engineering, multidisciplinary,materials science, biomaterials
What problem does this paper attempt to address?