Underwater motion target recognition using artificial lateral line system and artificial neural network method

A well-developed lateral line system enables fish to perceive their surroundings and make appropriate movements. Inspired by such a biological characteristic, an artificial lateral line system (ALLs) is proposed to differentiate the shape and dimensions of dynamic targets. A total of 25 distinct perception models approaching the carrier are studied to analyze the perceptual characteristics of the ALLs. The pressure distribution on the fish body surface is calculated and extracted using the computational fluid dynamics (CFD) method, and is verified by experimental tests. An artificial neural network (ANN) model is then constructed for building the relationship between the ALLs and the perceived target features. The implementation of a suitable network architecture results in the achievement of a more precise categorization outcome. The findings indicate that the artificial lateral line system, as described, demonstrates exceptional proficiency in perceiving underwater targets.