Robust Flow Field Signal Estimation Method for Flow Sensing by Underwater Robotics

The flow field is difficult to evaluate, and underwater robotics can only partly adapt the submarine environment. However, fish sense complex environment by their lateral line system. In order reveal sensing mechanism, a robust nonlinear signal estimation method based on Volterra series model with Kautz kernel function provided, which named KKF-VSM. around square target used as original signal. sinusoidal noise triangular obstacle are considered undesired signals, predicting performance of KKF-VSM analyzed after introducing them locally in signals. Compared radial basis neural network (RBF-NNM), advantages not its robustness but also higher sensitivity weak signals accuracy. It confirmed that even for strong such pressure responses field, more efficient than commonly RBF-NNM. provide reference application artificial system robotics, improving adaptability environments information.