Bionic Artificial Lateral Line Underwater Localization Based on the Neural Network Method

The lateral line system is an essential mechanosensory organ for organisms such as fish; it perceives the fluid environment in the near-field through the neuromasts on the lateral line system, supporting behaviors (e.g., obstacle avoidance and predation in fish). Inspired by the near-field perception ability of fish, we propose an artificial lateral line system composed of pressure sensors that respond to a target's relative position by measuring the pressure change of the target vibration near the lateral line. Based on the shortcomings of the idealized constrained modeling approach, a multilayer perceptron network was built in this paper to process the pressure signal and predict the coordinates on a two-dimensional plane. Previous studies primarily focused on the localization of a single dipole source and rarely considered the localization of multiple vibration sources. In this paper, we explore the localization of numerous dipole sources of the same and different frequency vibrations based on the prediction of the two-dimensional coordinates of double dipoles. The experimental results show that the mutual interference of two vibration sources causes an increase in the localization error. Compared with multiple sources of vibration at the same frequency, the positioning accuracies of various vibration sources at different frequencies are higher. In addition, we explored the effects of the number of sensors on the localization results.

Automated summary

This paper proposes an artificial lateral line system composed of pressure sensors for perceiving the relative position of a target. A multilayer perceptron network is built to process the pressure signal and predict the target's coordinates on a two-dimensional plane. It is found that the mutual interference of multiple vibration sources increases the localization error, and the positioning accuracy is higher for vibration sources with different frequencies.