Background With the rapid development of the low-altitude economy, the number and speed of aircraft in partially open airspace are sharply increasing, intensifying the challenges of collision avoidance and safety management. Real-time perception of high-speed targets is therefore essential. However, existing sensors cannot meet the strict constraints of power, payload, and cost, highlighting the need for lightweight and energy-efficient alternatives.

Purpose Aircraft accumulate electrostatic charges through triboelectric effects between blades or skins and atmospheric particles such as dust, ice crystals, or raindrops, as well as through jet charging from engine exhausts. These charges disturb the surrounding electrostatic field, which can be detected by electrostatic sensors. Previous studies have demonstrated trajectory and velocity detection using flat plates, cylindrical electrodes, vector field sensors, and even self-powered or deep learning–assisted approaches. Yet, such electrode arrays are typically bulky, aerodynamically intrusive, and unsuitable for low-altitude platforms. To address these gaps, this study introduces a bio-inspired electrostatic sensing approach.

Methods Inspired by sharks’ Lorenzini ampullae, a conformal hemispherical electrode array was deployed to provide directional diversity and enhance spatial resolution. Theoretical modeling examined the effects of electrode orientation, target distance, and velocity on induction signals. A laboratory system was constructed using an electromagnetic launcher and attitude control module to fire charged metal spheres within a 4 m arena. A robotic platform equipped with the hemispherical array collected signals under two modes: central crossing and regional coverage. A symbolic regression model is then trained on the dataset to reconstruct the trajectory direction of flying targets.

Results Models trained solely on central crossing data exhibited overfitting and poor generalization. In contrast, models trained on regional coverage data achieved consistent accuracy across datasets, with omnidirectional prediction errors below 7.71°and variations within 1°, demonstrating strong robustness and generalization.

Conclusions This work presents a bio-inspired electrostatic sensing method for detecting charged flying targets. The system achieves accurate omnidirectional trajectory prediction with low error and minimal energy demand. The approach is passive, lightweight, low-cost, and easy to integrate, offering considerable potential for low-altitude airspace safety management. Future studies will focus on improving trajectory prediction accuracy and extracting multidimensional flight information.