Bio-Inspired Electrostatic Detection Method for Threat Perception in Autonomous Platforms

Autonomous platforms have been widely adopted in both civilian and military contexts, achieving notable success. However, their energy resources, computational power, payload capacity, and cost are constrained and interdependent. These limitations prevent the integration of advanced threat detection systems, such as high-speed cameras and radar. Consequently, they are unable to detect high-velocity projectiles, significantly undermining their survivability in hostile environments. Developing low-power, cost-effective, and computationally efficient threat detection methods is of critical importance. This letter proposes a method for perceiving the flight trajectories of charged objects based on the principle of electrostatic induction. A physical model for electrostatic detection is established, and the effects of flight speed and trajectory on induction signals are analyzed. Inspired by sharks, a four-element array testing system with specific orientation and spatial distribution characteristics is designed. An indoor experimental setup is developed to simulate and test the flight of charged objects, generating a comprehensive dataset by varying parameters such as flight speed, incident angle, and testing distance. This dataset is then used to train a symbolic regression machine learning model, resulting in a mathematical model capable of predicting the incoming direction of charged objects with an error margin of less than 10 degrees. The model's generalization ability and the impact of the number of sensors are also discussed.