Spectrum optimization of light-emitting diode insecticide lamp based on partial discharge evaluation

The spectrum distribution of light-emitting diodes (LEDs) is a critical issue in adopting LEDs as an attractant in insect-catching apparatuses via phototaxis effects. Numerous research efforts have attempted to configure the LED spectrum in line with the insect-sensitive spectrum under designed test conditions. Although successful methods have achieved a stationary spectrum configuration, it is less effective when applied directly to an LED insecticide lamp due to the time-variant behaviors of insects in complex environments. Therefore, dynamic optimization for spectrum distribution of the LED insecticide lamp in practical fields remains essential to improve energy efficiency and the control quality of insects. In this paper, an online learning and dynamic control method has been proposed, where dead insects are evaluated by a partial discharge waveform and a cost function derived from the lamp's safety, efficiency, and effectiveness is optimized using a proposed jumping-predication algorithm. Then, a detailed procedure for developing a smart and effective LED insecticide lamp based on photovoltaic power supply is illustrated, and finally the LED insecticide lamp is validated. The suggested methodology and experimental results shed light on controlling insects and pests using an LED insecticide lamp in green agriculture.