Objective  To address the problems of low efficiency in traditional manual lettuce sowing, as well as difficulty in controlling furrow depth and uneven seed spacing caused by the small particle size of lettuce seeds, this study designed an intelligent control system for lettuce sowing. The system integrates multi-sensor technology and multi-motor coordinated control, aiming to fill the research gap in specialized seeding control systems for small-particle-size vegetable seeds (represented by lettuce) and improve the automation and precision levels of lettuce sowing.

Method An embedded architecture with an STM32 chip as the core was constructed. The hardware and software of the control system were developed using a modular design method. A multi-modal sensing system was integrated to achieve environmental perception data fusion. A multi-motor coordination control system and a dynamic switching system for encoder reading channels were designed based on the PCA9685 chip and CD4052 analog multiplexer. A single-cascade PID fusion algorithm with attitude feedback was established to ensure motion stability under complex terrains. An OLED display and Android BLE interactive terminal were integrated.

Result Field tests showed that, the ditch depth qualification rate of the lettuce seeder in the test fields reached 77.3%, the missed seeding rate was 5.9%, the seedling emergence rate reached 49.1%, the total operation time was 57 s, and the efficiency was approximately three times that of manual seeding. The cost required for the operation of the seeder was drastically lower than that of the traditional manual sowing.

Conclusion The design of this control system meets the expected requirements of the prototype, providing a practical and feasible design scheme for the intelligence of the lettuce seeders.