Precise variable spraying system based on improved genetic proportional-integral-derivative control algorithm

Abstract

Currently, the most efficient method of resolving the pollution problem of weed management is by using variable spraying technology. In this study, an improved genetic proportional-integral-derivative control algorithm (IGA-PID) was developed for this technology. It used a trimmed mean operator to optimize the selection operator for an improved searching rate and accuracy. An adaptive crossover operator and mutation operator were constructed for a rapid convergence speed. The weed density detection was performed through an image acquisition and processing subsystem which was capable of determining the spraying quantity. The variable spraying control sub-system completed variable spraying operation. The performance of the system was evaluated by simulations and field tests, and compared with conventional methods. The simulation results indicated that the parameters of the overshoot (1.25%), steady-state error (1.21%) and the adjustment time (0.157s) of IGA-PID were the lowest when compared with the standard algorithms. Furthermore, the field validation results showed that the system with the proposed algorithm achieved the optimal performance with spraying quantity error being 2.59% and the respond time being 3.84s. Overall, the variable spraying system based on an IGA-PID meets the real-time and accuracy requirements for field applications which could be helpful for weed management in precise agriculture.