基于改进遗传算法的苹果收获机器人三维路径规划

Abstract

In recent years, the problem of “rural labor shortage” in China has become increasingly serious, with a large number of young laborers going out to work, leading to an increasing amount of idle rural land. The intensification of population aging and the reduction of agricultural labor force in China resulted in an urgent demand for agricultural robots. With the rapid development of agricultural machinery and automation technology, agricultural robots have been continuously developing. They can better adapt to the development of biotechnology in agriculture, and traditional harvesting methods may undergo significant changes, with an increased focus on the cultivation of crops. Therefore, this paper introduces a new encoding scheme on the basis of traditional genetic algorithm (GA) and proposes an improved double encoding GA. This new encoding scheme is used on the crossover link, whereas the path node sequence encoding scheme is still used on the mutation link. The selection operation is placed after the mutation, and the merging sorting and elitist selection are performed on the parent population, crossover population, and mutation population before selection, thereby accelerating the convergence speed. On the basis of the improved GA, the three-dimensional path of the apple harvesting robot is designed and planned, with the addition of adaptive adjustment function during the progress. The experimental simulation results show that the three-dimensional path planning of the apple harvesting mobile robot based on the improved GA can minimize the number of paths and loops and well meet the operational requirements of the harvesting robot.