Abstract:In order to solve the problem of land grader in unmanned operation lacks local planning to realize on-line adjustment of leveling path, a three dimensional path planning method for automatic navigation of farmland leveling based on improved ant colony algorithm was proposed for the purpose of reasonable loading and unloading of earthwork and the shortest path in land leveling operation. Based on the three dimensional terrain model of farmland, the improved ant colony algorithm was used for three dimensional path planning. A new path search node was established based on the decision direction of earthwork transportation in leveling operation. The pheromone updating rules and heuristic functions of land leveling were set up by comparing the earthwork carried by the leveling shovel and the excavation and filling volume required by grid calculation, and then the optimal three dimensional path for earthwork transportation was obtained. The path was smoothed and optimized according to the steering constraints of land leveler, which was based on the kinematical model of the land leveler. And the effect evaluation standard of three dimensional path planning was established. The simulation results showed that compared with the original ant colony algorithm, the path planning effect of this method was improved by more than 33.3%, which could better guide the land leveler to realize the local leveling task. Moreover, this method shortened the path generation time and path length, making the path smoother, and was more suitable for automatic navigation of farmland leveling.