Page 110 - Read Online

Mai et al. Intell Robot 2023;3(4):466-84 I http://dx.doi.org/10.20517/ir.2023.37 Page 17 of 19

Figure 13. Convergence trend in medium complex environments.

Figure 14. Convergence trend in complex environments.

enhances the performance of path optimization, search efficiency, and convergence speed. This approach pre-
vents the occurrence of local optima and improves the path planning performance of the UAV. The simulation
results show that the path length of DSACO is reduced by 7.3%, 6.5%, and 4.3%, respectively, compared with
AS, MMAS, and improved AS. Compared with AS, MMAS, and improved AS, the fitness value of DSACO
decreased by 8.6%, 7.2%, and 2%, respectively. The algorithm reduces the number of iterations from 421, 407,
and 398 to 383, and the running time also increases slightly within the acceptable range.

However, theDSACOalgorithmstillhassomeshortcomings, suchasnotconsideringtheinfluenceofdynamic
obstacles when constructing the mountain model and not considering the complex kinematics and dynamics
constraints of UAVs, which limits its application. In the future, the path planning method can be improved
in the following aspects: (1) the current path planning method mainly deals with problems in static environ-
ments, and in the future, it should be extended to dynamic environments and consider the interaction effects
of the aircraft and other moving objects to realize smarter path planning; (2) machine learning technology can
be studied to be applied to path planning to improve the intelligence and adaptability of the path planning
algorithm by learning a large amount of historical path data and flight experience.