A diversified group teaching optimization algorithm with segment-based fitness strategy for unmanned aerial vehicle route planning. (15th December 2021)

• Record Type: Journal Article
• Authors: Jiang, Yuxin; Wu, Qing; Zhang, Guozhong; Zhu, Shenke; Xing, Wei
• Is Part Of: Expert systems with applications. Volume 185(2021)
• Publication Date: 2021-12-15
• Subjects: Unmanned aerial vehicle -- Group teaching optimization algorithm -- Three-dimensional route planning; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2021.115690
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: Proposing an improved group teaching optimization algorithm for UAV route planning. A mathematical model of UAV route planning with various obstacles is established. Introducing a segment-based fitness strategy to deal with the constraints. Performance of the method is assessed on 3 different UAV flight environment models. The proposed algorithm is very competitive and superior to the compared algorithms. Abstract: The complexity and diversity of the flight environment pose great challenges to unmanned aerial vehicle route planning, which demands feasible flight strategies and efficient route planning algorithms. To address the issue, this paper constructs a 3-D flight environment model with multiple obstacles, and designs a novel diversified group teaching optimization algorithm for the generation of flight routes of unmanned aerial vehicles. In the environment model, a variety of obstacles are taken into consideration to make the flying scenarios more realistic, including mountain, cuboid, cylinder and triangular prism, and corresponding strategies are presented for unmanned aerial vehicles to safely avoid these obstacles. In the proposed algorithm, three novel teaching methods are introduced to balance the exploitation and exploration phases. Besides, a novel constrained optimization strategy is adopted, in which constraints are incrementally added to the fitness function to avoid the premature phenomenon in the initial iteration stage of algorithm. The experimental results show that compared with several state-of-the-art optimization algorithms, the proposed algorithm is significantly superior and can always generate the optimal flight route in complicated environments.