Classical versus reinforcement learning algorithms for unmanned aerial vehicle network communication and coverage path planning: A systematic literature review

The unmanned aerial vehicle network communication includes all points of interest during the coverage path planning. Coverage path planning in such networks is crucial for many applications, such as surveying, monitoring, and disaster management. Since the coverage path planning belongs to NP-hard issues, researchers in this domain are constantly looking for optimal solutions for this task. The speed, direction, altitude, environmental variations, and obstacles make coverage path planning more difficult. Researchers have proposed numerous algorithms regarding coverage path planning. In this study, we examined and discussed existing state-of-the-art coverage path planning algorithms. We divided the existing techniques into two core categories: Classical and reinforcement learning. The classical algorithms are further divided into subcategories due to the availability of considerable variations in this category. For each algorithm in both types, we examined the issues of mobility, altitude, and characteristics of known and unknown environments. We also discuss the optimality of different algorithms. At the end of each section, we discuss the existing research gaps and provide future insights to overcome those gaps.