INTELLIGENT LEARNING DIVERSITY MECHANISM FOR UNMANNED AERIAL VEHICLES APPLICATIONS

Abstract

The increased use of drones and aerial vehicles in applications poses challenges of airspace safety for aviation organizations. It is important to ensure the safety of the airspace when a significant number of unmanned aerial vehicles are deployed by civilian users. A solution that meets this requirement is important to promote innovation in the commercialization of air space for civilian users deploying unmanned aerial vehicle. The discussion in this paper proposes a mechanism that uses artificial intelligence to address this challenge. The proposed mechanism utilizes a low altitude platform (LAP) and entities in terrestrial wireless networks. The low altitude platform (LAP) observes, develops insights and training data (with human aid). The training data is used to develop learning mechanisms which determine the suitable unmanned aerial vehicles flight parameters in different scenarios.  The use of the LAP reduces the burden of communicating with terrestrial base stations. The unmanned aerial vehicles have a reduced altitude between the LAPs in comparison to terrestrial base stations. This reduces the free space path loss and rain-induced attenuation.  The performance benefit of the proposed mechanism in comparison to existing solution is examined via MATLAB simulations. Evaluation shows that the proposed mechanism reduces the network access costs by up to 90% on average. The proposed mechanism also increases available flight power and improves airspace safety by 37.3% and up to 53.2% on average respectively.

http://dx.doi.org/10.4314/njt.v39i2.22