A small group of Unmanned Aerial Vehicles (UAV), each equipped with a communications payload, offers a possible means of providing broadband services over disaster regions. The UAVs are power limited so the number of mobile sub-scribers that can be supported by each UAV depends on its proximity to clusters of mobiles. One way of maximising the total number of mobiles supported within the available RF power is to periodically relocate each of the UAVs in response to the movement of the mobiles. This paper compares two approaches for optimally locating the UAVs. One approach employs a non cooperative game (NCG) as the mechanism to plan the next flying strategies for the group. The other uses evolutionary algorithms (EA) to evolve flying manoeuvres in a collaborative manner. Exemplar comparison results show that although both approaches are able to provide sufficient network coverage adaptively, they exhibit different flying behaviours in terms of flightpath, separation and convergence time. The non cooperative game is found to fly all aerial vehicles in a similar, balanced and conservative way, whilst the evolutionary algorithms enable the emergence of flexible and specialised flying behaviours for each member in the flying group which converge faster to a sufficient global solution.
|Title of host publication||2016 International Conference on Unmanned Aircraft Systems, ICUAS 2016|
|Number of pages||9|
|Publication status||Published - 30 Jun 2016|
|Event||2016 International Conference on Unmanned Aircraft Systems, ICUAS 2016 - Arlington, United States|
Duration: 7 Jun 2016 → 10 Jun 2016
|Conference||2016 International Conference on Unmanned Aircraft Systems, ICUAS 2016|
|Period||7/06/16 → 10/06/16|