Aerospace and Electronic Systems Magazine August 2016 - 9

Afonso et al. Figure 5. Mission plan in two-dimension (a) and in three-dimension (b), with the UAV prototype used for the flight tests. Gateway and Remote GCS module, such as IP addresses and the identification of the flight controller. The UAV Relay is used for the implementation of NAT traversal protocols, namely TURN. It can also be used for efficient IP multicast management between multiple vehicles and GCSs, with possible simultaneous radio connections. C C UAV Gateway behaves as an interface, intended to establish a connection between the local flight controller and the Remote GCS. This module, located in the vehicle, is responsible for NAT discovery procedures, UAV registration in the UAV Register module, verifying the connectivity to the network, and maintaining a continuous connection between the Remote GCS module and the flight controller. Remote GCS is the module that centralizes information from the vehicles, receiving telemetry and video streams. Similarly to the UAV Gateway module, it has several network functions, such as discovering the NAT and registering network information. Although designed with the aim of exploiting mobile radio networks, the proposed architecture is transparent to the access technology in the sense that any of the vehicles, through the UAV Gateway module, or in the case of the GCSs, through the Remote GCS AUGUST 2016 module, can use either 3G/4G, IEEE 802.15.4, Wi-Fi, a dedicated link, or even a wired connection (for a GCS). In fact, both the UAV Gateway and the Remote GCS can integrate multiple radio connections that can operate simultaneously in case of overlapping coverage and increase the reliability of the system. The possibility of dynamically choosing a communication link in order to improve reliability of UAV communications was also discussed in [14]. Although this architecture is presented in the scope of A2G communications, it can be extended for air-to-air communications between multiple UAVs, as this functionality can be implemented in the UAV Gateway. In this case, the UAV Gateway also acts as an interface between the ground radio network and an airborne network, which allows the adoption of a device-to-device (D2D) communication approach [15] that can extend the cellular coverage through multi-hopping as well as the support of coordinating and swarming functionalities between multiple UAVs [2]-[16]. EXPERIMENTAL SETUP AND RESULTS In order to assess the use of mobile networks, the UAS architecture proposed in Figure 4 was implemented and tested. Flight tests were performed according to the mission plan shown in Figure 5a, and using the hexacopter vehicle, shown in Figure 5b. The duration of each flight lasted around 5 minutes, covering a total path length of 1,760 m (4 laps of 440 m) with speeds varying between 1-8 m/s. The altitude ranged from 10 to 100 m above ground level, which are common values for civilian applications. In our implementation, we adopted the APM as the UAV flight controller and part of the experiment consisted in using the MAVlink msg_ping message to measure the round-trip time (RTT) at the Application Layer. A 4G smartphone was used for the implementation of the UAV Gateway. Three different wireless access technologies were tested, namely EDGE, HSPA+, and LTE, for two different situations: C C best case scenario, where the UAV and the GCS were connected to the same BS, communicating without a relay (Remote GCS with a public IP address and the vehicle behind a port restricted cone NAT); worst case scenario, where the UAV and the GCS were connected to different BSs and used the UAV Relay module to communicate. IEEE A&E SYSTEMS MAGAZINE 9

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine August 2016