QoS and QoE Aware Routing Protocol for Flying Ad-Hoc Surveillance Networks Using Fuzzy Inference Systems

— The Flying Ad-hoc Network (FANET) networks have often operated in regions that are difficult to reach and have no fixed infrastructure. The use of devices such as Unmanned Aerial Vehicles (UAV) enable aerial networks to be created that are extremely fast, although there are no specific routing protocols for this type of network that can make communication more efficient among these devices. For this reason, this paper sets out a specific routing protocol for FANET networks that is designed for the discovery of routes among UAV devices. Evidence of the benefits of this strategy has been obtained through simulation by the Network Simulator version 2.


I. INTRODUCTION
The traditional concept of Ad-Hoc Networks has been adapted to the creation of a new concept called the Flying Ad-Hoc Network (FANET). In this new context, FANETs consist of Unmanned Aerial Vehicles (UAVs) for the creation of networks in local areas where it is difficult to access terrestrial UAVs, particularly after natural disasters. UAVs are responsible for monitoring a certain area by capturing images that are sent to a base station on the ground [1].
One of the main challenges in FANETs is how to position the UAVs in a suitable way to enable them to monitor the region. The positioning of the UAVs is of strategic importance for the network as a means of establishing a connection among them. Moreover, owing to their mobility (UAV´s ability to move in different directions and speeds), the devices might either draw closer or keep a distance from one another and hence risk impairing communication among the UAVs. The degree of mobility is also an important factor since the UAVs might either fly too slowly or quickly, which would also jeopardize the connection [2].
Another important factor is the flight autonomy of the UAVs, since most of the devices have an average flight time of 30 minutes [3], [4]. A UAV with a low flight autonomy will have to stop being QoS and QoE Aware Routing Protocol for Flying Ad-Hoc Surveillance Networks Using Fuzzy Inference Systems Evaluate traditional routing protocol in FANETs [5] No Mechanism for data routing based on localization The position for UAV´s device [6] No Extension of OLSR protocol called P-OLSR Compare P-OLSR with OLSR [7] No Energy -efficiency algorithm.
To reduce the energy consumption [8] No Survey of the routing protocols Improve the routing performance [9] No There is now a type of network configuration that has evolved from the concept of Mobile Ad-Hoc Networks (MANET) and Vehicular Ad-Hoc Networks (VANET) as well as from Wireless Sensor Networks. This type of configuration has led to an increased number of resources and is constantly growing, especially in the production of new devices and systems that are complex and able to move and fly autonomously.
In FANET, the devices are generally referred to as Unmanned Aerial Vehicles (UAVs In extensive coverage areas, it may be impracticable to establish direct communication from the UAVs to the base station on the ground at certain times. However, this problem can be overcome through hop-by-hop communication, which requires the use of a routing protocol to discover the best route/path from the original source to the final destination [13].

A. Problem Statement and Major Contributions
One of the main problems in this type of network is to determine and maintain the routes, since the mobility of the UAVs can cause changes in the topology. For this reason, the main focus of this paper is to set out a specific routing protocol for FANET networks that can accomplish the task more efficiently [14].
The routing protocols are responsible for finding, establishing, and maintaining routes between two nodes that wish to communicate with each other. These protocols must generate the minimum possible overhead and the bandwidth consumed by them must also be small.
A routing protocol that targets FANET networks is more complex than fixed network protocols; this is due to several features of these networks such as their dynamic topology algorithm, mutual interference, restricted power and the limited resources available in the UAVs.
In FANET networks, if one UAV is not close enough to another to carry out the communication, it will have to make use of routing information to choose the best path. The communication among UAVs that is beyond the reach of transmission, is made in multiple hops through the collaboration of intermediary nodes; that is, the scope is not restricted to the radius of action of each device individually, but to the sum of the radius of action of all of the devices (Fig. 1). The mobility of UAVs and their spatial arrangement are also very important for determining the communication routes. As a result of the movement, these routes are usually re-arranged so that the interconnection between the UAVs can be continued. For this reason, the routing must be carried out dynamically by increasing the autonomy of the UAVs and reducing the delay in data delivery between a source node and a destination node [15].
Another main contribution of this paper is the adoption of a new communication network model used to provide connectivity in regions that are difficult to reach on land (especially regions after natural disasters). FANETs are easily established, as they are easily moved to a new region.
The frequent updating of the control information can ensure more accurate information; however, there is a need for a greater use of energy, since this reduces the autonomy of the nodes. FANETs are commonly employed to monitor regions by using sensors to capture images and/or videos Therefore, it is very important that the quality of the streamed video can be assessed using QoE metrics to ensure that, in fact, good data communication reflects a good user experience.
Therefore, this article performs a cross-layer evaluation involving the network and application layers to verify it.

IV. FUZZY ROUTING PROTOCOL SYSTEM
Fuzzy System allows the use of variables that are dynamic and imprecise, which makes it ideal for scenarios where values are often changing. As the values change, a new solution is found for the network. This paper examines three metrics for input: Mobility Level, Flight Autonomy and RSSI (These three entries are shared between UAV´s by routing protocol signaling packets).
In order to define the RSSI, mobility and flight autonomy intervals, the values imposed by the technology [3] were used as upper limit and, through simulation, adjustments were made in such a The ideal communication will be between drones that have a high flight autonomy, high RSSI and low mobility. This type of setting allows the route to be longer for the transmission. However, these conditions are not always possible, and in this case, the Fuzzy System will try to find out which route is closest to it.
The following table shows the main rules of the Fuzzy System which has an excellent or good route.
In these cases, the drone will select the route that will remain active for longer and hence achieve the best performance. The other situations that are not shown in the Table, correspond to TERRIBLE or REGULAR route that will rarely be chosen by the Fuzzy System, and then only in cases where GOOD or EXCELLENT routes are not available (Table II). The final decision is made in accordance with the inference value (resulting from the output of the Fuzzy System), and the highest inference value will be chosen for decision-making. During the implementation of the Fuzzy System, it was observed that the inference values equal to, or greater than, 0.6 represent the best routes and hence are very likely to be chosen. In the following graph, the yellow area represents the most likely routes to be chosen, since they correspond to the routes in which the drones have high flight autonomy, low mobility and high RSSI. In general, the part of the graph shaded in green, corresponds to the drones with medium mobility, medium RSSI and medium flight autonomy, and in this type of situation, there is little chance of the drone being chosen as a communication route. The blue region of the graph represents a drone with high mobility, low RSSI and low flight autonomy, and in this case the drone will not be chosen as the communication route (Fig. 3).
In general, for a UAV to select a transmission route, it transmits to the Fuzzy Inference machine the RSSI, mobility, and flight autonomy range information, received from neighboring UAVs. The Fuzzy Inference Machine, through the Fuzzy rule base, will generate an output value to evaluate the quality level of the route. The route with the highest rating and the highest outbound inference value will be chosen by the UAV to initiate transmission to the destination UAV (Fig. 4).   Table III. Due to random movements as well as the speed of the UAVs, they may fly in the same or in opposite directions. This aspect may make them approach or fly far from each other either faster or slower. These changes in the topology require a fast response from the routing protocols and if this does not occur, the network performance will be impaired.
The throughput graph shows the best performance of the Fuzzy Adaptive Protocol compared with other protocols. Owing to the topology variations, AODV and OLSR protocols were unable to update the routes quickly and efficiently, and as a result, none of the protocols was able to keep the route active. For this reason, both protocols break the transmission during part of the simulation. The Fuzzy Adaptive Protocol was able to update the route quickly and efficiently, while at the same time always keeping the route active and preventing any break in the connection (Fig. 5).  The best performance of the proposed protocol can be found when the throughput averages are compared. The OLSR protocol had an average throughput rate of 0.26 Mbps, the AODV protocol had an average throughput rate of 0.45Mbps, while the Adapted Fuzzy protocol had an average throughput rate of 0.72Mbps. The adapted fuzzy protocol had a 176% gain over OLSR and a 60% gain over AODV. The following graph shows the average throughput rates for each protocol (Fig. 6). The throughput directly interferes with the number of frames received by the user and hence affects the quality of the video. The interruption of the throughput during a certain period of time leads to a reduction in the number of frames that are normally received. As mentioned earlier, the video had 2000 frames. Figure 7 shows that the OLSR protocol received 400 frames, the AODV protocol received 1400 frames and the Fuzzy Adaptive protocol received 2000 frames. Metric (VQM). The PSNR evaluates the signal-noise of the video with regard to the following features: brightness, noise and color. In the following graph the OLSR protocol had a PSNR average of 18dB (considered to be a bad video), the AODV protocol had an average of 27dB ( a regular video) and the Fuzzy Adaptive protocol had an average of 41dB (and thus is thought to be an excellent video) (Fig.8).  The VQM metric measures the following features: color, brightness, intensity, and frame distortion.
It was also used to evaluate the video received by the user compared with the original video. The range of the VQM metric is on a scale of 0 to 5; in this case the value closer to 0 means a better video quality, that is, the distortion is less than in the original video. The OLSR protocol had a VQM average of 4.1 (i.e. a low-quality video), the AODV protocol had an average of 3.3 (a regular video) and the Fuzzy Adaptive protocol had an average of 1.4 (a video of excellent quality). Graph X displays the VQM values for each protocol (Fig. 10). In video simulation, the protocol achieved a 127% gain over the OLSR protocol and a 51% gain over the AODV protocol. In addition to evaluating the video received through the Quality of Experience metrics, this paper has also made a visual evaluation of the received video frames using the MSU Video Quality Measurement Tool Software [19]. On the basis of the visual evaluation, it can be stated that the Fuzzy Adaptive System had a better performance than the other protocols. Figure 11 makes it possible to analyze the frame of the video transmitted through the OLSR Protocol. In this case, the frame is quite distorted, and there are several defects since it is a frame of low quality. With regard to frame of the video transmitted through the AODV Protocol, it is clear that there is an improvement in quality when compared with the previous frame. However, the frame still has minor distortions and defects. Although it is not a poor-quality frame, it cannot be concluded that it is an excellent frame. The video has thus been classified as having a regular quality (Fig. 12). When the frame of the video transmitted with the Fuzzy Adaptive Protocol is analyzed, it can be seen that there is a great improvement of quality with regard to the other two protocols. In Figure 13 below, as the frame has no distortions or defects like the previous frames, the video transmitted with the proposed protocol is rated as being of excellent quality.

VI. CONCLUSION
The aerial networks have been the object of a good deal of research owing to the ease of creating and implementing them. Although the FANETs networks provide a number of benefits (as mentioned earlier) there are also a number of challenging tasks that have to be carried out.
One of these is finding the most efficient communication system among the UAVs by means of a routing protocol. For this reason, this paper has recommended a specific routing protocol for FANET networks which involves using a Fuzzy System to improve the route discovery process, while taking into account the RSSI, mobility level and in particular, the flight autonomy.
The proposed routing protocol was compared with the traditional Ad-Hoc routing protocols: AODV and OLSR. The comparison was carried out by means of traditional QoS and QoE metrics and the proposed routing protocol achieved a better performance (around 35%) than the other two routing protocols referred to.
In future work the authors intend to employ new artificial intelligence techniques, and include new parameters for decision making, as well as making use of new wireless technologies and new propagation models.