Following the rise of the Internet of Things (IoT), the market of IP-based video surveillance systems has been growing exponentially. Wireless Video Sensor Networks (WVSNs), based on IEEE 802.11 (Wi-Fi) mesh networks, are a suitable and widely employed solution to build such systems. Despite their flexibility, coverage area, easiness of installation and low cost, these WVSNs suffer from three major problems: poor performance in transmissions, throughput unfairness and energy inefficiency. While the first two compromises the scalability of the system, the energy inefficiency, besides reducing the system lifetime when the nodes are battery powered, also thwarts the current trend of green networks, as well as the attempt of reducing the ever-increasing Internet’s carbon footprint.
The majority of the solutions found on the state of the art to address the above-mentioned problems are proposed for the general context of Wireless Mesh Networks (WMNs), not being suitable for the particular and more demanding case of WVSNs. Moreover, few solutions combine approaches to tackle all these problems at once, normally they only focus on a single problem at a time.
Therefore, the goal of this thesis is to develop and implement an all-around solution in order to obtain a fully functional prototype of a WVSN with better performance, throughput fairness, energy-efficiency and scalability. This solution consists on a scheduling mechanism that uses the radio data system (RDS), in the FM band, as an always-on, point-to-multipoint control channel to turn off the nodes’ Wi-Fi radio interfaces whenever they are not needed to transmit, receive or relay data, as proposed in [2].
To conclude this work, the prototype built will be exhaustively tested, allowing the characterization of the implemented solution behavior in real-world scenarios, the quantification of the improvements attained and a comparison with the state of the art solutions.
Motivation
Within a video monitoring system, an increase in the number of surveillance cameras to monitor a certain location demands the network to be both flexible and capable of supporting numerous flows of video transmitted towards a monitoring center where the video is stored, processed and analyzed. Due to their flexibility, low cost and large bandwidth, the IEEE 802.11 wireless ad hoc networks are an attractive solution for this problem.
Nevertheless, there are three major problems with multi-hop wireless networks: poor performance in transmissions, throughput unfairness, and energy inefficiency. While the first two problems compromise the scalability of the system, the energy inefficiency, besides reducing the system lifetime when the nodes are battery powered, also thwarts the current trend of green networks, as well as the attempt to reduce the ever-increasing Internet’s carbon footprint.
The performance problem is related to the medium access control mechanism used by Wi-Fi – the Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) – which was designed for single-hop networks and cannot efficiently avoid collisions within systems operating in ad hoc mode.
The multi-hop nature of the network is responsible for the throughput unfairness. In a multi-hop scenario, nodes closer to the gateway monopolize the medium and cause other nodes to starve [12].
Finally, the energy inefficiency problem is related to the high energy consumption of Wi-Fi interfaces, allied with the high rate of frame collisions and the subsequent retransmission process.
Objectives
The main objectives of this work are the following:
- Adaptation of the FM RDS into a reliable communication channel, for control purposes;
- Improvement of the proposed FM-WiFIX control mechanism to be applied on real video monitoring scenarios;
- Design and development of a proof-of-concept prototype;
- Characterization of the solution’s behavior in real operation scenarios.
The successful deployment of the proposed prototype, using the FM-WiFIX solution, is expected to significantly improve WVSNs’ needs for large bandwidths, tight QoS levels and energy efficiency, addressing the lack of suitable solutions found in the state of the art. Furthermore, the control channel developed, based on the RDS, could have much more applications since it is frequently useful to employ an out-of-band control channel on distributed systems.
