Wireless sensor networks [1] are composed of small devices, called sensor nodes, which cooperate to forward collected data to a sink node that either uses the data locally or forwards it to other networks through a gateway, like for example, the Internet. Sensors are resource-limited devices composed of sensing, processing, transceiver, and power units.
The addition of low cost multimedia hardware to sensors fostered the development of Wireless Multimedia Sensor Networks [2], allowing the retrieval of multimedia streams, and/or scalar sensor data. Wireless sensor networks have many application areas [2] such as multimedia surveillance sensor networks, storage of potentially relevant activities, traffic avoidance, enforcement and control system, and many more.
Routing protocols [3] in WSNs can be classified according to the network structure, protocol operation, how routing information is acquired and maintained. In terms of network structure, routing protocols can be divided into flat-based routing, hierarchical-based routing and location-based routing. In flat-based routing, typically nodes have similar roles, whereas in hierarchical-based routing nodes have different roles. In location-based routing, location information is used to route data in the network. According to protocol operation, these protocols can be classified as multipath-based, query-based, negotiation-based, QoS-based, or coherent-based routing techniques. In multipath-based routing, multiple paths are maintained between a source-destination pair. In query-based routing, the destination node sends a query through the network and the node with this data, sends an answer. In negotiation-based routing, high level data descriptors are used to eliminate redundant data transmissions through negotiation. In QoS-based routing, certain QoS metrics have to be satisfied while routing data through the network. In coherent-based routing, sensors cooperate in processing data flooded throughout the network. According to how routing information is acquired and maintained, they can be classified into proactive, reactive, and hybrid. In proactive protocols, nodes compute routes before they are needed. In reactive protocols, nodes compute route on demand. Hybrid protocols combines ideas of both.
Multimedia applications have different QoS requirements such as, bounded latency or delay, throughput, jitter, availability, and energy consumption. Since energy efficiency is considered as the main goal of most WSNs routing protocols, the majority of these protocols does not perform well when applied to QoS-constrained WMSN. Routing techniques in WMSN can be classified similarly to those of WSNs. In [4] another categorization for WMSN routing protocols is presented. Protocols are classified based on the handled data types, data delivery model types, classes of algorithms adopted, and the used hole-bypassing approach.
Many routing schemes [4] and [5] have been proposed to address QoS requirements. In most of these schemes, only one of the desired objectives is optimized, while others are assumed as problems’ constraints [6]. In certain applications, a meta-heuristic approach [7] and [8] using a multi-objective optimization (MO) algorithms that can provide several optimal solutions may be preferred, since single design objective algorithms ignore other relevant objectives. By considering all objectives simultaneously, a set of optimal solutions can be generated, also known as the Pareto solutions [9] of the multi-objective problem. It is also known from [10] that finding optimal routes for multiple objectives in networks (multi-constrained QoS routing), is a NP-complete problem, hence efficient heuristic search algorithms based on reduced-complexity Evolutionary Algorithms (EAs) [11] are necessary.
The Expected Transmission Count (ETX) [12] metric is an estimation of the expected total number of transmissions (including retransmissions) required to deliver a packet to the destination node successfully. ETX allows finding high throughput paths on a multi-hop wireless network, and incorporates the effects of link loss ratios, asymmetry in the loss ratios between the two directions of each link, and the interference among the successive links of a path.
This paper proposes a new multi-objective approach for the WMSN routing problem that takes into account QoS parameters such as delay and ETX. A comparison of the proposed approach with two alternative routing protocols was also presented.