Collaborative Cross-Layer Design in Wireless Sensor Networks

Energy efficiency is a fundamental requirement in Wireless Sensor Networks (WSNs) due to its critical importance in a wide range of applications where recharging/replacing sensor batteries is unfeasible. In WSN applications such as field surveillance, environmental monitoring, disaster response, and traffic control, the cost of recharging batteries or replacing sensors may exceed the cost of deploying a new network. This led energy efficient operations to be the overarching goal of protocols employed at different layers of the network stack. In response, a plethora of research efforts have been pursued to achieve this goal, many of which utilize collaboration either between different layers of the network stack or between sensor nodes. In the literature, the former is known as Cross-Layer (CL) design and the latter is often referred to as Collaborative WSNs (CWSNs). Both CL design and CWSNs explore benefits of creating a new dimension of awareness by sharing information that is otherwise hidden and we refer to the former as intra-nodal collaboration and to the latter as inter-nodal collaboration. The benefits from such collaborations are mainly dependent on the context in which the new information is processed and utilized. Existing state of the art in this diverse area lack a unified design framework that allows seamless CL design for communication and processing operations in WSNs. Towards this end, we propose several energy-efficient collaborative CL schemes with various contexts of interest. First, we develop CL-MAC, a Cross-Layer Medium Access Control protocol for synchronous WSNs. CL-MAC takes advantage of routing layer information and inter-nodal collaboration in order to efficiently handle multi-packet, multi-hop and multi-flow traffic patterns while adapting to a wide range of traffic loads. CL-MAC's scheduling is based on a unique structure of Flow Setup Packets (FSPs) that efficiently utilize routing information to transmit multiple data packets over multiple multi-hop flows. Unlike other MAC protocols, supporting construction of multi-hop flows, CL-MAC considers all pending packets in the routing layer buffer and all flow setup requests from neighbors, when setting up a flow. This allows CL-MAC to make more informed scheduling decisions, reflecting the current network status, and dynamically optimize its scheduling mechanism accordingly. Second, we investigate supporting multi-hop and multi-packet routing in asynchronous WSNs in order to improve end-to-end latency and overall power consumption in such traffic patterns. We propose extending the knowledge of asynchronous MAC schemes to utilize a combination of routing information (intra-nodal) and duty-cycling information (inter-nodal) to temporarily synchronize nodes on the routing path between the transmitter and receiver. This idea can be applied to most asynchronous MAC schemes and was tested on RI-MAC (one of the recent energy efficient asynchronous MAC protocols). Third, we study collaboration between MAC, routing and application layers in order to eliminate communication redundancy and improve load balancing across the entire network via forming data-dependent virtual clusters. We propose a CL Dynamic Virtual Clustering-based Data Gathering technique called DVC-DG. DVC-DG integrates overhearing at the MAC layer with data being processed/communicated at upper layers (i.e., routing and application layers) to realize implicit virtual clusters and eliminate data redundancy. Unlike most existing clustering-based data gathering solutions, which employ an explicit data-independent clustering algorithm to choose cluster heads and members, DVC-DG eliminates the need for special cluster head election mechanisms and distributes common centralized cluster-head responsibilities (e.g., data collection and aggregation) over all cluster members and eliminates data redundancy at its very source rather than at the cluster-head. Finally, we propose a Cross-Layer Application-aware Paradigm (CLAP) that realizes and facilitates seamless collaboration across layers (intra-nodal) and between nodes (inter-nodal). CLAP allows any layer in the network stack to impact the behavior of other layers, according to the context of interest. The key is a new means of exchanging and processing CL information, called the Information-Layer (I-Layer). The I-Layer allows each layer of the network stack to publish its local information to be shared with other layers and subscribe other layers' shared information. Moreover, we augment CLAP into the SIDnet-SWANS simulator via a new API design which eliminates the need for hacking/bypassing conventional design hierarchies and simulator architectures. This greatly reduces the design and implementation complexities of CL protocols. Furthermore, to demonstrate CLAP’s unique capabilities, we utilize it to develop a sample CL protocol, which constantly monitors the application's current demands (i.e., contexts of interest) and re-configures underlying protocols accordingly.