Mohamed Hefeida

CL-MAC: A Cross-Layer MAC protocol for heterogeneous Wireless Sensor Networks

In duty cycled MAC protocols, multi-packet, multi-flow and multi-hop traffic patterns experience significant latencies, which are partially due to duty cycling. Several cross-layer routing/MAC schemes have been proposed to mitigate this latency. However, they utilize routing information from a single flow and/or a single packet perspective, thus limiting their adaptation to varying traffic loads and patterns. In this paper, we propose a novel Cross-Layer MAC protocol (CL-MAC) for WSNs, to efficiently handle multi-packet, multi-hop and multi-flow traffic patterns while adapting to a wide range of traffic loads. CL-MACs scheduling is based on a unique structure of flow setup packets 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. We evaluate CL-MAC through extensive ns-2 simulations and compare its performance to the state of the art, over various networks and for a wide variety of traffic loads and patterns. In all our experiments, CL-MAC substantially reduces end-to-end latency, increases delivery ratio while reducing the average energy consumed per packet delivered.

BulkMAC: a cross-layer based MAC protocol for wireless sensor networks

This paper presents a duty cycling based cross layer MAC protocol for wireless sensor networks (WSNs), referred to as BulkMAC, to support the transmission of multihop multiple packet flows during a single sleep period. We show that without the proposed cross-layered approach, the sensor nodes will spend significant energy and induce longer delays. The proposed protocol cleverly schedules the channel allocation using the upper routing layer information. We implement our protocol in ns2.29 and compare it against RMAC (Routing Enhanced MAC Protocol) and PRMAC (Pipelined-RMAC). On the average, BulkMAC improves data delivery by a factor of 2.26 and 1.67 compared to RMAC and PRMAC, respectively, for data collection in random networks.

Context modeling in collaborative sensor network applications

The design of context aware protocols in Wireless Sensor Networks (WSNs) is an emerging challenge. The interpretation of the sensed information greatly depends on the context and for efficient processing and communication, context awareness can play a major role. In most of the existing WSN protocols, context awareness is exploited in a single dimension and is captured either at the application layer or at the routing layer using a single context parameter. In this paper, we develop a new WSN context model to efficiently capture multiple context parameters in multiple dimensions (i.e. context from/to different layers of the network stack) and adjust the network behavior accordingly while simultaneously balancing the network load. The new model not only considers context parameters reflecting run-time application demands from a node, but also takes into consideration the current state of the node as well as the state and demands of neighboring nodes (inter-nodal context sharing). The new model: (a) represents context demands from each layer; (b) reflects the current individual state of each layer; (c) communicates (a) and (b) to the neighboring nodes to impact the decision process; and (d) locally distributes available resources aiming to achieve an optimal load balance. We show the application of this model in realizing a cross-layer optimized protocol for routing multi-hop and multi-packet traffic in WSNs.

Supporting multi-hop and multi-packet transmission in asynchronous WSNs

Cross-layer design approaches explore the benefits of information exchange between different layers of the network stack. In this paper, we present a cross-layer design approach to support multi-hop and multi-packet routing in asynchronous networks. This is achieved by realizing implicit temporary synchronization at the MAC layer while using multi-hop, multi-packet routing information, and current duty-cycling information. We formally present the cost of information sharing and maintenance at different levels of granularity across network layers and among neighboring nodes. The benefits of information sharing and communication are tremendous. The proposed cross-layer approach can be applied to any asynchronous MAC scheme. For example, when applied to RI-MAC (one of the most efficient and recent asynchronous MAC protocols), it reduces its latency by 34% and improves power consumption by 32% on a 16-node clique network, without compromising throughput. The proposed approach was also tested on a 10-node chain and a 25-node cross chain at various traffic loads and flows; it showed great improvements in latency without affecting power consumption or delivery ratio.

Interconnect wire length estimation for stochastic wiring distributions

This paper presents an improved stochastic wiring distribution model. The model is an improvement and modification of the models presented in some prior works available in literature. The proposed model shows 28% - 50% reduction in error when estimating the average interconnect wire length compared to existing models. This paper also investigates the effect of Rents exponent on the average wire length estimation. This investigation leads to the identification of some limitations of the approximations found in recent models.

A cross-layer approach for context-aware data gathering in Wireless Sensor Networks

Successful deployment of Wireless Sensor Networks (WSNs) depends on energy efficiency in computations and networking operations. Significant research efforts have been pursued over the last decade to realize techniques aimed at data gathering while prolonging network lifetime. Within data gathering applications, there is a class of applications that do not reconstruct the entire sensing field but mainly focus on monitoring and event/anomaly detection scenarios. In such applications, only representative data values (or distinct within a given threshold) are desired from different geographical (spatial) regions. Utilizing existing data gathering techniques in this type of applications yields communication-inefficient solutions, and therefore expensive in terms of energy cost. In this paper we investigate a cross-layer approach to reduce the number of communication operations in such specialized data gathering/monitoring applications. We explore the use of overhearing at the MAC layer in modifying the behavior of the application layer to realize Dynamic Virtual Clusters (DVCs). DVCs reduce data redundancy and dynamically distribute cluster head responsibilities, thus balancing the load and prolonging the network lifetime. Despite the cost of overhearing, our results show a reduction in the number of communication operations by a factor of up to N-1, where N is the number of nodes in a neighborhood.

Energy conservation in WSNs: A Collaborative Information Processing approach

We revisit the problem of data redundancy in Wireless Sensor Networks (WSNs) from a Collaborative Signal and Information Processing (CSIP) perspective. We propose an Information Processing and Communication Reduction (IPCR) scheme that utilizes spectrum sensing to detect and eliminate data redundancy. IPCR adapts its functionality according to data-space correlations and is independent of spatial and temporal field correlations. Its operation is also independent of the underlying Medium Access Control (MAC) scheme and does not require location information. Compared to spatial/temporal correlation-based signal and information processing techniques, IPCR can achieve up to N-fold reduction in communication complexity, where N is the number of nodes in a neighborhood.

Diamond frequency domain inter frame motion estimation for HEVC

This paper presents an efficient algorithm for motion estimation to reduce High Efficiency Video Coding (HEVC) standard encoding complexity. Phase correlation is initially utilized as a preprocessing step to indicate an approximation of the shift between coding units in the current frame and the reference frame. This is followed by a 9-point diamond search centered on the shift found in the initial step, in order to refine the best matching block. The proposed method has the potential to yield substantial improvements in terms of execution time and resulting video quality in comparison to the traditional search methods.

Data Aware Communication for Energy Harvesting Sensor Networks

We propose a Data Aware Communication Technique (DACT) that reduces energy consumption in Energy Harvesting Wireless Sensor Networks (EH-WSN). DACT takes advantage of the data correlation present in household EH-WSN applications to reduce communication overhead. It adapts its functionality according to correlations in data communicated over the EH-WSN and operates independently from spatial and temporal correlations without requiring location information. Our results show that DACT improves communication efficiency of sensor nodes and can help reduce idle energy consumption in an average-size home by up to 90 % as compared to spatial/temporal correlation-based communication techniques.

A Variable-Length Network Encoding Protocol for Big Genomic Data

Modern genomic studies utilize high-throughput instruments which can produce data at an astonishing rate. These big genomic datasets produced using next generation sequencing (NGS) machines can easily reach peta-scale level creating storage, analytic and transmission problems for large-scale system biology studies. Traditional networking protocols are oblivious to the data that is being transmitted and are designed for general purpose data transfer. In this paper we present a novel data-aware network transfer protocol to efficiently transfer big genomic data. Our protocol exploits the limited alphabet of DNA nucleotide and is developed over the hypertext transfer protocol (HTTP) framework. Our results show that proposed technique improves transmission up to 84 times when compared to normal HTTP encoding schemes. We also show that the performance of the resultant protocol (called VTTP) using a single machine is comparable to BitTorrent protocol used on 10 machines.