Ambient Intelligence Context-Based Cross-Layer Design in Wireless Sensor Networks

Yang Liu, Boon‐Chong Seet, Adnan Al‐AnbukySensors 2014 · DOI

TL;DR

This paper introduces a novel cross-layer (CL) design framework for Wireless Sensor Networks (WSNs) that leverages Ambient Intelligence (AmI) user context to optimize network performance. By integrating AmI context into MAC and network layer protocols, the framework significantly improves throughput, packet delivery, delay, and energy efficiency in WSNs.

What Problem Does This Paper Address?

Wireless Sensor Networks (WSNs) are crucial for Ambient Intelligence (AmI) applications, enabling smart environments by collecting and processing sensor data to infer user context. However, existing WSN designs often overlook the rich context information available from AmI applications, which could be used to optimize network performance. Traditional cross-layer (CL) optimization in WSNs typically focuses on interactions between lower network layers (physical, MAC, network) and rarely incorporates application-level context. This limits the potential for WSNs to adapt intelligently to user needs and environmental situations, especially in energy-constrained WSNs that are vital for real-time monitoring. The paper aims to bridge this gap by proposing a framework that uses AmI context to drive CL optimization in WSNs.

Methods

The core of this work is a generic and customizable Cross-Layer (CL) framework designed to integrate AmI context into WSNs. This framework has three main components:

  1. Communication Mechanism: A data-centric publish/subscribe (pub/sub) paradigm is used. AmI systems act as publishers of inferred contexts, while WSN nodes are subscribers. A Virtual Broker (VB) entity, formed by a cooperative group of sensor nodes, facilitates context storage and retrieval.
  2. Node Architecture: Each WSN node is equipped with hardware (sensor, transceiver, storage) and software modules for pub/sub control, broker management, and context-aware CL optimization. This allows nodes to act as both publishers and subscribers, enabling them to receive AmI context and adapt their protocol parameters accordingly.
  3. Ontology-Based Context Modeling and Reasoning: An ontology model is used to standardize and formalize AmI context information (e.g., user location, activity, personal attributes). This model allows sensor nodes to interpret context and configure protocol parameters for desired performance.

As a use-case, the framework was applied to modify two existing protocols:

  • Dynamic Reconfiguration MAC (DR-MAC): The backoff behavior of this MAC protocol is adapted based on AmI context to prioritize high-priority sensor data.
  • Delay Aware AODV-Multipath (DAAM): The path selection mechanism of this routing protocol is modified to prioritize low-latency paths for high-priority data based on AmI context.

Datasets

Not discussed in the paper. The evaluation relies on simulations using the OPNET simulator.

Key Insights and Ideas

  • AmI Context is a Powerful Network Optimizer: The paper demonstrates that high-level AmI context, derived from user behavior and environment, can significantly enhance the performance of low-level WSN protocols. This moves beyond traditional CL optimization that relies solely on network-level metrics.
  • Bridging the Application-Network Divide: By creating a mechanism for WSN nodes to understand and act upon AmI context, the framework effectively bridges the gap between application requirements and network-layer operations.
  • Ontology as a Universal Context Language: The use of an ontology provides a standardized way to represent and reason about diverse AmI contexts, making the framework adaptable to various AmI scenarios.
  • Prioritization is Key: The core idea is to prioritize sensor data based on its importance to the AmI system, allowing WSN nodes to dynamically adjust their communication strategies. This prevents unnecessary transmissions and conserves energy.
  • Context-Awareness Benefits Both MAC and Network Layers: The study shows that integrating context-awareness at both the MAC and network layers leads to synergistic improvements, outperforming optimizations at a single layer.

Key Results

The proposed context-aware CL framework, when applied to DR-MAC and DAAM, showed substantial improvements:

  • Throughput: Context-aware DR-MAC achieved up to 36% improvement over CSMA/CA and 22% over original DR-MAC.
  • Packet Delivery Ratio (PDR): Context-aware DAAM achieved up to 26% improvement over AODV and 13% over original DAAM under normal traffic, and up to 44% and 18% respectively under high traffic.
  • End-to-End Delay: Context-aware DAAM reduced delay by up to 38% and 32% over AODV and original DAAM, respectively, under normal traffic.
  • Energy Efficiency: Context-aware protocols generally showed better energy efficiency, with context-aware DAAM reducing energy cost by 24% and 15% over AODV and original DAAM, respectively, under high traffic.
  • Overall Performance: The combined context-aware DR-MAC and context-aware DAAM protocol set (Set 9) consistently ranked highest across multiple metrics, demonstrating superior performance in throughput, PDR, and delay.

Limitations and Future Work

  • Limitations:
    • The paper focuses on a specific AmI scenario (intelligent event notification) and a limited set of WSN protocols.
  • The complexity of the ontology model needs to be balanced with the resource constraints of sensor nodes.
  • The simulation results are based on specific parameter settings and traffic conditions.
  • Future Work:
    • Extend the framework to the Internet of Things (IoT) paradigm, which involves a massive number of diverse networked devices.
  • Adapt the context-aware CL approach for optimizing IoT communications in AmI environments.
  • Explore more complex AmI scenarios and a wider range of WSN protocols.
  • Investigate the scalability of the framework with a larger number of nodes and higher network densities.

Summary updated

Abstract

By exchanging information directly between non-adjacent protocol layers, cross-layer (CL) interaction can significantly improve and optimize network performances such as energy efficiency and delay. This is particularly important for wireless sensor networks (WSNs) where sensor devices are energy-constrained and deployed for real-time monitoring applications. Existing CL schemes mainly exploit information exchange between physical, medium access control (MAC), and routing layers, with only a handful involving application layer. For the first time, we proposed a framework for CL optimization based on user context of ambient intelligence (AmI) application and an ontology-based context modeling and reasoning mechanism. We applied the proposed framework to jointly optimize MAC and network (NET) layer protocols for WSNs. Extensive evaluations show that the resulting optimization through context awareness and CL interaction for both MAC and NET layer protocols can yield substantial improvements in terms of throughput, packet delivery, delay, and energy performances.
BibTeX
@inproceedings{liu2014ambient,
  author = {Yang Liu and Boon‐Chong Seet and Adnan Al‐Anbuky},
  title = {{Ambient Intelligence Context-Based Cross-Layer Design in Wireless Sensor Networks}},
  booktitle = {Sensors},
  year = {2014},
  doi = {10.3390/s141019057},
}
Figure