Presenter

Ms Thi Kieu Ha Phung - ETRO dept. and Hanoi University of Science and Technology

Abstract

In the last decade, wireless sensor networks (WSNs) have evolved from the idea that a large amount of small wireless sensors can be used to collect information from the physical environment and transport the data wirelessly to a collection point. Initially, WSNs have been used for applications with low-rate data collection. Recently, wireless sensor networks, enabling the connection between the physical world and the digital world, have become an important component of the Internet of Things (IoT). Based on the IoT architecture, the information captured by WSNs is available to the user through the internet. Several applications in the IoT (e.g. health care monitoring, tracking, industrial monitoring) require the efficient and timely collection of larger amounts of data. Due to the limited bandwidth of the radios on the sensor nodes, the interference and contention over the wireless medium, the limited processing capacity and energy resource of the battery-operated devices, fulfilling the newly established data-collection requirements, becomes a challenging task.
In this research, we exploit the multi-channel operation capability of the sensor radios to provide the higher network throughput while keeping energy consumption low. Efficient scheduling parallel transmissions on different channels can eliminate collision, idle-listening or over-hearing, which are consequences of non-coordinated transmissions. Our work focuses on scheduling the regular traffic that is periodically transmitted and at adapting the schedule to additional traffic that can be requested at some point in time.
First, we analyze factors, which influence network throughput, including interference and contention of the wireless medium and the half-duplex nature of radio transceivers. Based on that, we propose a joint model of transmission schedule and route selection to partially exploit the routing topology in multi-channel coordination for increasing the network throughput. The model applies the schedule-based approach and the channel-hopping scheme to organize multi-channel coordination in order to achieve energy-efficient and reliable transmission.
Second, we propose a scheduling algorithm to produce a collision-free channel coordination and traffic-adaptive schedule for every sensor node. The algorithm based on Reinforcement Learning (an AI technique) operates in a distributed manner to make the nodes learn how they should use the shared wireless medium to efficiently delivery their data. The results of extensive simulation experiments show that the algorithm can achieve collision-free parallel transmissions over different channels to provide high throughput and high delivery ratio while meeting the crucial energy efficiency requirement.
Finally, we present the complete design for deploying the schedule-based multi-channel protocol in real applications. The design includes a communication procedure that is necessary for sensor nodes to communicate among them in order to form a network and to propagate sensed data to the collection point. Furthermore, we propose a low-overhead time synchronization scheme that is critical for a schedule-based protocol. We demonstrate the feasibility of the protocol and the time synchronization scheme on a laboratory-scaled test-bed.

Short CV

Lecturer, School of Electronics and Telecommunications Hanoi University of Science and Technology, Vietnam