Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.733248
Title: Micro-scale monitoring in an urban area using vehicular sensor network
Author: Milojevic, Milica
ISNI:       0000 0004 6496 9865
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
One of the emerging problems facing populated urban areas represents high level of the air pollution. Current monitoring approaches assume processing of pollution data in a centralised manner, however due to external factors (e.g. current atmospheric conditions) the air pollution level can have fast fluctuations inside a street which might lead to disparity of the pollution levels in neighbouring streets. This thesis proposes a decentralised monitoring framework that enables harvesting of the pollution data, its processing, and dissemination of early warnings to the Static Monitoring Units (SMU) when the onsets of hazardous air pollution concentration are detected at the street (micro-scale) level. The proposed air pollution monitoring framework relies on a Vehicular Sensor Network infrastructure where vehicles have limited on-board resources, such as processing, battery power and storage. Three aspects of the micro-scale air pollution monitoring have been investigated. The Decentralised data Dissemination and Harvesting (DDH) is a single-hop dissemination mechanism which enables the nodes to decide whether they should harvest the data based on their current position and movement direction. This mechanism identifies the streets that nodes need to monitor by comparing the amount of stored pollution data with the one harvested from other nodes in the network. The Decentralised Data Fusion (DDF) algorithm is developed to fuse the delayed pollution data that arrives from other nodes in the network. The algorithm uses the Delayed State Information Filter to refine sensor measurements and applies an interpolation technique to interpolate the missing data in the time gaps which occur as a consequence of receiving delayed data. The proposed algorithm augments the pollution data history at the slight cost of data accuracy. The Decentralised Dissemination of Warnings (DDW) is a beaconless, multihop dissemination mechanism designed to relay early warnings to the SMUs. Mobile nodes calculate the time interval (waiting time) that they need to wait before they rebroadcast the warning message. The waiting time is calculated by taking into account the distance between sending and receiving nodes, and nodes’ distances to the SMUs. The DDW mechanism ensures that high number of non-duplicated warnings is received at the collection points (SMUs). The proposed framework enables efficient utilisation of node’s on-board resources in terms of transmitting and processing activity, and data storage. It reduces the number of mobile nodes required for monitoring purposes without losing the volume of the relevant collected pollution data. Using uncontrolled mobile nodes and their mobility, the framework enables reporting the hazardous pollution levels in near real-time.
Supervisor: Barria, Javier Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.733248  DOI:
Share: