Browsing by Author "Bagula, Antoine"

Now showing 1 - 19 of 19
  • Thumbnail Image
    Item
    Open Access
    Computational intelligent systems : evolving dynamic Bayesian networks
    (2009) Osunmakinde, Isaac Olusegun; Bagula, Antoine
    In this thesis, a new class of temporal probabilistic modelling, called evolving dynamic Bayesian networks (EDBN), is proposed and demonstrated to make technology easier so as to accommodate both experts and non-experts, such as industrial practitioners, decision-makers, researchers, etc. Dynamic Bayesian Networks (DBNs) are ideally suited to achieve situation awareness, in which elements in the environment must be perceived within a volume of time and space, their meaning understood, and their status predicted in the near future. The use of Dynamic Bayesian Networks in achieving situation awareness has been poorly explored in current research efforts. This research completely evolves DBNs automatically from any environment captured as multivariate time series (MTS) which minimizes the approximations and mitigates the challenges of choice of models. This potentially accommodates both highly skilled users and non-expert practitioners, and attracts diverse real-world application areas for DBNs. The architecture of our EDBN uses a combined strategy as it resolves two orthogonal issues to address the challenging problems: (1) evolving DBNs in the absence of domain experts and (2) mitigating computational intensity (or NP-hard) problems with economic scalability. Most notably, the major contributions of this thesis are as follows: the development of a new class of temporal probabilistic modeling (EDBN), whose architecture facilitates the demonstration of its emergent situation awareness (ESA) and emergent future situation awareness (EFSA) technologies. The ESA and its variant reveal hidden patterns over current and future time steps respectively. Among other contributions are the development and integration of an economic scalable framework called dynamic memory management in adaptive learning (DMMAL) into the architecture of the EDBN to emerge such network models from environments captured as massive datasets; the design of configurable agent actuators; adaptive operators; representative partitioning algorithms which facilitate the scalability framework; formal development and optimization of genetic algorithm (GA) to emerge optimal Bayesian networks from datasets, with emphasis on backtracking avoidance; and diverse applications of EDBN technologies such as business intelligence, revealing trends of insulin dose to medical patients, water quality management, project profitability analysis, sensor networks, etc.
  • Thumbnail Image
    Item
    Open Access
    Context-aware handoff support for wireless access networks
    (2010) Mokhesi, Lekometsa; Bagula, Antoine
    The phenomenal emergence of several heterogeneous wireless networks and technologies has allowed users to have access to IP services anywhere, at anytime, from any network and with whatever terminal they use. This computing platform has been driven by the rapid evolution of mobile devices that are equipped with multiple network interfaces and the development of IP based applications.
  • Thumbnail Image
    Item
    Open Access
    Cooperative Behaviours with Swarm Intelligence in Multirobot Systems for Safety Inspections in Underground Terrains
    (2014) Yinka-Banjo, Chika; Osunmakinde, Isaac O; Bagula, Antoine
    Underground mining operations are carried out in hazardous environments. To prevent disasters from occurring, as often as they do in underground mines, and to prevent safety routine checkers from disasters during safety inspection checks, multirobots are suggested to do the job of safety inspection rather than human beings and single robots. Multirobots are preferred because the inspection task will be done in the minimum amount of time. This paper proposes a cooperative behaviour for a multirobot system (MRS) to achieve a preentry safety inspection in underground terrains. A hybrid QLACS swarm intelligent model based on Q-Learning (QL) and the Ant Colony System (ACS) was proposed to achieve this cooperative behaviour in MRS. The intelligent model was developed by harnessing the strengths of both QL and ACS algorithms. The ACS optimizes the routes used for each robot while the QL algorithm enhances the cooperation between the autonomous robots. A description of a communicating variation within the QLACS model for cooperative behavioural purposes is presented. The performance of the algorithms in terms of without communication, with communication, computation time, path costs, and the number of robots used was evaluated by using a simulation approach. Simulation results show achieved cooperative behaviour between robots.
  • Thumbnail Image
    Item
    Open Access
    Design and optimisation of a low cost Cognitive Mesh Network
    (2016) Nleya, Sindiso M; Keet, Maria; Bagula, Antoine
    Wireless Mesh Networks (WMNs) have been touted as the most promising wireless technology in providing high-bandwidth Internet access to rural, remote and under-served areas, with relatively lower investment cost as compared to traditional access networks. WMNs structurally comprise of mesh routers and mesh clients. Furthermore, WMNs have an envisaged ability to provide a heterogeneous network system that integrates wireless technologies such as IEEE 802.22 WRAN, IEEE 802.16 WiMAX, IEEE 802.11 Wi-Fi, Blue-tooth etc. The recent proliferation of new devices on the market such as smart phones and, tablets, and the growing number of resource hungry applications has placed a serious strain on spectrum availability which gives rise to the spectrum scarcity problem. The spectrum scarcity problem essentially results in increased spectrum prices that hamper the growth and efficient performance of WMNs as well as subsequent transformation of WMN into the envisaged next generation networks. Recent developments in TV white space communications technology and the emergence of Cognitive radio devices that facilitate Dynamic Spectrum Access (DSA) have provided an opportunity to mitigate the spectrum scarcity problem. To solve the scarcity problem, this thesis reconsiders the classical Network Engineering (NE) and Traffic Engineering (TE) problems to objectively design a low cost Cognitive Mesh network that promotes efficient resources utilization and thereby achieve better Quality of Service (QoS) levels.
  • Thumbnail Image
    Item
    Open Access
    The design of an intelligent parking system using wireless sensor networks and multi-protocol label switching
    (2009) Mwebaze, Anthony; Bagula, Antoine
    The challenge of parking management has increasingly posed the need for smart solutions. Motorists in today's busy world seek the best option in locating available parking points. The need for an efficient parking system stems from increased congestion, motor vehicle pollution, driver frustration and fatigue to mention but a few. This study was conducted at a time when the world was experiencing a financial crisis and more than ever motorists needed intelligent parking systems to reduce the cost of gas spent driving around to find parking. Indeed, the time spent driving around would be beneficial if used to do work that would put one at an advantage in the credit recession. The study was also conducted at a time when South Africa was preparing to host the 2010 soccer world cup. In the preparation to manage motor vehicle congestion, this study was a viable solution to manage the expected challenge of parking. This study presents the design and illustrates the performance of an intelligent parking system based on an integrated architecture where (1) Wireless Sensor networks (WSNs) using Small Programmable Object Technology (SPOT) motes are launched into parking places to monitor the activity of the parking area through light intensity sensing and (2) the sensed information is gathered and channeled through a gateway into databases used for parking space visualization and information dissemination over the World Wide Web technology and mobile devices via a Multi Protocol label Switching (MPLS) network. Using an illustrative simulation model of a small parking system built around a new generation of SUNspot motes, the study demonstrates how a real life smart parking iv system can be deployed to benefit motorists in today's busy World and serves as a foundation to future work on how this emerging generation of motes can be used to provide better ways of finding parking.
  • Thumbnail Image
    Item
    Open Access
    Development of cooperative behavioural model for autonomous multi-robots system deployed to underground mines
    (2015) Yinka-Banjo, Chika Ogochukwu; Bagula, Antoine; Osunmakinde, Isaac Olusegun
    The number of disasters that occur in underground mine environments monthly all over the world cannot be ignored. Some of these disasters for instance are roof-falls; explosions, toxic gas inhalation, in-mine vehicle accidents, etc. can cause fatalities and/or disabilities. However, when such accidents happen during mining operations, rescuers find it difficult to respond to it immediately. This creates the necessity to bridge the gap between the lives of miners and the product acquired from the underground mines by using multi-robot systems. This thesis proposes an autonomous multi-robot cooperative behavioural model that can help to guide multi-robots in pre-entry safety inspection of underground mines. A hybrid swarm intelligent model termed, QLACS, that is based on Q-Learning (QL) and the Ant Colony System (ACS) is proposed to achieve cooperative behaviour in a MRS. The intelligent model was developed by harnessing the strengths of both QL and ACS algorithms. The ACS is used to optimize the routes used for each robot while the QL algorithm is used to enhance cooperation among the autonomous robots. The communication within the QLACS model for cooperative behavioural purposes is varied. The performance of the algorithms in terms of communication was evaluated by using a simulation approach. An investigation is conducted on the evaluation/scalability of the model using the different numbers of robots. Simulation results show that the methods proposed in this thesis achieved cooperative behaviour among the robots better than state-of-the-art or other common approaches. Using time and memory consumption as performance metrics, the results reveal that the proposed model can guide two, three and up to four robots to achieve efficient cooperative inspection behaviour in underground terrains.
  • Thumbnail Image
    Item
    Open Access
    Drivable region detection for autonomous robots applied to South African underground mining
    (2012) Falola, Omowunmi Elizabeth; Bagula, Antoine
    This dissertation focuses on enhancing autonomous robots' capability to identify drivable regions in underground terrains. A system model that compares the drivability analysis of underground terrains using the entropy model and statistical region merging (SRM) was developed, with a view to presenting an analysis of 2D and 3D results. The approach involves standard image-processing techniques, such as colour and texture feature extraction and region segmentation for underground image classification. A probabilistic method based on the local entropy was employed. The entropy is measured within a fixed window on each frame in order to compute features used in the segmentation process. This research compares the results obtained from the entropy method and SRM approach. Performance evaluation is carried out to provide useful qualitative and quantitative conclusions.
  • Thumbnail Image
    Item
    Open Access
    An energy-efficient routing protocol for Hybrid-RFID Sensor Network
    (2011) Boneke Nokonoko, Dulce-Maira; Dlodlo, Mqhele E; Bagula, Antoine
    Radio Frequency Identification (RFID) systems facilitate detection and identification of objects that are not easily detectable or distinguishable. However, they do not provide information about the condition of the objects they detect. Wireless sensor networks (WSNs), on the other hand provide information about the condition of the objects as well as the environment. The integration of these two technologies results in a new type of smart network where RFID-based components are combined with sensors. This research proposes an integration technique that combines conventional wireless sensor nodes, sensor-tags, hybrid RFID-sensor nodes and a base station into a smart network named Hybrid RFID-Sensor Network (HRSN).
  • Thumbnail Image
    Item
    Open Access
    An interactive mobile lecturing model: enhancing student engagement with face-to-face sessions
    (IGI Global, 2013) Boyinbode, Olutayo; Ng'ambi, Dick; Bagula, Antoine
    Although use of podcasts and vodcasts are increasingly becoming popular in higher education, their use is usually unidirectional and therefore replicates the transmission mode of traditional face-to-face lectures. In this paper, the authors propose a tool, MOBILect, a mobile lecturing tool that enables users to comment on lecture vodcasts using mobile devices, and aggregated comments become an educational resource. The vodcasts are generated through Opencast Matterhorn and YouTube. The tool was evaluated at the University of Cape Town with students’ own devices. The paper reports on the architecture of the MOBILect, its framework for student-vodcast interaction, and evaluation results. The paper concludes that the MOBILect has potential for use as a supplement to the traditional face-to-face lectures especially in scenarios of large classes, or where the medium of instruction is not the students’ mother tongue.
  • Thumbnail Image
    Item
    Open Access
    Internet of things : least interference beaconing algorithms
    (2014) Tuyishimire, Emmanuel; Bagula, Antoine; Sanders, J W
    The emerging sensor networking applications are predicting the deployment of sensor devices in thousands of computing elements into multi-technology and multi-protocol platforms. Access to information will be available not only anytime and anywhere, but also using anything in a first-mile of the Internet referred to as the internet-of-things (IoT). The management of such a large-scale and heterogeneous network, would benefit from some of the traditional IP-based network management techniques such as load and energy balancing, which can be re-factored to achieve efficient routing of sensor network traffic. Research has shown that minimizing the path interference on nodes was necessary to improve traffic engineering in connection oriented networks. The same principle has been applied in past research in the context of the IoT to reveal that the least interference beaconing protocol (LIBP); a protocol derived from the least interference beaconing algorithm (LIBA) outperforms the Collection Tree Protocol (CTP) and Tiny OS Beaconing (ToB) protocol, in terms of energy efficiency and lifetime of the sensor network. However for the purpose of efficiency and accuracy, it is relevant, useful and critical to revisit or reexamine the LIBA algorithm in terms of correctness and investigate potential avenues for improvement. The main contributions of this research work are threefold. Firstly, we build upon formal methods to verify the correctness of the main principles underlying the LIBA, in terms of energy efficiency and interference minimization. The interference is here defined at each node by the number of routing paths carrying the sensor readings from the motes to the sink of the network that traverse the node. Our findings reveal the limitations in LIBA. Secondly, building upon these limitations, we propose two improvements to the algorithm: an algorithm called LIBA+ that improves the algorithm performance by keeping track of the energy usage of the sensor nodes, and a multi-sink version of the algorithm called LIBAMN that extends the algorithm to account for multiple sinks or gateways. These enhancements present preventive mechanisms to include in IoT platforms in order to improve traffic engineering, the security of network protocols and network stability. Lastly, we present analytical results, which reveal that the LIBA algorithm can be improved by more than 84% in terms of energy balancing. These results reveal that formal methods remain essential in the evaluation and performance improvement of wireless sensor network algorithms and protocols.
  • Thumbnail Image
    Item
    Open Access
    ITIKI: Bridge between African indigenous knowledge and modern science on drought prediction
    (2012) Masinde, Euphraith Muthoni; Bagula, Antoine; Muthama, Nzioka
    The now more rampant and severe droughts have become synonymous with Sub-Saharan Africa; they are a major contributor to the acute food insecurity in the Region. Though this scenario may be replicated in other regions in the globe, the uniqueness of the problem in Sub-Saharan Africa is to be found in the ineffectiveness of the drought monitoring and predicting tools in use in these countries. Here, resource-challenged National Meteorological Services are tasked with drought monitoring responsibility. The main form of forecasts is the Seasonal Climate Forecasts whose utilisation by small-scale farmers is below par; they instead consult their Indigenous Knowledge Forecasts. This is partly because the earlier are too supply-driven, too ""coarse"" to have meaning at the local level and their dissemination channels are ineffective. Indigenous Knowledge Forecasts are under serious threat from events such as climate variations and ""modernisation""; blending it with the scientific forecasts can mitigate some of this. Conversely, incorporating Indigenous Knowledge Forecasts into the Seasonal Climate Forecasts will improve its relevance (cultural and local) and acceptability, hence boosting its utilisation among small-scale farmers. The advantages of such a mutual symbiosis relationship between these two forecasting systems can be accelerated using ICTs. This is the thrust of this research: a novel drought-monitoring and predicting solution that is designed to work within the unique context of small-scale farmers in Sub-Saharan Africa. The research started off by designing a novel integration framework that creates the much-needed bridge (itiki) between Indigenous Knowledge Forecasts and Seasonal Climate Forecasts. The Framework was then converted into a sustainable, relevant and acceptable Drought Early Warning System prototype that uses mobile phones as input/output devices and wireless sensor-based weather meters to complement the weather stations. This was then deployed in Mbeere and Bunyore regions in Kenya. The complexity of the resulting system was enormous and to ensure that these myriad parts worked together, artificial intelligence technologies were employed: artificial neural networks to develop forecast models with accuracies of 70% to 98% for lead-times of 1 day to 4 years; fuzzy logic to store and manipulate the holistic indigenous knowledge; and intelligent agents for linking the prototype modules.
  • Thumbnail Image
    Item
    Open Access
    Managing economic value and uncertainty on software projects : an empirical study with the CASSE framework
    (2009) Balikuddembe, Joseph Kibombo; Bagula, Antoine
    Lack of adaptive-predictor models in software development renders the decision-making process complex, principally when evaluating investment options. Prior work has presented various approaches which are still non-integrated, requiring rigour to embrace overall economic value and uncertainty management. In this study. a complex Adaptive Software Engineering Framework which uses the Actor objective Dependency Analysis (AOD) technique is proposed as a feasible option. It is aimed at providing a useful technique for monitoring and controlling value propositions in terms of cost, schedule, and progress of complex issues in software.
  • Thumbnail Image
    Item
    Open Access
    Multi-layered security in the Internet of the Things
    (2014) Ngqakaza,Lutando; Bagula, Antoine
    It is well discussed and understood that there is still a need for suitable security for the Internet of Things. It is however still not clear how existing or emerging security paradigms can be effectively applied to a network of constrained nodes in a lossy communications environment. This thesis provides a survey into what routing protocols can be used with network security in mind. What will also be discussed, is an implementation, that in conjunction which a robust routing protocol, can provide security for a network of constrained devices with a certain level of confidence. The implementation and design involves including communications encryption and centralized non-cryptographic methods for securing the network. This thesis basically explores the use of multiple security mechanisms in an Internet of Things environment by using Contiki OS as the platform of choice for simulations and testing.
  • Thumbnail Image
    Item
    Open Access
    Natural language interface to relational database: a simplified customization approach
    (2016) Mvumbi, Tresor; Keet, Maria; Bagula, Antoine
    Natural language interfaces to databases (NLIDB) allow end-users with no knowledge of a formal language like SQL to query databases. One of the main open problems currently investigated is the development of NLIDB systems that are easily portable across several domains. The present study focuses on the development and evaluation of methods allowing to simplify customization of NLIDB targeting relational databases without sacrificing coverage and accuracy. This goal is approached by the introduction of two authoring frameworks that aim to reduce the workload required to port a NLIDB to a new domain. The first authoring approach is called top-down; it assumes the existence of a corpus of unannotated natural language sample questions used to pre-harvest key lexical terms to simplify customization. The top-down approach further reduces the configuration workload by autoincluding the semantics for negative form of verbs, comparative and superlative forms of adjectives in the configuration model. The second authoring approach introduced is bottom-up; it explores the possibility of building a configuration model with no manual customization using the information from the database schema and an off-the-shelf dictionary. The evaluation of the prototype system with geo-query, a benchmark query corpus, has shown that the top-down approach significantly reduces the customization workload: 93% of the entries defining the meaning of verbs and adjectives which represents the hard work has been automatically generated by the system; only 26 straightforward mappings and 3 manual definitions of meaning were required for customization. The top-down approach answered correctly 74.5 % of the questions. The bottom-up approach, however, has correctly answered only 1/3 of the questions due to insufficient lexicon and missing semantics. The use of an external lexicon did not improve the system's accuracy. The bottom-up model has nevertheless correctly answered 3/4 of the 105 simple retrieval questions in the query corpus not requiring nesting. Therefore, the bottom-up approach can be useful to build an initial lightweight configuration model that can be incrementally refined by using the failed queries to train a topdown model for example. The experimental results for top-down suggest that it is indeed possible to construct a portable NLIDB that reduces the configuration effort while maintaining a decent coverage and accuracy.
  • Thumbnail Image
    Item
    Open Access
    A Novel Epidemic Model for the Interference Spread in the Internet of Things
    (2022-04-02) Tuyishimire, Emmanuel; Niyigena, Jean de Dieu; Tubanambazi, Fidèle Mweruli; Rutikanga, Justin Ushize; Gatabazi, Paul; Bagula, Antoine; Niyigaba, Emmanuel
    Due to the multi-technology advancements, internet of things (IoT) applications are in high demand to create smarter environments. Smart objects communicate by exchanging many messages, and this creates interference on receivers. Collection tree algorithms are applied to only reduce the nodes/paths’ interference but cannot fully handle the interference across the underlying IoT. This paper models and analyzes the interference spread in the IoT setting, where the collection tree routing algorithm is adopted. Node interference is treated as a real-life contamination of a disease, where individuals can migrate across compartments such as susceptible, attacked and replaced. The assumed typical collection tree routing model is the least interference beaconing algorithm (LIBA), and the dynamics of the interference spread is studied. The underlying network’s nodes are partitioned into groups of nodes which can affect each other and based on the partition property, the susceptible–attacked–replaced (SAR) model is proposed. To analyze the model, the system stability is studied, and the compartmental based trends are experimented in static, stochastic and predictive systems. The results shows that the dynamics of the system are dependent groups and all have points of convergence for static, stochastic and predictive systems.
  • Thumbnail Image
    Item
    Open Access
    Situation recognition using soft computing techniques
    (2012) Machaka, Pheeha; Bagula, Antoine
    The last decades have witnessed the emergence of a large number of devices pervasively launched into our daily lives as systems producing and collecting data from a variety of information sources to provide different services to different users via a variety of applications. These include infrastructure management, business process monitoring, crisis management and many other system-monitoring activities. Being processed in real-time, these information production/collection activities raise an interest for live performance monitoring, analysis and reporting, and call for data-mining methods in the recognition, prediction, reasoning and controlling of the performance of these systems by controlling changes in the system and/or deviations from normal operation. In recent years, soft computing methods and algorithms have been applied to data mining to identify patterns and provide new insight into data. This thesis revisits the issue of situation recognition for systems producing massive datasets by assessing the relevance of using soft computing techniques for finding hidden pattern in these systems.
  • Thumbnail Image
    Item
    Open Access
    Towards an interactive mobile lecturing model a higher-level engagement for enhancing learning
    (2013) Boyinbode, Olutayo Kehinde; Ng'ambi, Dick; Bagula, Antoine
    The use of mobile devices has grown in recent years and has overtaken the proliferation of desktop computers with their dual affordances of small size and easy connectivity in diverse fields. The usage of these devices has not been widespread in higher education. Mobile technology is a new and promising area of research in higher education. The affordance of mobile technologies has prompted their adoption as a means of enhancing face-to-face (f2f) learning. In this thesis, mobile lecturing is presented as a means of achieving mobile learning. The availability of mobile devices has positively enabled the mobile lecturing process. F2f lectures are recorded and distributed as lecture vodcasts using mobile devices. The vodcasts are generated through Opencast Matterhorn and YouTube. Currently, there are few descriptive models of mobile lecturing that can be used to enhance learning in Higher Education Institutions (HEIs). This thesis has several contributions: first I propose a “MOBLEC” theoretical model of mobile lecturing; mobile lecturing represents a new paradigm in mobile learning which enhances students’ engagement with lecture vodcasts to foster deep learning. The second contribution of this thesis is a mobile lecturing tool, MOBILect. MOBILect is developed in HTML5 for cross-platform solution across most mobile devices. This tool enables students to use mobile devices to actively interact with lecture vodcasts and with peers using the vodcast. Finally, I use different case studies to evaluate the MOBLEC model to explore the effectiveness of mobile lecturing in enhancing learning in HEIs. The MOBLEC model is proposed to define mobile lecturing: it describes mobile lecturing as a process resulting from the convergence of mobile technologies, learning engagements and learning interactions. The case studies are evaluative, relying on a group of students to evaluate the MOBLEC by accessing MOBILect. Empirical data was acquired through triangulation method involving focus group discussions, open-ended questions and interviews. All the questions were based on the MOBLEC model. The result of the studies provided positive indicators as to the usefulness and effectiveness of mobile lecturing in engaging students to enhance and foster deep learning. Mobile lecturing, through augmenting and accessing lecture vodcasts on students’ mobile devices anywhere and at any time, with an affordance to comment and respond to comments, has potential for empowering students who might be struggling to understand f2f sessions and the aggregated comments become a valuable educational resource. The thesis also outlines areas for future research work.
  • Thumbnail Image
    Item
    Open Access
    Trajectory planing for cooperating unmanned aerial vehicles in the IoT
    (2022-02-24) Tuyishimire, Emmanuel; Bagula, Antoine; Rekhis, Slim; Boudriga, Noureddine
    The use of Unmanned Aerial Vehicles (UAVs) in data transport has attracted a lot of attention and applications, as a modern traffic engineering technique used in data sensing, transport, and delivery to where infrastructure is available for its interpretation. Due to UAVs’ constraints such as limited power lifetime, it has been necessary to assist them with ground sensors to gather local data, which has to be transferred to UAVs upon visiting the sensors. The management of such ground sensor communication together with a team of flying UAVs constitutes an interesting data muling problem, which still deserves to be addressed and investigated. This paper revisits the issue of traffic engineering in Internet-of-Things (IoT) settings, to assess the relevance of using UAVs for the persistent collection of sensor readings from the sensor nodes located in an environment and their delivery to base stations where further processing is performed. We propose a persistent path planning and UAV allocation model, where a team of heterogeneous UAVs coming from various base stations are used to collect data from ground sensors and deliver the collected information to their closest base stations. This problem is mathematically formalised as a real-time constrained optimisation model, and proven to be NP-hard. The paper proposes a heuristic solution to the problem and evaluates its relative efficiency through performing experiments on both artificial and real sensors networks, using various scenarios of UAVs settings.
  • Thumbnail Image
    Item
    Open Access
    Ubiquitous mesh networking : application to mobile communication and information dissemination in a rural context
    (2014) Maliwatu, Richard; Bagula, Antoine
    ICT has furthered the social and economic development of societies but, rural African communities have lagged behind due to issues such as sparse population, low household income, a lack of electricity and other basic infrastructure that make it unattractive for telecommunication service providers to extend service provision. Where the service is available, ubiquitous service coverage has not translated into ubiquitous access for individuals because of the associated costs. A community-wide WMN offering VoIP using fixed telephone handsets has been deployed as a viable alternative to the cellular service provider. The effectiveness of this WMN VoIP service springs from the mobile phone usage statistics which showed that the majority of calls made are intra-community. This dissertation has been an effort towards improved communication and access to in- formation for the under-served communities. Key contributions include, mobile VoIP support, translation gateway deployment to make textual information accessible in voice form via the phone, IP-based radio for community information dissemination. The lack of electricity has been mitigated by the use of low-power devices. In order to circumvent the computational challenges posed by the processing and storage limitations of these devices, a decentralised system architecture whereby the processing and storage load are distributed across the mesh nodes has been proposed. High-performance equipment can be stationed at the closest possible place with electricity in the area and connectivity extended to the non-electrified areas using low-power mesh networking devices. Implementation techniques were investigated and performance parameters measured. The quality of service experienced by the user was assessed using objective methods and QoS correlation models. A MOS value of 4.29, i.e. very good, was achieved for the mobile VoIP call quality, with the underlying hardware supporting up to 15 point-to-point simultaneous calls using SIP and the G.711 based codec. Using the PEAQ algorithm to evaluate the IP-based radio, a PEAQ value of 4.15, i.e. good, was achieved. Streaming audio across the network reduces the available bandwidth by 8Kbps per client due to the unicast nature of streaming. Therefore, a multicast approach has been proposed for efficient bandwidth utilization. The quality of the text-to-voice service rendered by the translation gateway had a PESQ score of 1.6 i.e. poor. The poor performance can be attributed to the TTS engine implementation and also to the lack of robustness in the time-alignment module of the PESQ algorithm. The dissertation also proposes the use of the WMN infrastructure as a back-haul to isles of WSNs deployed in areas of interest to provide access to information about environ- mental variables useful in decision making.