Dejene Ejigu

Pattern recognition and knowledge discovery from road traffic accident data in Ethiopia: Implications for improving road safety

This research tries to view accident data collection and analysis as a system that requires a special view towards understanding the whole and making sense out of it for improved decision making in the effort of reducing the problem of road safety. Under the umbrella of an information architecture research for road safety in developing countries, the objective of this machine learning experimental research is to explore and predict the role of road users on possible injury risks. The research employed Classification and Adaptive Regression Trees (CART) and RandomForest approaches. To identify relevant patterns and illustrate the performance of the techniques for the road safety domain, road accident data collected from Addis Ababa Traffic Office is exposed to many sided analyses. Empirical results showed that the models could classify accidents with promising accuracy.

A balanced battery usage routing protocol to maximize network lifetime of MANET based on AODV

Energy efficiency is a critical issue for battery-powered mobile devices in ad hoc networks. Failure of node or link allows re-routing and establishing a new path from source to destination which creates extra energy consumption of nodes, sparse network connectivity and a more likelihood occurrences of network partition. Routing based on energy related parameters is one of the important solutions to extend the lifetime of the network. In this paper, we are designing and evaluating a novel energy aware routing protocol called a balanced battery usage routing protocol (BBU) which uses residual energy, hop count and energy threshold as a cost metric to maximize network life time and distribute energy consumption of Mobile Ad hoc Network (MANET) based on Ad hoc on-demand Distance Vector (AODV). The new protocol is simulated using Network Simulator-2.34 and comparisons are made to analyze its performance based on network lifetime, delivery ratio, normalized routing overhead, standard deviation of residual energy of all Nodes and average end to end delay for different network scenarios. The results show that the new energy aware algorithm makes the network active for longer interval of time once it is established and fairly distribute energy consumption across nodes on the network.

Learning the Classification of Traffic Accident Types

This paper presents an application of evolutionary fuzzy classifier design to a road accident data analysis. A fuzzy classifier evolved by the genetic programming was used to learn the labeling of data in a real world road accident data set. The symbolic classifier was inspected in order to select important features and the relations among them. Selected features provide a feedback for traffic management authorities that can exploit the knowledge to improve road safety and mitigate the severity of traffic accidents.

Mining traffic accident features by evolutionary fuzzy rules

Traffic accidents represent a major problem threatening peoples lives, health, and property. Traffic behavior and driving in particular is a social and cultural phenomenon that exhibits significant differences across countries and regions. Therefore, traffic models developed in one country might not be suitable for other countries. Similarly, attributes of importance, dependencies, and patterns found in data describing traffic in one region might not be valid for other regions. All this makes traffic accident analysis and modelling a task suitable for data mining and machine learning approaches that develop models based on actual real-world data. In this study, we investigate a data set describing traffic accidents in Ethiopia and use a machine learning method based on artificial evolution and fuzzy systems to mine symbolic description of selected features of the data set.

Delegation Theory in the Design of Cross-Platform User Interfaces

The amalgamation of various technologies to support the needs of new computing models has become prevalent in computing environments like ubiquitous computing. Amalgamation means here heterogeneity caused by not only the coexistence of various devices in the same computing environment, but also the diversity between software, users as well as interaction modalities. The platform heterogeneity together with additional needs of interaction modalities and the proliferation of new technologies pose unique challenges for user interface (UI) designers and developers. We consider the problem of heterogeneity as a demand of collaboration between platforms (device and system) that are owned or controlled by a human user. Hence, we drive the concept of delegation to be implemented in a peer-to-peer model, where one peer (known as delegator) delegates another peer (known as delegatee) to run a UI (or a single interaction-modality) on its behalf. Thus, the delegatee uses its own capabilities to present the required UI or interaction-modality.

Towards green networking: Gossip based balanced battery usage routing protocol to minimize energy consumption of MANETs

Energy conservation is a critical issue in battery powered mobile nodes of mobile ad hoc networks (MANETs). Most MANETs routing protocols use some form of flooding to discover routes among mobile nodes. Despite various optimizations, many route discovery messages are still propagated without considering a coalesce effect of node density and residual energy. During route discovery process, each node of MANET should not blindly broadcast because malfunction of node or link might occur and allows establishing a new path from source to destination which creates extra energy consumption of nodes, sparse network density and a more likelihood occurrence of network partition. In this paper we are developing and evaluating an energy aware routing protocol called a gossip based balanced battery usage routing protocol (GBBU) which integrates minimum residual energy and node degree as cost metric to minimize and distribute energy consumption of MANETs based on Ad hoc on demand distance vector (AODV). The performance of the protocol is measured based on reachability, energy consumed per packet delivered, delivery ratio, average end-to-end delay, and network lifetime using network simulator-2.35. The simulation results show that GBBU routing protocol minimizes energy consumption per packet and fairly distribute energy usage across mobile nodes.

A Generic Layered Architecture for Context Aware Applications

Abstract As pervasive computing is at its preliminary stage a number of computing solutions were proposed and more are still under experimentation. In this paper a four layered generic architecture is proposed with a recommended components so as to support context awareness for pervasive applications. Detail of each layer is described sufficiently to insight the process of context generation from low level contextual data up to high level context reporting scheme in a given pervasive environment. The implementation section of this article describes how the context reasoning component employee different level of reasoning techniques such as knowledge acquisition, context based rule execution and the application of ontology in pervasive computing.

A novel road safety information architecture (RSIA): An enterprise view

This research tries to bring in an enterprise view to road safety information management towards understanding the whole and making sense out of it for improved decision-making in the effort of reducing the problem of road safety. Under the umbrella of an information architecture research for road safety, the objective of this research is to identify the role of collision related factors to the severity of an accident in explaining road safety situations and define or construct road safety information architecture based on existing enterprise frameworks. The research followed a Design Science research paradigm and Zachman Enterprise Framework was used to guide the development of the road safety information architecture. The research also employed classification techniques using Decision tree and Rule induction approaches. Results showed that the architectural representation guided by the selected framework can provide a holistic view to the management of road safety data. Moreover the identified patterns in a form of rule can supplement the previous severity prediction experiments with promising performance.

Novel range-free immune to radio range difference (IRRD) geo-localization algorithm in wireless networks

This paper presents a novel range-free immune to radio range difference (IRRD) geo-localization algorithm in wireless networks. The algorithm does not require the traditional assumption of anchor (location aware) nodes that have the same communication range as it works with anchor nodes having homogeneous and/or heterogeneous communication ranges. It is rang-free - it utilizes node connectivity to estimate the position of unknown (location unaware) nodes using two or more anchor nodes. The algorithm works in two steps: in the first step, the True Intersection Points (TIPs) forming the vertices of the smallest communication overlap polygon (SCOP) of the anchor nodes are found. In the second step, it estimates the position of the unknown node at the center of the SCOP which is formed from these TIPs. The problem is first geometrically and mathematically modeled, then new localization approach that does not assume anchor nodes have the same radio range is proposed.

Knowledge discovery from road traffic accident data in Ethiopia: Data quality, ensembling and trend analysis for improving road safety

a sel y Abstract: Descriptive analysis of the magnitude and situation of road safety in general and road accidents in particular is important, but understanding of data quality, factors related with dangerous situations and various interesting patterns in data is of even greater importance. Under the umbrella of information architec- ture research for road safety in developing countries, the objective of this machine learning experimental research is to explore data quality issues, analyze trends and predict the role of road users on possible injury risks. The research employed TreeNet, Classification and Adaptive Regression Trees (CART), Random Forest (RF) and hybrid ensemble approach. To identify relevant patterns and illustrate the performance of the techniques for the road safety domain, road accident data collected from Addis Ababa Traffic Office is subject to several analyses. Empiri- cal results illustrate that data quality is a major problem that needs architectural guideline and the prototype models could classify accidents with promising accu- racy. In addition, an ensemble technique proves to be better in terms of predictive accuracy in the domain under study.