View Full Document

The Influence Mobility Model - A Novel Hierarchical Mobility Modeling Framework



View the full content.
View Full Document
View Full Document

38 views

Unformatted text preview:

The Influence Mobility Model A Novel Hierarchical Mobility Modeling Framework Muhammad U Ilyas1 2 and Hayder Radha1 1 Department of Electrical Computer Engineering 2120 Engineering Building Michigan State University East Lansing MI 48824 USA ilyasmuh radha egr msu edu Abstract Practical mobile ad hoc systems include heterogeneous classes of mobile nodes such as people and vehicles The simultaneous presence and movement of these classes influence the mobility pattern of their members in a variety of random and deterministic manners In addition there is an inherent hierarchical and multi resolution structure to the mobility patterns In this paper we present a novel hierarchical mobility framework which incorporates the fact that the movement of certain classes of mobile nodes e g people is affected by their surroundings and the movement of other forms of mobile nodes The proposed model takes into account that mobile nodes movement is neither completely random nor a mere function of their routing decision and or the trip s source and destination points The proposed model is neither classified as microscopic nor macroscopic Instead it is categorized as a multi scale mobility framework capable of modeling mobility scenarios of different scales i e it is equally capable of modeling the movement of mobile nodes in a street intersection as it is to model the movement of nodes between a number of population centers in a state The proposed mobility framework presented in this paper integrates and extends an influence model and hierarchical graph based representations of the mobility area to achieve this goal Our simulation results show that this framework accurately captures the influences among different groups of mobile nodes and the constraints imposed by their surroundings Keywords Simulations Mobility Modeling Ad hoc Network Simulation Stochastic processes Markov Chains Influence Model Graph theory I INTRODUCTION Mobility models play a crucial role in the design development and implementation of traditional mobile systems and emerging ad hoc networks Consequently a variety of mobility models have been proposed and analyzed by many previous efforts see for example 1 3 These models have many strengths and advantages however they fail to capture some aspects of practical mobile scenarios In particular emerging ad hoc systems include heterogeneous classes of mobile nodes such as people and vehicles that interact and influence each other s mobility in a hierarchical and multi resolution manner In this paper we propose a novel mobility model that we refer to as the hierarchical influence mobility model HIMM The proposed model is a novel mobility modeling framework that integrates and extends crucial aspects of an influence model 8 and a graph theoretic representation of the simulation area The HIMM addresses the effect the movement of one group of mobile nodes has on the movement of other groups while the latter graph based representation restricts 2 Department of Computer Science School of Arts Sciences Lahore University of Management Sciences Lahore Punjab 54792 Pakistan muilyas lums edu pk movement due to the surroundings and resources available to the mobile nodes The framework s flexibility allows us to apply it to mobility scenarios of different scales i e it is possible to apply the same mobility model to successively more detailed views of the same simulation or apply it to entirely different mobility problems spanning geographical areas of widely varying magnitudes The rest of the paper is organized as follows Section II is a list of shortcomings of previous mobility models that are addressed by the HIMM Section III details the HIMM by explaining the graph based representation of the simulation plane and the application of the binary influence model Section IV describes how the proposed HIMM model is applied to two different scenarios The first scenario is a pedestrian crossing across a busy one way street on which traffic is controlled by a traffic signal The second example models the movement of travelers using the interstate highway network and air links within a large state These two scenarios illustrate the ability of HIMM to capture the influences among heterogeneous nodes at different scales of mobility Section V concludes the paper II PREVIOUS WORK In this section we provide a brief overview of previously proposed mobility models and their shortcomings that our HIMM addresses 1 Task Based Movement A notion of tasks and starting and destination points People do not move around at random Their movement is usually task based Kumar et al make use of this feature in their model which they derived from the gravity model in 1 Most of our movement is restricted to neighboring places i e moving around in an office or around the house It is rarely the case that once a person reaches a location he she immediately moves on to another distant location The longer the distance traveled to reach a location the longer the stay If a person moves from one room to another he she may stay there for only a few seconds minutes or a few hours but if a person moves from one part of the city to another it is safe to assume that he she will not stay at his her destination for just a few minutes but stay for a few hours Similarly if a person moves from San Francisco to New York we can again assume that he will not just stay for a few hours but probably for a few days We believe that there exists a correlation between the distance traveled between two points and the time that a person spends in the vicinity of his destination 0 7803 8966 2 05 20 00 C 2005 IEEE 2 Path Selection People do not tend to walk around aimlessly and hope to reach their destination by chance People select routes to their destination on the basis of spatial path congestion cost distance and travel time We believe that this aspect of characterizing movement may be reducible to a routing problem for which there already exists a well established theory of routing algorithms 3 Mobile Node Classification It is possible to perform classification of mobile nodes into classes based on some features such as their location and speed Nodes in the same class move together along similar paths e g pedestrians and cyclists have a tendency to stay on the sidewalk and only rarely mix with fast driving cars on the road and vice versa 4 Class Transition A mobile node may switch classes i e a person descending from a car and joining the stream of


Access the best Study Guides, Lecture Notes and Practice Exams

Loading Unlocking...
Login

Join to view The Influence Mobility Model - A Novel Hierarchical Mobility Modeling Framework and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view The Influence Mobility Model - A Novel Hierarchical Mobility Modeling Framework and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?