Approximating Travel Times Using Opportunistic Networking

Abstract

Determining the time and means by which to travel from location A to location B for a person utilising a MANET requires the dissemination of both realtime and historic data. In this paper, we describe an approach to collecting and analysing the time costs related to geographic tracks, utilising opportunistic networking, local knowledge approximation and commonality of query. We call this approach the Short-Pass-Feedback (SPF) protocol. Peers record their own tracks, recording the times between known waypoints. Tracks and the durations of walks are shared between peers to approximate and update the local time-cost knowledge of a community of local peers. SPF seeks to maximise the sharing of common travel data as it exists between individuals in such geographical settings. We present an application for opportunistic networking, show and outline the algorithms used to aquire context and geographical time-cost data, analyse the benefits and pitfalls of a distributed approach through simulation and finally discuss how this knowledge can be used by middleware or applications for the provision of mobile services. We also discuss that such an approach is feasible and can provide significantly useful information to a community of peers.