A non-intrusive technique for asserting user identity in pervasive computing environments

Pervasive computing environments are an exciting new computing scene due to therecent advances in wireless communications and digital electronics. Theseenvironments are small smart worlds that have many advantages, such ascontinuous and distributed interaction with people using a variety of wirelessdevices, which are ubiquitous and sharable, to provide useful services. In addition,each user has wide interaction with a huge number of entities. It would beimpractical to require people to authenticate themselves every time they crossvarious network boundaries, as the frequent authentication process would disruptthe users' normal activities. The use of frequent authentication contradicts theobjectives of pervasive computing in creating seamless environments and deliveringdistributed services. Due to the characteristics of these environments there is achallenge in dealing with asserting user identity. As a result, the need for a systemto apply a non-intrusive authentication technique to assert user identity across theenvironment, and for every interaction within these environments, becomes one ofthe main challenges in these environments.This research aims to develop a technique for verifying user identity wheninteracting with such environments. User identity information is used by the systemin pervasive environments to retrieve a variety of contextual information, which isnecessary for providing the required services. In doing so, a new approach forasserting user identity that is non-intrusive and adaptable is developed. Theproposed approach, called Non-Intrusive Identity Assertion System (MAS), wouldbe aware of the intentions of the user and authenticate her/him continuouslythroughout the day to maintain confidence in user identity. It is assumed that usersin smart environments carry identification devices which are used by NIAS todetect when users attempt to access resources. Although these devices provide abasic mechanism for identifying users, they are not sufficient to assert users'identity, for they could easily be picked up by other people. Therefore, NIASattempts to assert users' identity by monitoring a minimum amount of theiractivities instead of extensive tracking. The system then uses these activities to inferuser events. User events inference is achieved using an unsupervised clusteringtechnique, which gathers a group of user activities and creates an event. Then eachuser event has to be converted to a numeric value and passed to the system to assertuser identity.The system should be able to cope with various situations, such as users losing theirsmart tags or people impersonating other people and raise an alarm to block anintruder from being asserted as a legitimate user. NIAS provides an alternative tothe intrusive periodic user authentication process and, at the same time, minimizethe risk of false identity in pervasive computing environments.

Level dc:type.qualificationlevel

Doctoral (Level 8)

Year dc:date.issued

2012

Author dc:creator

• AI-Karkhi, AA

Language dc:language

en

Identifier

oai:salford-repository.worktribe.com:1338372

OAI identifier oai:identifier

oai:salford-repository.worktribe.com:1338372