Improving the LBS QoS through Implementation of QoS Negotiation Algorithm Renato Filjar, Lidija Bušić, Petar Pikija Ericsson Nikola Tesla d. d. Krapinska 45, 10000 Zagreb, Croatia Telephone: +385 1 365 3098, Fax: + 385 1 365 35 48 E-mails: {renato.filjar, lidija.busic, petar.pikija}@ericsson.com Abstract - Location-based Services (LBS) utilise location awareness for delivery of useful information/content to the user. When considering the Quality of Service (QoS) for LBS, few parameters defined in standards can offer the basic insight into the actual quality of provided service. Usually, the pattern is to strive to the best values for each defined parameter. It is however unnecessary, both from technological and economical point of view, to demand an excellent QoS performance for every LBS. It is much more efficient to use the service and user preferences and construct the appropriate level of QoS for LBS in question. This paper introduces the negotiation algorithm for improving the LBS QoS. Proposed algorithm makes it possible to determine the most suitable level of QoS for particular LBS. Practical validation of the algorithm is also presented, based on the developed prototype. I. INTRODUCTION Location-based services have finally emerged as a very prosperous group of telecommunication services [1]. Aimed to provide content and services based on user’s location (knowledge of the user’s place in a physical space enhanced by information about geospatial relations with the other neighbouring objects), it is the appropriate establishment of Quality of Service (QoS) for LBS that drives the business success [2]. Industrial standards [3] have determined the LBS QoS parameters and current common practice has aimed for obtaining as best QoS as possible, regardless of the actual and common needs of particular LBS. Such an approach usually overloads network resources without benefit for either service provider, content provider, network operator or end-user. Here the improvement in the LBS QoS is presented, achieved through a special LBS QoS Negotiation Algorithm (NA) implementation. The introduction of the LBS QoS NA reduces usually overstretched requirements on positioning performance as the foundation of the LBS QoS, thus providing the most suitable QoS level in accordance with requirements of invoked LBS and the user personal profile. Overall, proposed LBS QoS NA provides efficient utilisation of all resources involved in LBS provision with firm guarantees on the quality of the LBS provided. II. LBS QUALITY OF SERVICE As any other group of telecommunication services, LBS provide services with a certain level of quality [2]. International industrial standards [3] define the fundamental parameters of LBS QoS, as presented in Table 1. Although these parameters do not cover all aspects of LBS QoS and thus do not provide the complete LBS description, their international acceptance allows for at least provisional distinction between various location- based services based on their quality. TABLE I LBS QoS PARAMETERS LBS QoS PARAMETERS Horizontal position determination accuracy Vertical position determination accuracy Response time In practice, achieving the requested level of LBS QoS [3] means the implementation of a particular position estimation method. Existing position determination methods can be outlined in two essential groups: - (telecom) network-based position determination methods - terminal-based position determination methods. Network-based position determination methods [3] are based on utilisation of signals and hardware installations of public mobile communication networks. The network is responsible for performing the appropriate measurements and calculation of the user’s position estimate. User’s equipment (mobile phones) serve in position determination process as auxiliary and assistance devices, supplying common control signals to the network. Common network- based position determination methods are [3]: Cell Global Identity (CGI) method and Extended Observed Time Difference (E-OTD) method. In principle, network-based position determination methods are robust and widely available (availability is determined by network coverage), but not accurate (~150 m, at their best) [3].