Predicting the movements of a mobile phone

Predicting the movements of mobile cell phone communications systems, thus providing a guarantee of the quality of service (QoS) and offering locating services, as well as recognising imperfect texts on a dictionary basis, are two of the applications demonstrated by José Javier Astrain Escola, engineer in the Public University of Navarre. The tool that makes this task possible is one that employs fuzzy state automatas with empty chain transitions.

At the beginning of the study, this engineer found that there was very little literature on the subject: “As regards the theory fuzzy state automatas with empty chain transitions, there was no work in the literature. Regarding the classification of chains, they were presented as the most common techniques, as probabilistic types although in this PhD, we have shown that fuzzy techniques enable improved final recognition rates than these”.

It is in the mobile cellular communications systems that, “currently great efforts are being made to provide locating mechanisms, quality of service and recognition of the mobility of users with latest–generation systems such as UMTS”.

Thus, José Javier Astrain’s PhD puts forward “a tool that enables the measurement of the similarity between chains, working with imperfect chains of symbols. In this way, we have presented a fuzzy automata which gives rise to various families of values as a function of the parameters employed by said autómata.”

This new automata works with fuzzy state sets and fuzzy symbols chains. The automata was validated “by means of a series of experiments that enabled us to compare the results obtained with other techniques proposed in the literature”.

In this context, the most relevant contribution of the work is “the capacity to work with errors of edition and the handling by the automata of imperfect decisions inherited from the previous stages of classification”.

On terminating the investigation, José Javier Astrain underlined the improvements achieved, “the deformed fuzzy state automata achieving recognition rates of 99% when the observed texts show error rates of 80%”.

Improving service quality

Once the capacity of the fuzzy automatas and deformed fuzzy are shown to tackle the problems of the imperfect recognition of symbols, “we proposed taking on board an industrial problem of great current interest, the application of techniques developed for the management of quality of service in mobile communications systems”.

As regards QoS management in mobile cell phone communications systems, “we offer a model of mobility based on chains of symbols with three different types of coding. The chains of symbols which represent the displacement of the mobile terminal throughout its movements are obtained by the mobile by the triangulation of adjacent base stations. These chains are compared by the fuzzy automata of the terminal with those of a dictionary that represent the most frequent routes within the cell in which the terminal is found. On determining the route followed, this makes it possible for the terminal to initiate the reserve of resources in the target cell with sufficient advance warning thanks to the prediction of location calculated by the automata”.

Thus, “a tool is proposed that enables the satisfactory resolution of these needs at a low computational cost, an interchange of messages between base stations and terminals and that can integrate itself into new designs of these types of networks”.