Adaptive wireless access environment in transport solutions

This paper presents new attitude to adaptive control of the multi-path and multi-technology technology wireless access communications solutions. Presented approach is response on transport processes management requirements to provide complex seamless communication services in the selection of different modes coverage and guaranteed quality. Such solutions are related to communications of the vehicle to infrastructure as well as vehicle to vehicle. Such communications networking must include inter-vehicle communication interconnect, as well. CALM based system represent relevant data routing/switching with vertical RM OSI compatible communications architecture, however, with horizontal hierarchical management structure. Genuine decision processes in defined communications CALM routing/switching structures are, however, in literature discusses quite rarely. Decision processes measurable i.e. precisely enumerated system requirements are qualified by the performance. Based on application analysis performance indicators are demarcated into the acceptable tolerance ranges. In this paper we propose decision processes based on Bayes statistics as “classical” alternative to e.g. concept of Policy-based Management (PBM) traditionally and widely applied within the IP based networks. In proposed alternative measured data are processed by Kalman filters to separate reasonable part of noise and to predict individual parameters behavior in its near future. Than selftrained classification algorithm processes filtered measured data which are combined with deterministic parameters like the services economy, company policy etc. Training data block represent parameters vectors time line extended by correct decisions, i.e. assignment to the appropriate class. Even though these tools are designed specifically for the multi-path communications system, such principles are open for future penetration into whole telematic system adaptive management. These results were obtained within projects e-Ident, DOTEK and SRATVU which are elaborating results of project CAMNA.