Multisensorial modular system of monitoring and tracking with information fusion techniques and neural networks

The article presents a model of a monitoring system with high benefits, based on two complementary sensorial systems (images and acoustics), that allow the information received to fuse, and to prioritize the computational resources towards the zones with considered risks. The application uses components developed completely in the Multimedia Technology Group of the ETSI of Telecommunication in Valladolid (Spain), such us: a) active acoustic radar based on a phase array of microphones, with techniques of beam conformation and with multibeam detection schemes, and b) efficient compression algorithms and video transmission in real time. The system consists of an arbitrary number of sensors (monitoring probes), distributed by the monitoring space, a data network, and a central server (monitoring manager). The sensors consist of a microprocessor with its memory and three main modules: i) acoustic radar; ii) video capture card;, and iii) network card. The flexibility of the system allows new modules to be added such as: personal detector card; sound card; and cards for telecontrol, etc. The sensorial microprocessor preprocesses the information received by the modules, packs and transmits it to the central server through a data network. The system uses the IP protocol and SNMPv2 for management.