Neural network based for automatic vehicle classification in forward scattering radar

The paper is dedicated to the continuation and improvement of the vehicle classification method of SISAR micro-sensors for ground vehicles previously presented in RADAR2004 and RADAR2005 [1–2]. In spite of a number of theoretical research efforts in the application of SISAR for target classification [1–4] , there are only few research concentrate on the classification processing to confirm the feasibility of SISAR´s practicality. This paper begins with an overview and summary of the authors´ previous research. Then a new research topic in the improvement of the classification performance for various scenarios using Neural Network is proposed. Finally experimental results, conclusions and recommendations are presented.