We present in this paper a new microwave sensor developed for autonomous vehicle navigation and perception in extensive outdoor environments. Radar offers advantages as a robotic perception modality, because it is not vulnerable to the functioning conditions found in natural outdoor environments: dust, rain, fog, snow, light conditions, wind turbulences, etc. The developed radar is based on Frequency Modulated Continuous Wave (FMCW) technology, that is well intended for short and medium-range applications. The paper describes the spatial integration that is used in order to reduce the speckle effect in radar image. This anti-speckle filtering is obtained through a dead reckoning phase, using two proprioceptive sensors (odometer and compass heading). The radar sensor is used to develop a Simultaneous Localization And Mapping (SLAM) application. In order to fit the needs of radar image analysis complexity, the SLAM process is purely radar based. This process makes no landmark assumptions, and avoids the data association problem by storing a detailed map instead of sparse landmarks.