Intelligent or Autonomous Vehicles are not visionary technologies that may be present in a far away future. From high tech military unmanned systems and automatic reverse parallel parking to cruise control and Antilock Brake Systems (ABS), these technologies are becoming embedded in people's daily life. The methodologies applied in these applications vary widely across areas, ranging from new hardware development to complex software algorithms. The purpose of this work is to present a survey about the benefits and applications of autonomous ground vehicles and to propose a new learning methodology, which provides the vehicle the capacity of learning maneuvers with a human driver. Four main areas will be discussed: system's topology, instrumentation, high level control, and computational vision. Section 1 will present the most common architecture of autonomous vehicles. The instrumentation used in intelligent vehicles such as differential global positioning system (DGPS), inertial navigation system (INS), radar, ladar and infrared sensor will be described in section 2, as well as the techniques used for simultaneous registration and fusion of multiple sensors. Section 3 presents an overview of some techniques and methods used for visual control in autonomous ground vehicles. Section 4 will describe the most efficient techniques for autonomous driving and parking, collision avoidance and cooperative driving. Finally, section 5 will propose a new algorithm based on Artificial Immune Systems, where a fuzzy system for autonomous maneuvering will be learnt by a data set of actions taken by a human driver. In order to validate the proposed method, the results of its application in an automatic parallel parking maneuver will be showed.