This thesis focuses on two important aspects of autonomous vehicles:  Control of the vehicle’s longitudinal and lateral motion, and recognition of the objects in the vehicle’s drivable space.  In terms of vehicle control, we first drilled down to the existing kinematic and dynamical models used to describe the motion of the vehicle.  Subsequently, we developed appropriate lateral and longitudinal controllers to be used in the CARLA vehicle simulator. For longitudinal control we developed a PID controller, while for lateral control we developed a Stanley controller, and we implemented both in the Python language. The controllers generated brake, throttle and accelerator commands to drive the vehicle dynamical model in the CARLA environment. For longitudinal control, we tuned the gains of the PID controller to achieve satisfactory performance.  Τhe results indicated that a PD controller  appropriately tuned resulted in good performance. In terms of visual perception, we drilled down on aspects of existing related methods and techniques and outlined how they can be used to achieve this very complex task.  Subsequently, we developed Python routines to process the semantic segmentation output of a deep neural network and perform relatively simple tasks, such as ground plane estimation, lane marking identification, object recognition within predefined bounding boxes and distance estimation between the recognized objects and the vehicle. Possibly the most significant contribution of this thesis is the systematic presentation of existing fundamental knowledge in the above two areas in a way that one can build upon to develop new, improved concepts for vehicle control and visual perception.