This project deals with the design and construction of an autonomous car able to video map an unknown room. The main issues are analysis of the optimal way to dimension and map a room, design and construction of the car, and designing a controller for the car. The analysis is based on finding the best suited robot for mapping the room, and how to optimize the use of a spy cam in accordance to making a complete video map of the room. The car is designed to weight less that $500~g$ and to be able to estimate the dimensions of the room. In the design of the car a tilting device for the camera has been made to get full use of the cameras vertical view angle. Two controllers for the car have been designed. One model-based and one nonmodel-based. Both use the classic control theory, but two different design methods have been used. Root-Locus for the model-based and Ziegler-Nichols for the nonmodel-based controller. Their performances have been compared and the nonmodel-based controller had the best performance of the two. The nonmodel-based controller can only be used for simulations though, while the model based controller can be used for implementation.