This chapter looks at how Self-driving vehicles (SDVs) perceive the world around them. Perception is the single most important function of an SDV. The aim of the perception function is to achieve as complete and accurate understanding of the vehicle’s environment as possible, in order to provide a basis for decision making in the subsequent navigation function. Global Navigation Satellite Systems are a popular global localization technique as they offer a simple and inexpensive way for vehicles to localize. Localization is the process of determining the vehicle’s position and orientation based on a map. In general, perception in dynamic environments, i.e., environments with moving objects, can be decomposed into two major sub-functions: Simultaneous Localization and Mapping (SLAM) and Detection and Tracking of Moving Objects. In graph-based SLAM, a pose-constraint graph is constructed and the task is to find the configuration that is maximally consistent with that graph. Some open-source implementations of graph-based SLAM are available on the Internet.