Introduction

This chapter covers how robots estimate where they are and what the environment looks like. You will start by seeing why wheel-encoder odometry alone drifts over time, then learn how external landmarks and feature-based maps correct that drift. From there you will confront the SLAM paradox, the chicken-and-egg problem of needing a map to localize and needing a position estimate to map, and see how robots solve it by treating position and landmark locations as probability distributions that sharpen with every sensor observation. The chapter closes with two hands-on simulations that let you experiment directly with how drift accumulates and how periodic landmark corrections keep it in check. Together, mapping and localization provide the state estimates needed for reliable autonomy.