This project is the latest generation of robotic person following robots that have been implemented at the Intelligent Machines and Robots Laboratory. The goal of the thesis is for a robot to follow a person using a cellphone carried by the person. In this work we use the accelerometer and gyroscope data of the cellphone device to estimate the distance of the person to the robot. The accelerometer signal transmitted from the cellphone to the robot provides an indication of the movements of the person. We have made a number of preliminary experiments that show the strength (amplitude) of the acceleration signal and periods between signal peaks are related to the speed of the person's walk and the traveled distance. However, the accelerometer signal is extremely noisy. As a result we use Kalman filtering to reduce the noise. Several signal processing techniques such as peak detection and Fourier transform are then applied to extract useful information for estimating the distance traveled by the person. This information will be used in a fuzzy expert system to estimate the distance. The cellphone gyroscope is also used to complement and enhance the robot control to follow the person closely. We believe that methods developed in this thesis have important applications other than person following, such as precise localization of a person walking in a building for which the GPS has very poor accuracy and performance.