In this paper we present the development of a patrol robot system for home security with some considerations on its interaction functionalities. The system integrates a variety of sensors to gather environmental information and detect abnormal events including fire alarm, intruder alert and lethal gas leakage. To facilitate a security robot to live with people in a home environment, we implement some dedicated human-robot interactions, including a face mask with several facial expressions and a force feedback steering wheel controller. All of the sensor measurements and the dedicated interactions can be remotely accessed via the Internet. We further design an indoor patrol algorithm that can autonomously command the mobile robot to move randomly, avoid obstacles, or steer along a wall baseline based on ultrasonic and vision data fusion. Finally, we apply the patrol robot system to patrol the second floor of EE building of our campus and provide several experimental results for validating its performance.