Human-robot interaction in indoor

During the preparation of this thesis, the studies in the field of HRI were examined and the methods used were investigated. Inspired by these studies, dialogue models that can be used in robotic systems have been studied in the field of HRI. In this thesis, finite state machine and content independent grammar methods are used for different dialogue models. In this context, finite state machines and content free grammar methods were examined. Dialogue modeling was developed with these methods. These dialogue models were tested in real time on the robot. There are many stages in applications. For each application, firstly image processing techniques were used, then the following systems were used, such as speech detection systems. Two experimental areas related to HRI were created. In the first experiment, when the robot finds the human, he starts to follow the head by mimicking the facial movements of the human being. In the initial state, the robot expects interaction from the other person. If the person in front of the robot does not interact with the robot within the specified time period, it starts the interaction by telling the person in front of the robot how to use itself. If the human interacts, the robot answers the questions asked by human.The interaction is completed according to the Finite State Machine (FSM) specified by the programmer who programming the robot. Three different FSM have been proposed. These models are called "Human control model", "Head tracking model" and "Robot eye control model". The two experimental media were crossed with the proposed three models and six experiments were performed with a total of five replicates. In the experiments, successful interactions between FSM models were determined. The most suitable model was determined. In the future studies, it is aimed to increase this ratio even more by analyzing the age, gender and emotion of the people.