To finish our project, we prepared a presentation, recorded a video demo and wrote a report to summarize and reflect on our work. At the end of the project, we are proud to have achieved our goals and to have created a fully functional robotic bartender. We are also happy to have learned a lot about the different aspects of the project, from the conception of the idea to the implementation of the final product. We can conclude that we have successfully completed our project and are satisfied with the result. Botender - the robotic bartender - is ready to serve you a drink!

Our program is designed to identify and address people by their names, thereby raising the user experience to a new level of engagement. This was achieved by enabling our AI system to offer personalized greetings.

Our robotic bartender has now mastered expressing emotions. Botender can convey feelings ranging from happiness to concern.

Botender’s behavior is adapted in real-time to respond appropriately to the user’s emotions and the context of the interaction. The Interaction Component integrates perception data with a state machine approach.

The main part of the user perception component is the prediction of the user’s emotion. This is done by a machine learning model that uses the facial expression of the user as input. The model is trained with a dataset of images of faces and the corresponding emotion.

To make the interaction with the robot more natural, we had the plan to implement a gaze controller. The gaze controller should be able to detect the location of the customer’s face and then move the robot’s head to look at the customer.

After laying the foundational architecture of the Botender framework, our next step is to implement the User Perception Framework.

Framework Components and Their Roles

• The EmotionDetector predicts the customer’s emotional state by utilizing a Support Vector Machine (SVC) model. The model is trained based on the dataset DiffusionFER provided on Studium. It is loaded from a pre-trained file.

After defining the control flow and the system architecture, we started to implement the Botender framework. The goal of the framework is to encapsulate the different subsystems to be able to develop them independently. The framework should also provide a common interface for the subsystems to communicate with each other.

The goal of our second meeting was to define the control flow of our dialogue with Botender and the high-level system architecture. This was required to get a common understanding of our application and define the tasks accordingly.

The goal of our first meeting was to create the specification document, which outlines the objectives of our project, alongside the resulting deliverables. We also identified possible future issues and stepping stones. Lastly we began to structure all tasks into a timeline in order to gain an overview of the project.