udayrockzz / experiment2_restaurant_

Application for Pepper to be deployed in a Restaurant as FOH

Experiment2_Restaurant_

Application for Pepper to be deployed in a Restaurant as FOH

This repo shows code from my disserations HRI and Navigation in hospitality.

Robot interaction was explored by developing applications for Pepper to be deployed in a customer service role through conversational capabilities and built-in tablet functionality such as using branching conversation trees and a comprehensive tablet application for food and beverage orders with tablet-touch voice prompts.

The report can be found here.

The results obtained from interaction with Pepper show a positive review for implementation of conversation and tablet use when being deployed at conferences and as a FOH robot. Pepper was successfully able to communicate with users through basic conversational skills and tablet interaction and fulfil the objectives set for this research.

The biggest challenge with HRI is that Pepper does not have efficient voice recognition or capabilities for accents and complex conversation. Adding new features outside NAOqi on Pepper was outside the scope of this project but where future work is concerned, the integration of more complex chatbot capabilities may produce better results. Examples of this are to use web services and machine learning to continuously improve Pepper’s understanding and response or, implementing a “wake-word” to initialise conversation such as the robot’s name “Pepper” which may be useful for recognition of when a person wants to engage in conversation.

On a different topic, many of the participants interacting with Pepper, particularly in the second experiment, had limited experience with humanoid robots. Therefore, introducing state-of-the-art methods for customer service to a wider range of the public and obtaining feedback will improve the understanding of human reactions and therefore may provide better solutions for introducing humanoid robots into hospitality.