This thesis discusses the challenge faced in the Netherlands, caused by an increasing shortage of human resources to provide care for the growing population of individuals aged 65 years and older. To tackle this challenge, HIT is developing Project ROSE, a semi-autonomous care robot capable of performing basic and generic tasks in care facilities. However, the default navigation solution to move around in the care facility while continuously interacting with humans leads to unsafe situations. As a solution, this study aims to create a socially aware navigation system that considers social norms to make robots more socially acceptable in public spaces.
Primary social norms are identified by analyzing human and robot behaviour in an indoor environment, and a new navigation design has been developed. Primary social norms are staying on the right side while navigating in a corridor and respecting personal space. Additionally, norms were derived from analysing socially unacceptable robot navigation behaviour, which included not scaring people around intersections and avoiding forbidden zones. The new design's performance has been evaluated using objective and subjective metrics. Objective metrics include the robot's path length to destination, number of recoveries, and time to destination. Subjective metrics include a survey among test subjects exposed to a range of simulated scenarios. The objective metrics indicate that socially aware navigation is more effective than default navigation behaviour, significantly reducing the number of recoveries and unpredictable behaviours.
Socially aware navigation does take a bit longer to reach the destination. The survey results indicate that the new navigation design is deemed more socially acceptable than the default navigation behaviour. This study contributes to social robotics by providing a simple approach to a new navigation design that considers social norms, making robots more acceptable and adaptable in public spaces.