Enhancing Human-Robot Cohesion through HAT Methods: A Crowd-Avoidance Model for Safety Aware Navigation

In the world of autonomous robots, achieving complete accuracy remains a challenge, underscoring the importance of human intervention, especially regarding safety. In Human-Autonomy Teaming (HAT), ensuring safe and effective human-robot cooperation in dynamic indoor settings is crucial. This paper introduces a framework designed to address this accuracy shortfall, enhancing safety and robotic interactions in these environments. Central to our approach is a hybrid graph system that melds the Generalized Voronoi Diagram (GVD) with spatio-temporal graphs, harmonizing human feedback, environmental elements, and key waypoints. An essential component is the improved Node Selection Algorithm (iNSA), which utilizes the revised Grey Wolf Optimization (rGWO) for enhanced adaptability and performance. Human insights are instrumental, from supplying initial environmental data and determining key waypoints to stepping in during unexpected challenges or dynamic changes in the environment. Extensive simulation and comparison tests confirm the reliability and effectiveness of our model, highlighting its unique advantages in the realm of HAT.