A review of social-aware navigation frameworks for service robot in dynamic human environments

Abstract

The emergence of service robot into human daily life in the past years has opened up various challenges including human-robot interaction, joint-goal achievement and machine learning. Social-aware navigation also gains vast research attention in enhancing the social capabilities of service robots. Human motions are stochastic and social conventions are very complex. Sophisticated approaches are needed for a robot to abide to these social rules and perform obstacle avoidance. To maintain the level of social comfort and achieve a given task, the robot navigation is now no longer a search for a shortest collision-free path, but a multi-objective problem that requires a unified social-aware navigation framework. A careful selection of navigation components including global planner, local planner, the prediction model and a suitable robot platform is also required to offer an effective navigation amidst the dynamic human environment. Hence, this review paper aims to offer insights for service robot implementation by highlighting four varieties of navigation frameworks, various navigation components and different robot platforms.