Emerging Situation Awareness of Drivable Routes for Autonomous Robots Using Temporal probabilistic Reasoning

Researchers and practitioners have stressed that autonomous navigation in complex environments is an ongoing key challenge for robotic vehicles. Detection of drivable routes is often used as one of the important safety key operations to address some of the issues associated with autonomous navigation. While a number of conventional detection methods have been developed for such navigation; awareness of drivable routes by alleviating robot short-sightedness -without being trapped in uncertain dead-end problems, and to facilitating global navigational planning have received little attention. Finding a solution to these uncertainty problems is a challenge. In this paper, temporal probabilistic reasoning of the Emergent Situation Awareness (ESA) technology is proposed as a supportive strategy for autonomous navigation. The ability to reveal uncertainties over time is a drivable route awareness strategy of hidden paths embedded in the complex environments. Experimental evaluations of the ESA on real life and publicly available road frames outperform the classical statistical baseline methods in handling uncertainties over time. Our awareness results reveal to robotic vehicles that all ground planes are not traversable routes