Gesture Driven Remote Robot Manoeuvre for Unstructured Environment

摘要

This article proposes a novel framework for robot navigation in unstructured environments based upon gestural features. In such an environment, it is not possible for a human being to be present in person to accomplish any task. Hence it is a good alternative to direct a robot to navigate to a specific position and execute a specific task. Without efficient instructions there is a high probability of the task remaining incomplete due to the robot not reaching its goal or a possible damage of the robot because of the unknown environment. Gestures are considered to be a wise modality for robot control in unstructured environment because of the high accuracy that the task calls for. For this purpose, an Interval Type 2 Fuzzy based gesture recognition scheme has been employed for the classification phase. This is done to register the uncertainty quotient in the extracted gestural features from different subjects at different time instances. This article addresses an extremely important issue of robot navigation in perilous environments which are not accessible to the human beings. The scheme has shown promising potential for future applications in terms of performance metrics as compared to other existing state of the art techniques.