Abstract:Dynamic and complex indoor environments present a challenge for mobile robot navigation. The robot must be able to simultaneously map the environment, which often has repetitive features, whilst keep track of its pose and location. This chapter introduces some of the key considerations for human guided navigation. Rather than letting the robot explore the environment fully autonomously, we consider the use of human guidance for progressively building up the environment map and establishing scene association, learning, as well as navigation and planning. After the guide has taken the robot through the environment and indicated the points of interest via hand gestures, the robot is then able to use the geometric map and scene descriptors captured during the tour to create a high-level plan for subsequent autonomous navigation within the environment. Issues related to gesture recognition, multi-cue integration, tracking, target pursuing, scene association and navigation planning are discussed.

Autonomous Navigation For Mobile Robots With Human-Robot Interaction