An extended Kalman filter-simultaneous localization and mapping method with Harris-scale-invariant feature transform feature recognition and laser mapping for humanoid robot navigation in unknown environment

摘要

This article proposes a novel method that combines real-time image recognition with choosing target object real-time localization in unknown environment. The position information of the object recognized as a target is provided by laser, which avoids collecting all the objects position information in the environment. This can save the computation resource and improve the real time. Furthermore, this article realizes the autonomous back and forth navigation in unknown environment and companies voice prompt and intelligent searching of the target object. This will set up a basis on building a complete environment map. Finally, the experiments show the validness of the proposed method and improve the intelligence of the environment information collection.