The indoor SLAM using multiple three dimension sensors integration

Abstract

The high performance mobile robots nowadays should operate themselves in many applications. One attractive task is to build the map, meanwhile estimate their poses for the locomotion in real time. The Simultaneous localization and mapping (SLAM) technique is once technique which can solve this demand. The selecting sensors are essential factors for generation the high accuracy output. This paper present, the applying a new 3D sensor namely the Photonic Mixer Devices (PMD) provides a real-time capturing of surrounding volume. However the output of PMD is very excellent depth metric, to recognize the complex objects is still faultily because of the low resolution output (64×48 pixels). The 2D high resolution camera inspires to compensate the PMD camera's drawback. The high resolution image output is registered on the 3D volume. The visual input from the 2D camera does not delivers only high resolution texture data on 3D volume but it is also used for object recognition. Finally, the Iterative Closest Point (ICP) algorithm is used for the image registration in order to yield the real time 3D data frames registration.