Abstract:

As technology advances, robots have gained in importance in manufacturing industries but have also started taking a crucial role in human life. For robots moving into human life, robots must understand humans and the environment. This means they should be equipped with cognitive abilities to understand the way humans act and the relevance of events. Hence, the focus of this thesis will be the ability of robots to recognize objects in an environment. This thesis aims at reconstructing 3D environments, transferring them into a specific data structure for robots to understand the environments and compute the data with path planning algorithms. The study used Truncated Signed Distance Function (TSDF) to integrate the 3D maps from Kinect before saving and visualising them by hash table. Then the data was imported into ICP algorithm to be aligned with existing information to generate a map which can be comprehended by robots. Once the cognitive map was generated, the robot was able to plan the motion based on it. This thesis further proposes that the motion planning can be applied in a bigger space after being improved based on DAO*.