SIMULTANEOUS LOCALIZATION AND MAPPING SYSTEM IN UNMANNED GROUNDED VEHICLE

摘要

Robot navigation in the environments inaccessible for humans demands techniques for simultaneous localization and mapping (SLAM). Navigation is the process of moving the robot to a goal location in the environment steering around obstacles along the way. The biggest issues with attempts to implement such algorithms in the real world are establishing the right transforms and obtaining accurate data from the sensors. Recent advances in sensor technology have allowed such research to achieve ground breaking results that hold viability as a possible commercial use. The proposed invention is also based on SLAM principle for robots working as unmanned guided vehicles (UGV). The proposed system is a very economic and reliable tool for environment mapping in the hostile environment. The upper hand of the designed system is that it does not need to have prior topological information of the environment. This invention is developed on the existing framework which sequentially revolves around feature extraction from RGB, projection of depth data onto the features, association of features with previous and current map, estimation of current position of robot (localization), instruction to the robot for movement and further updation in the environment map. The designed system has used low cost Kinect sensor and a Hokuyo URG-04LX range-finder for acquiring the environment information. This invention has built a robotic platform that has the capability of autonomous navigation based on SLAM concept.