| With the advancement of technology, Home Service Robots (HSRs) willhave a broader market. This paper studies how HSRs recognize objects in the indoorenvironment and grasp objects. After receiving orders, HSRs are capable of movingautonomously to the location near the target object through navigation by a starpositioning system, recognizing and locating the object by the long-range coloridentification plus close-feature recognition, and grasping the object through themechanical arms. Specifically, the research covers the following aspects:1.Built the platform that integrates motion control, mechanical arms control,information collection, object recognition and positioning on the VS2010andOpenCV software platform.2.Studied the object identification algorithm SURF (Speeded Up RobustFeatures algorithm). Due to the difficulty to capture characteristics at a longerdistance, the object recognition strategy adopted in this study is one integrating coloralgorithm, which is to find objects at a distance using color recognition algorithmand to confirm the object’s position using SURF at a nearby place.3. Created maps of the indoor environment based on the StarGazer starpositioning and navigation system. Therefore, we solve the navigation issue duringthe process of finding the object.4. Studied the grasping techniques of mechanical arms. Proposed solutions togain end executors’ expected positions from features of different angles based on thepinhole imaging model. Calculated the torque angles of the joints of mechanical armsbased on the mechanical arm motion formulas.With this application platform, the HSR can successfully complete the tasks ofobject recognition and grasping. The experiment indicates that efficiency andaccuracy of recognition can be improved through the combination of color andcharacteristic recognitions; that star navigation can enable robots to move to targeted position by itself; that by turning at calculated angle, the mechanical arms can graspobjects successfully. |
PDF Full Text Request