The Precise Positioning Of The Domestic Robots Based On Multimodal Fusion

In the context of pervasive computing, with the vigorous development of robot technology and improve people’s quality of life, the global market size of home service robot is rapidly expanding. In the field of household robots, positioning technology is one of the most basic, the most important research, researchers at the mobile robot localization has done a lot of research work.In the practical application of domestic robots, only make sure robots can accurately positioning, to other behaviors, such as the fetch, obstacle avoidance, camera and so on.In order to make the robot’s movement more accurately, this paper made the following several aspects.First of all, the paper analyzes the domestic robots commonly used indoor positioning technology research status, and analyzes the data fusion of multimodal fusion technology research present situation and the technical difficulties of the current integration technology.Second, in indoor WiFi positioning technology, first of all localization algorithms are classified, and then analyzes the WiFi is commonly used in RF radio frequency localization and RFID positioning.In WiFi positioning of the robot, this paper improves a fuzzy clustering position fusion algorithm of fingerprint orientation method, relative to the hard clustering algorithm, can effectively improve the positioning precision of the indoor positioning.In the RFID localization of robot, this paper puts forward to tag, improve the efficiency of positioning, real-time and the positioning error was tested. Third, in the study of kinematic positioning, analyzes the household robot kinematics, coordinate system of the model theory is firstly analyzed, the kinematics coordinate system can be divided into positive movement model and inverse model, then the kinematics of the robot positioning that is commonly used in ultrasonic positioning and odometer positioning principle is expounded, finally completed in the experiment of ultrasonic distance measurement sensor calibration, the calibration and the receiving of the ultrasonic range was tested.In positioning results show that the indoor positioning accuracy can meet the demand of experiment. Fourth, in the multi-channel data fusion of positioning, analyzes the DS evidence theory, for each channel to obtain the data between the uncertainty and the compatibility and mutual exclusivity, improve a kind of secondary adjusting weighted based DS evidence theory to optimize the results of positioning, and set up a sign in the program, according to the data source to determine logo, in each of the channels are involved in the fusion, will this method comparing with classical DS evidence theory fusion results, the results showed that the DS evidence theory based on secondary adjusting weighting can improve the positioning precision of the home robots, meet the requirements of indoor location.