A multisensory data fusion method of the two-wheeled self-balanced robot

To recognize the running state of the robot efficiently, a SVM-based multisensory two-graded data fusion method is presented The running state recognition is realized from the view of classification. The problem that the classified accuracy is low is solved The method is applied to the two-wheeled self-balanced robot and the experiments of the running state recognition are conducted When the individual sample number of each running state exceeds twenty the accuracy of fusion method will be above 98%. Experimental results demonstrate that the running state of the two-wheeled self-balanced robot could be recognized efficiently and reliably. The real-time requirement will be suitable in the fast and maneuverable process.