| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 107 === To implement the human and robot collaborations, the speech enhancement based (SEB) adaptive hierarchical fixed-time saturated control (AHFTSC) of omnidirectional mobile robot (ODMR) is designed. From the outset, the features of nine speech commands are extracted and then trained by multiclass support vector machine. Moreover, the background noise is addressed and filtered by the suitable design of filtering mechanism. Comparisons among with or without background noise and filtering are given. Due to advantageous feature of ODMR (i.e., synchronous translation and rotation), the speech command based controls are more reliable and satisfactory. For the faster pose tracking ability, ODMR are divided as indirect and direct modes to design the AHFTSC. To null track the indirect output (desired pose of the ODMR) in fixed time, the adaptive fixed-time virtual desired pose (AFTVDP) with nonlinear switching gains is designed. To make the direct output (motor current) null track the AFTVDP in fixed time, an adaptive fixed-time saturated control (AFTSC) with nonlinear switching gains is employed to execute high frequency motions of speech command. In summary, the proposed SEB-AHFTSC contains speech enhancement’s recognition and classification, AFTVDP, and AFTSC.
|
|---|
