Aimed at uncertainties and model's impreciseness, nonlinearity and time-variability of depth control system in autonomous underwater vehicle (AUV), a depth predictive control method was put forward based on rough ...Aimed at uncertainties and model's impreciseness, nonlinearity and time-variability of depth control system in autonomous underwater vehicle (AUV), a depth predictive control method was put forward based on rough set (RS) and least squares support vector machine (LSSVM). By using RS theory, the monitor data attribute of AUV was reduced to eliminate the redundant information and to improve efficiency. Then, LSSVM model was trained by using the reduced rules, and its parameters were optimized by using chaos theory for the higher accurate control. Taken an AUV typed NPS Phoenix as an example, its depth step response, horizontal rudder and pitch change were simulated. The simulation results show that the method improves the model's accuracy and has better real-time response, fault-tolerant ability, reliability and strong anti-interfere capability.展开更多
文摘Aimed at uncertainties and model's impreciseness, nonlinearity and time-variability of depth control system in autonomous underwater vehicle (AUV), a depth predictive control method was put forward based on rough set (RS) and least squares support vector machine (LSSVM). By using RS theory, the monitor data attribute of AUV was reduced to eliminate the redundant information and to improve efficiency. Then, LSSVM model was trained by using the reduced rules, and its parameters were optimized by using chaos theory for the higher accurate control. Taken an AUV typed NPS Phoenix as an example, its depth step response, horizontal rudder and pitch change were simulated. The simulation results show that the method improves the model's accuracy and has better real-time response, fault-tolerant ability, reliability and strong anti-interfere capability.