Online Optimal Energy Distribution of Composite Power Vehicles Based on BP Neural Network Velocity Prediction

In this paper, an online optimal energy distribution method is proposed for composite power vehicles using BP neural network velocity prediction. Firstly, the predicted vehicle speed in the future period is obtained via the output of a BP neural network, where the current vehicle driving state and elapsed vehicle speed information is used as the input. Then, according to the predicted vehicle speed, an energy management method based on model predictive control is proposed, and online real-time power distribution is carried out through rolling optimization and feedback correction. Cosimulation results under urban drive cycle show that the proposed method can effectively improve the energy efficiency of composite power sources compared with the commonly used method with the assumption of prior known driving conditions.

A Predictive Control Model for Master Slave Robotic Manipulator with RBF Neural Network

In recent years, manipulator control has been widely concerned, and its uncertainty is one of the focuses. As we all know, the manipulator is a MIMO nonlinear system, which has the characteristics of severe variable coupling, large time-varying amplitude of parameters and high degree of nonlinearity. Therefore, a lot of uncertain factors must be considered when designing the control algorithm of manipulator system. The predictive control algorithm adopts online rolling optimization, and in the process of optimization, feedback correction is carried out by the difference between the actual output and the reference output. It can iterate the predictive model and suppress the influence of some uncertain disturbances to a certain extent. Therefore, the design of predictive controller for robot is not only of theoretical significance, but also of great practical significance. The trajectory tracking problem is proposed in this paper, and a predictive control method for master slave robotic manipulator with sliding mode controller is designed. In addition, when external disturbances occurred, the approximation errors are compensated by the proposed control method. Finally, The results demonstrate that the stability of the controllers can be improved for the trajectory tracking errors.