Man-machine interaction coordination control strategy based on multi-model predictive control

Abstract

In order to solve the problem which cannot be solved by traditional MPC complex automatic driving control under multiple working conditions, the invention provides novel multi-model predictive control(Multi-model predictive content, MMPC) algorithm, and the algorithm is applied to a man-machine cooperative control system of an intelligent automobile. The method comprises steps of adopting Gospson-Kessel algorithm to perform clustering analysis on the real vehicle test data to obtain a clustering center and training sample data of each typical steering working condition; utilizing LS-SVM to build a multi-model structure, and enabling the built sub-models of all types to serve as prediction models of the MPC algorithm; constructing an objective function of multi-objective optimization, andadopting NSGA-II algorithm to solve a multi-objective optimization problem, and obtaining an optimal control quantity, so as to solve a problem that a conventional MPC cannot adapt to the automatic driving control under the conditions of strong parameter time-varying characteristics and complex multiple working conditions.

Description

technical field [0001] The invention belongs to the technical field of unmanned driving, and in particular relates to a human-computer interaction coordination control strategy based on multiple model predictive control. Background technique [0002] After more than half a century of development and evolution, the theory of model predictive control (MPC) has gradually formed a relatively complete theoretical system, and has developed applications in many fields such as industrial control and robotics. At the same time, MPC also provides better solutions for intelligent driving control. However, the human-machine cooperative control process of smart cars is a relatively complex nonlinear control process. With the development of automatic driving, vehicle control tends to become more and more complicated. Autopilot control under conditions. [0003] Since the mathematical model established in the traditional MPC algorithm is fixed, and the intelligent car is affected by many...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention aims to propose a human-computer interaction coordination control strategy based on multiple model predictive control to solve the problem that the traditional MPC cannot adapt to the automatic driving control under complex and multi-working conditions with strong time-varying parameters

Images