A Control Method To Avoid Obstacles For An Intelligent Car Based On Rough Sets And Neighborhood Systems

Intelligent vehicle has become a research hotspot.Many automobile manufacturers and IT companies have invested heavily and some progress has been made.But its theory and technology are confidential.It is of great significance to study the theory and technology of intelligent vehicle.Due to the complexity of the road,there is a certain degree of difficulty to achieve complete intelligence.Therefore,the study of part of intelligent function of the car is a very available plan in the short term.It is an important part for the automatic driving technology that the obstacle avoidance and path selection in the road with obstacles.In this paper,we study it based on the related theory of neighborhood system and rough set.Firstly,in order to distinguish single parameter fuzzy set and multi-parameter fuzzy set,the concept of multi-dimensional fuzzy set is proposed.It is proved theoretically that the projection of multi-dimensional fuzzy set in some parameter's domain is a single parameter fuzzy set.At the same time,the relation between fuzzy soft set and multi-dimensional fuzzy set is discussed.Furthermore,the embodiment and function of the relation in frizzy comprehensive evaluation is studied in theory,with which to point out the diversity of fuzzy comprehensive evaluation.For studying the dynamic control model of intelligent vehicle in the road with obstacles,firstly,according to the idea of differentiation,a mathematical model of differential neighborhood in three-dimensional space is proposed.And apply it to create planar feasible neighborhood of intelligent vehicle in the road with dynamic obstacles.In this method,according to the changing trend of the key points in the neighborhood boundary,and considering the width and length of the neighborhood,the feasible neighborhood of intelligent vehicle is obtained.For realizing the algorithm easily,this neighborhood is covered by means of a finite number of trapezoid neighborhoods,whose lower approximation set in the sense of the covering rough set is used as a trapezoid feasible neighborhood.For static obstacles,the obstacle is covered by means of a finite number of trapezoidal neighborhoods.And take its the upper approximation set in the sense of the covering rough set as the obstacle region.Thus,irregular obstacles which are not easy to be described mathematically can be done easily.Compared with the dynamic obstacles,the method of creating the feasible neighborhood of the intelligent vehicle under the condition of the static obstacles is relatively simple.In this method,based on the neighborhood related parameters,a trapezoid feasible neighborhood system is obtained by using fuzzy soft set and multi-dimensional fuzzy set.Then the selection of trapezoid feasible neighborhood is given relative to regular obstacles based on the trapezoidal feasible neighborhood system.No matter the obstacles are dynamic or static,once the trapezoidal feasible neighborhood of the intelligent vehicle is determined,then the neighborhood is covered by means of a finite number of standard trapezoidal feasible neighborhoods.And its the lower approximation set in the sense of the covering rough set is considered to approximate the neighborhood.Furthermore,the corresponding control rules of the standard trapezoidal feasible neighborhood are obtained to make decision.Finally,the simulation experiment is carried out for obstacle avoidance control algorithm under the obstacle road by Matlab.The effectiveness of the above method is illustrated.