Research On Object Management And Task Scheduling In Vehicle Controllers Of Electric Vehicles

At present,the application of electric vehicles has become more and more complex and functions become increasingly intelligent,and the mode of developing its vehicle controller still adopts the traditional “V” mode.The streaking software generated through this mode cannot perform effective device management and task scheduling when facing a large number of controlled devices.This has caused the problem that the code of the electric vehicle controller is not universal,the development cycle of the new function is too long,and the intelligent progress is slow.In order to solve the above problem,this paper does the following work:In view of the lack of a digital(object-oriented)model of vehicle control equipment in the abstract layer of the vehicle controller for electric vehicles,a “data flow method” was proposed to classify on-board equipment;The “change rate determination method” calculates the hierarchical relationship between device objects and realizes the object-oriented processing of vehicle equipment.Aiming at the problem of a large number of controlled objects in electric vehicles and the current in-vehicle software's incapability of managing a large number of objects efficiently,a "jump hash net" object management method is proposed.Consolidate all the objects into a one-dimensional linear table,consecutively number them,and divide them into child nodes.Then add the child node number values to all the parent nodes in the one-dimensional linear table to form a skip exit,and then generate a hash through the parent node.The table,at the end of the jump address,forms the structure of the entire table.With this structure,adjusting the number of parent nodes and child nodes in the grid can achieve the goal of efficient management of all nodes without occupying too many resources.Aiming at the problem that the number of control software in electric vehicles is very large and the current vehicle software has no real-time task scheduling function,a task scheduling method based on time stamp compensation and soft interrupt is proposed.By measuring the timestamp of the last execution of the task and the execution timestamp of the current task,the periodic delay deviation is calculated,and the average delay of the entire operating system kernel environment is counted,and the obtained delay time is used as the compensation value to perform the periodic property of the task.Compensation,and according to the average delay value and priority value to determine whether the task needs to be deprived of scheduling through soft interrupts.Through this mechanism,real-time scheduling with a cycle precision of 8 microseconds for a large number of tasks can be realized.Finally,in a certain type of pure electric passenger vehicle vehicle controller,the above algorithm is transplanted in the operating system of the vehicle,and the feasibility of this research is verified through the algorithm test experiment.