The Research On Key Technology Of Grid Computing For Vehicle Driving Simulators

Vehicle Dynamics which research all things related to automotive system movement mostly research the interaction of the force during car moving and the law of car moving. With the development of info-technology, the Vehicle Dynamics study has been expended an integrated system, that is, man-car-road. In recent years, the step of automotive information has further expedited based on the development of requirement and computer simulation.Vehicle Dynamics study faced on more complex environment. Automobile movement involved lots of movement assembly, which bring complex interaction and coupling relationship. In view of research scope, automobile movement is not only the movement of car itself, but also the interrelationship of man-car-road integrated environment. These actual situations bring some difficulty to Vehicle Dynamics study. The effective approach to resolve these problems is that import the modern computer simulation technology into Vehicle Dynamics study field, which shortened form Vehicle Dynamics Simulation.Computer Simulation is a kind of digital technology testing on the computer. It has three mainly elements including system, model and computer. System is the research object, model is system abstract, modeling or discrimination, and computer is that we achieve the testing and testify goals by computer simulation with setting up the algorithm.Aimed at the Vehicle Dynamics field, the research object is Vehicle Dynamics system, the model means to automobile system dynamics model. Vehicle Dynamics Simulation is based on the automobile theory and control theory, making use of computing simulation technology and automobile system dynamics to experiment, and carrying out analysis of results from experiment with the help of expert knowledge and the statistics data.With the development of automotive industry, vehicle density, travelling high-speeding, cars household and driver non-occupation have become an inevitable trend. Thus, the overall research of vehicle appears to more important, the status of vehicle dynamics simulation also become more prominent. Vehicle Dynamics Simulation not only has the relative independence, but also closely coalesces with several professional technologies such as light, machine, electricity, sound, especially information, and so on. It is interdisciplinary, comprehensive, furthermore, also has some advantage including repeat, security, economy, controllability, without restriction of clime condition, which are other methods can not be compared.Vehicle Driving Simulators plays an increasingly important role in the Vehicle Dynamics Simulation. Due to constraints of domestic automobile manufacturers'resources, the capacity of development and experiment of large Vehicle Driving Simulators is not fully utilized. So how to make the best sharing of the limited resources in Vehicle Dynamics Simulation has become an urgent issue to resolve. Given the success of grid computing applications, Vehicle Dynamics Simulation program become one of feasible solutions, which contains key technology such as resource scheduling strategy and service model. It is of great importance theoretically and practically to research these technologiesGrid computing is an integrated computing and resource environment. It can fully absorb all kinds of computing resource and transform into a sort of everywhere acquirable, reliable, standard and economic computing capacity. In addition to various computers, there are network communication capacity, data, apparatus, human-being and other relative resources. Grid computing, as a newly arisen important foundation facility, takes on distribution, self-similarity, dynamic diversity, multiplicity of management and other important characteristics. Globus is currently one of the most influential grid computing projects in some general purpose grid computing study. On the strength of grid function, it can be divided into computing grid, data grid and information grid etc.Computing grid polymerizes all kinds of which distributed on the network such as isomorphism and heterogeneous computer, workstation, fleet, database, advanced apparatus, storage equipment etc. It forms a comparative transparent, virtual high performance computing environment. Computing grid is defined as a wide area within the scope of integration and the integration of collaborative computing environment. The mostly purpose is to use the existing network resources to implement the effective polymerization of high performance computing environment and to sustain the wide-area distribution of high performance collaborative computing. Computing grid constructs the middleware layer to polymerize the computing resources which are wide distributed via the autonomous computer system and constitute an extensive distribution of super computing based on Internet. It has become a new extensive scientific and engineering computing platform.Based on analysis of necessity of building Vehicle Dynamics Simulation environment with grid computing, we focus on the key technologies for implementing it. Grid computing, the research of which started early and relatively mature in its structure and function, is the basis of network environment. So our researches focus on it. Inherent characteristics have been inherited from the existed parallel systems and distributed systems by grid computing. Since resources are generally heterogeneous and dynamic, new challenging technical problems which do not exist in the parallel systems and distributed systems appear. In allusion to the critical resource scheduling algorithm problems, the thesis draws some new algorithms and important conclusions in depth study from different angles.Our work: 1. Propose a multi-constrained scheduling algorithm for grid. Put up an extended-assignment probability concept. That is setting weight according to the impact of load balancing assignment probability and expected execution time, mapping task to nodes in accordance with extended-assignment probability. Task scheduling model is divided into request information collection module, scheduling management module, task assignment module, user query information module, resource monitoring module and resources collection module by function. According to the each node of entire grid and the whole loading capacity, we get a loading balancing assignment probability concerning about time. On the basis, adding the effect parameters of the expected execution of the mission at each node, we design a function which makes the load balancing of the grid system within the expected range, as the same time, let the makespan of the mission when it executes keep a reasonable minimum.2. Design an aggregation service model based grid architecture referring to the hybrid mode in P2P architecture. The aggregation service based grid is the expansion of grid services. It is divided into four roles, general service providers, polymerization service providers, registration service center, service requestor and six operations which is general service registration, general service binding, general service searching, discovery of service and service searching. Based on the general service, congregate aggregation service and add into grid by the form of grid service, moreover, map the aggregation service model to the refined model, design a service usage model based on the RPC semantics and catalog service semantics.3. Introduce cooperation game in the grid resource scheduling. The unit that the task is allocated to execute is a union of several nodes. In order to keep stable and equitable, we allocate tasks and benefits by the Shapley value. Scheduling algorithm allocates task by union band value, union current capacity proportion and assurance ratio provided, and maps task to union. This module includes seven parts, module to query information which contains task information and resource information, module to maintain a union, module to schedule tasks, module to get instant information, module to allocate tasks to union, module to allocate tasks to members, module to allocate benefits.4. Propose a hierarchical adaptive grid scheduling algorithm. The auctioneers allocate tasks to server-level union according to assignment probability and the Banzhaf Index of Power. Based on the principle of load balancing at the internal of the union, in accordance with communication delays and dynamically adjust the load, and in accordance with Banzhaf value distribute benefits. Introduce the Service Level league reputation degrees, with the implementation of mission, the value is change constant, therefore, the scheduling algorithm is an evolving and learning process.5. Discuss the operation mode of distributed real-time simulation middleware (RSI), with the application of Vehicle Dynamics Simulation as an example. Vehicle Dynamics Simulation refers to multiple independent simulation module worked together, which are dynamic computing server, visual simulation system, sound simulation system, touch simulation system, movement simulation system, and data collection system. These modules are located on the independent grid host computer each other with high speed network linking. Each simulated host installs the support simulation run-time environment (RSI), the provider write simulation model program in accordance to RSI interface criterion.