On Intelligent System For Technical Examination Of Vehicles In Accident

Scientific traffic management entails the identification & compensation for road traffic accident, in which the technical examination of vehicles in accident plays a pivotal role. Through the technical examination, we can get a general picture of how the vehicle acts in the traffic accident, thus objectively analyzing the reasons behind the accident so as to identify the party's liability. Nowadays vehicles are numerous, and the number of road traffic accidents is on the rise, which calls for more experts as well as more scientific examinations indeed. However, the status quo is far from satisfactory. Actually, the technical examinations of vehicles are mainly carried out by experts based on their experience over years, which is subjective for sure. So it is of great significance to develop an intelligent system for technical examination of vehicles in accident.Therefore, an intelligent system for technical examination of vehicles in accident is introduced and studied in this paper. Here, all the historical cases of vehicle examination proceeded by experts are kept on file, from which the system extracts knowledge for reference, and if possible, applies it to the new cases.To begin with, the structures of traditional DSS and IDSS are discussed, and then, after the analysis of case-based reasoning, the structure of an intelligent system for technical examination of vehicles is advanced. Specifically speaking, it comprises case base, case retrieval knowledge base, examination reference database, modifying rule base and user interface etc., capable of retrieving cases, selecting plans, modifying plans, implementing plans and evaluating results. On the whole, the system has the following characteristics: First, it uses cases to illustrate the problem as well as the ways of solving the problem, and these cases provide more accurate information than rules do. Second, since it tries to find the solution to the new problem based on the historical cases, it avoids the match conflict and combination explosion and can be more efficient to solve the problem; Third, its scope gradually gets wider along with the increasing cases; Last, it is more practical for it is more easy to get cases than rules, and full domain model is not needed as well.In the light of CBDT, by analyzing decision makers' habits, the paper brings forth the decision-making method based on the similarity with alterable weights. Theoretically it holds the attributes' weights corresponding to different acts are different, which tallies with people's behavior in real life. As a result, it proves efficient in the technical examination of vehicles in accident.As to the case revising in CBR, the case revising rule, gained by applying intelligent retrieving method, is to help decision maker revise the examination scheme. Generally it consists of secondary revising rules and primary revising rules. Firstly the rough set theory is applied to get the secondary revising rules. Secondly, with the case revising transaction set obtained from experts as reference, the association rules mining method is employed to get the primary revising rules. Lastly, by combining the two types of rules, the author expounds the detailed case revising procedure.Besides, the structure of the examination reference database is discussed, and a multi-attribute evaluation method is put forward in this paper. Therein the attributes' weights and values are all expressed in the form of linguistic values, based on which the final result can be aggregated by the LIA. To get the objective linguistic weights, a linguistic weights optimization model, with the minimum discrepancy as its target, is introduced. The Genetic Algorithm is applied to solve this model so as to get the final optimal linguistics weights.In order to get the information about the technical examination of vehicles in accident from the document, and also, to provide the foundation for the autogeneration of examination report in natural language, the information retrieval techniques are studied, and an information retrieval technique based on the dependency relationship of sentences is produced. It constructs the template according to the dependency relationship between retrieving words and frequently-used words. The template, as a set of dependency characteristics in the form of logic expression, is simple and direct. Based on these templates, we can select the word tallying mostly with the dependency characteristics of the template as the final retrieving word, so the whole process is straight and speedy.