Research On Existing Horizontal Alignment Parameter Estimation And Development Of CAD System

Abstract:At the end of2013, highway mileage have reached4.23million kilometers, and railway running mileage have reached0.1million kilometers in China. Therefore, there will be a heavy and hard task for the renovation of old road and extension construction of existing railway. Researching on Existing Line of Railway and Highway CAD Techology are of great engineering significances. Taken the size of adjusting distance as evaluation criterion for the quality of restoring project, which is based on the coordinate principle, and starts from measuring points to line elements, to the whole associated linear optimization, so this research designed the linear iterative optimization model under the improved simulated annealing algorithm. The Optimization Algorithm Module in this paper include linear elements recognition and linear iterative optimization, and accordingly put forward "the algorithm of trend discrimination approach on the orthogonal fitting curvature of five-point","the concept of boundary between confusion region", and "the linear iterative optimization model based on the improved simulated annealing algorithm". The optimized alignment which fits the actual measurements of centerline best and satisfies the design constraints and technical requirement is built intelligently. Alignment has achieved global objective function optimization. The main contents and research results are as follows:(1) By the method of coordinate principle, a new algorithm of the plane linear estimate parameters optimization which is based on the absolute coordinate of center line has been proposed. To achieve the research objective, this paper consists of two parts:preliminary identification of line element module and linear iterative optimization module.(2) Based on the method of line elements recognition, using the principle of orthogonal least square helps improving the "Trend Discrimination Approach on the Curvature of Three-point Algorithm", and thus proposing the "Trend Discrimination Approach on the Orthogonal Fitting Curvature of Five-point Algorithm" to improve effectively measuring point section identification’s accuracy and stability. The concept of "Fuzzy Area of Line Elements Boundary" are put forwarded to reduce the assignment errors of measuring point because of line elements piecewise, and then measuring points in fuzzy area are auto-reassigned through the linear elements iterative optimization model to meet the objective requirements.(3) A linear iterative optimization model which is based on the improved simulated annealing algorithm has been proposed. Establishing the linear iterative optimization model with the minimum of sum of squares as the objective function, as well as line elements and measuring points as independent variables. By introducing the intelligent algorithm, the improved simulated annealing algorithm are combined with linear iterative optimization model. The search strategy are based on the preliminary identification of alignment.At last, the optimized alignment which fits the actual measurements of centerline best and satisfies the design constraints and technical requirement is built intelligently, and the realignment achieved global objective function.(4) The plane linear optimization auxiliary design system are developed by using the platform of CAD.In VC++6.0environment, by using of CAD secondary development technology of ObjectARX and database accession ADO, the plane linear optimization auxiliary design module are developed. The module generates optimal alignment and parameters automatically, then verifies the correctness and feasibility of it by projects.