Improved Particle Swarm Optimization Applied In Calibrating And Registering The Triaxial Measuring System

Particle Swarm Optimization (PSO) is an intelligent optimization algorithm based on cooperation and competition among swarms. This algorithm is simple. easy to realize and has compact code, but it is easy to fall into a locally optimized point and the search precision is low. Under the precondition of keeping the basic advantage of PSO, an improved PSO with sell-learn ability is presented to improve the detection and development performance from the angle of social psychology and biological behavior. When a particle needs to update its speed. it can determine the next step of evolution according to self and group condition. Also, the self-learning mechanism can increase diversity of particle swarm dramatically, reduce human intervention during evolutionary computation and improve the self-adaptability of system. Moreover, this new algorithm can significantly improve the quality of solutions and convergence speed. which can be verified by the typical test functions.The calibration and registration of triaxial measuring system has been widely used in laser communication, navigation and position, attitude control and so on, which is one of the fundamental issues in 3D measurement. In the experimental measurement of 3-axis angles, the non-orthogonal error, gain error and zero drift are generally existed in three-axis measuring system, and the registration error existing between two triaxial orthogonal coordinate system due to the corresponding axis are not parallel, all of which have enormous influence on the high precision measurement. It is impractical to deal with these problems by improving fabrication techniques, and it is difficult to calibrate the measurement system due to nonlinearity. Also, it is not easy to calibrate the sensor in three dimensions by precise instrument, which could increase the work and cost of scientific research. Therefore, algorithm compensation is a feasible and effective method to rectify the errors in three-axis measuring system.In this paper, an off-line method is studied for calibrating non-orthogonal error, gain error and zero drift existing among three measuring axes by using improved PSO based on self-learning mechanism. And an error parameters model is built by the mapping relationship between the measure system and a simplified perfect orthogonal system. In this paper, the corrected measuring system is considered as an ideal orthogonal system, and errors of corresponding axis shaft will not be changed once double triaxial measuring system is completed. So the main emphasis of registration algorithm research will focus on orientation bias correction between two triaxial measuring systems. The improved PSO is applied to establish and estimate an error model, and then the error mode was given with conversion of the two coordinate systems, so as to realize the registration and calibration of double triaxial measuring system. Simulation results indicate that the effect of calibration and registration can be very well by this algorithm, and the proposed algorithm can well advance the effectiveness of the algorithm that improve the accuracy of the optimization and increase the speed of convergence.