The Application Of Multi-sensor Information Fusion In Tower Crane Positioning

The institutions and structure of tower crane will bear strong impact in itsstarting, braking and coupling motion, the accurate transportation of tower craneis very difficult, so the accurate positioning of tower crane is a major premise toflexible and efficient control. The arm, car and load of tower crane can bepositioned accurately by information fusion technology every moment, thenfeed back the optimal estimation to the control system of tower crane timely andaccurately, the production efficiency can be improved and the number ofaccidents reduces.The positioning system of tower crane based on Kalman filteringinformation fusion technology is researched after the analysis of structure oftower crane and development trend. The specific research contents are asfollow:Firstly, the ideal linearized dynamics model, the equations of state andmeasuring are set up with simple variable relationship based on the analysis oftower crane’s structure. The federal Kalman filtering technology is introduced,and the positioning system of tower crane based on the federal Kalman filteringis set up. The using of distribution coefficient for federal Kalman filteringtechnology is improved: changing following the noise. It greatly restrains thenoise influence and complete the accurate positioning of tower crane’s linearsystem.Secondly, after depth analysis of tower crane’s structure，we find thatlinear positioning is only for variables can be detected directly and accurately.In order to solve this problem, the nonlinear dynamic model of tower crane isestablished, the Uncented Kalman Filtering (UKF) is introduced, the federalKalman filtering and UKF are combined, the missing data is largely reduced,the complex calculation of Jacobian matrix is avoided, the positioning precisionis improved.Thirdly, the tower crane load is difficult to positioning because of itscomplex three dimensional motion features. We introduce the least square method, and combine it with Kalman filtering to complete the three dimensionalpositioning. We make out the improved double least square Kalman filteringmethod, and each of the three coordinates makes Kalman filtering alone. Theamount of calculation is reduced, the positioning precision of the load isimproved.