摘要：
在河蟹养殖水草清理过程中,为降低养殖户劳动强度和提高导航定位精度,研究结合DGPS和视觉导航的优点,设计一种用免疫粒子群算法(IPSO)来优化无迹卡尔曼滤波(UKF)的组合导航定位方法,并应用于水草清理作业船.首先通过建立组合导航模型,得到系统的状态方程和量测方程;为解决UKF对导航模型滤波存在的发散问题,再通过粒子群算法(PSO)优化UKF,并引入免疫算法避免PSO的早熟现象;最后得到滤波后新的位置坐标.为获取视觉信息,对采集的图像采用相应的图像处理技术确定导航路径.导航实验结果表明,所提方法相比DGPS导航和组合导航,纬度误差分别下降22.69％、9.14％,工作时间分别减少4.77％、4.32％,进一步提高了作业船工作效率.%In the aquatic plants cleaning process of crab culture,in order to reduce labor intensity of the farmers and improve the positioning accuracy of navigation,a kind of DGPS and vision integrated navigation positioning method was designed with immune particle swarm optimization (IPSO) to optimize the trace of Kalman filter,which combined the advantages of DGPS and visual navigation,and was applied to aquatic plants cleaning workboat.Firstly,the integrated navigation model was established,and then the state equation and observation equation of the system were obtained.In order to solve the divergence problem of UKF filtering for navigation model,PSO was used to obtain new particles,and immune algorithm was introduced to avoid premature phenomenon of PSO.Combining with UKF,the navigation model was filtered,and the new position coordinates were obtained.At last,the comparative experiment was conducted by simulation and navigation experiment.Simulation experiment results showed that the root mean square error (RMSE) at east and north positions of the proposed method were reduced by 46.09％ and 71.51％ compared with DGPS navigation,and reduced by 23.92％ and 58.26％ compared with integrated navigation,respectively.Navigation experiment results showed that in the same longitude position the latitude error of proposed method was reduced by 22.69％ and 9.14％ compared with DGPS and integrated navigation,respectively.The results showed that the navigation time of the proposed method was reduced by 4.77％ and 4.32％ compared with DGPS and integrated navigation,respectively.
