“Experts have proposed and implemented an adaptive visual inertial geomagnetic tight coupling positioning system to address the issues of absolute heading information loss and attitude divergence in visual inertial integrated navigation systems. This innovative design utilizes the principle of joint calibration of internal and external parameters of a three-axis magnetometer. By constructing global and inter frame constraint residuals of geomagnetic information, dynamic adjustment of fusion weights is achieved. A visual/inertial/geomagnetic tight coupling system is designed using nonlinear optimization methods, effectively estimating its own motion state. In outdoor experiments in campus environments, the system operated stably under the interference of magnetic fields from some buildings and vehicles, with a positioning accuracy of better than 0.8% (RMSE). Compared with traditional VINS systems, the average position error was reduced by about 24%, and good real-time performance was demonstrated. This breakthrough research not only improves the positioning performance of existing visual inertial navigation systems, but also provides highly reliable real-time positioning results for unmanned systems, marking important progress in this field.”