The Experimental Study of MIMU/BeiDou Integrated Navigation System for Land Vehicle Applications in Highly Poor Weather Conditions

The introduction of Micro-Electro-Mechanical-System based IMU (MIMU) is a bold and meaningful conception for small unmanned aerial vehicle (UAV) autonomous approaching and landing. The adoption of a low-cost MIMU with high noise, which is suitable for UAV applications with advantages in cost-effectiveness, lightweight, miniature design, low power consumption, and survivability. However, complicated flight dynamics and MIMU noise pose a great challenge for precise navigation of UAV autonomous landing. In this paper, a precise GNSS/MIMU integrated navigation scheme is derived to achieve autonomous landing, based on STIM-300 MIMU and BeiDou/GPS multi-frequency multi-mode precise relative positioning. The precise MIMU error model is established based on Allan variance analysis to resist the influence of noise uncertainty of MIMU. A real-time improvement is introduced in the conventional data fusion algorithm considering the delayed observations and time synchronization. Moreover, a novel in-flight coarse alignment of MIMU is adopted to guarantee accurate and reliable fine alignment afterward. The land vehicle test results indicate the proposed algorithm can complete accurate and reliable navigation (CatIIIc precision requirements) even in GNSS outages (<= 12 s). As a cost-effective scheme, it has significant application potential for small UAV autonomous approaching and landing in highly poor weather conditions such as heavy fog and snow.