A Novel Approach for Attitude Estimation Based on MEMS Inertial Sensors Using Nonlinear Complementary Filters

In this paper, a new method for attitude estimation using low-cost micro-electro-mechanical system (MEMS) inertial sensors is developed, in order to obtain accurate attitude parameters when other external sensors, such as the magnetometer and GPS, are limited in operation. The method is based on specific rotation of the inertial measurement unit, which is independent from rigid body motion. The problem of heading angles estimation using just inertial sensors without incorporating additional sensors is studied. An algorithm is developed by incorporating a nonlinear observer in order to fuse measurements from both the gyroscope and the accelerometers. The algorithm provides bias estimation of the gyroscope and better estimation results. Both numerical simulations and real experiments validate the effectiveness of the proposed method.