Navigation is the science and art that answers the questions of knowing where you are at the current moment and where you will be in the next moment. Modern navigation systems are based mainly on satellite and inertial sensors. Inertial sensor systems are becoming very popular in navigation systems because they are self contained sensors. The goal of this research is to develop novel approaches for improving the performance of inertial sensor systems and their integration algorithms with external sensors such as global positioning system (GPS) and magnetometers. The standalone inertial navigation system (INS) is dependent on the inertial measurement unit (IMU). An IMU is traditionally composed of three orthogonal gyroscopes and three orthogonal accelerometers.
In the inertial sensors side, we focus on the use of distributed accelerometers for inferring the angular motion from the angular information contained in their measurements. There exists a variety of reasons for conducting this research. In short, high quality gyros have high cost, high power consumption, large weight and large volume. On the other hand, accelerometers are less costly, easier to manufacture, have less power consumption and less weight than gyros. We developed different fusion approaches for benefiting from the angular information vector (AIV) resulting from the distributed accelerometers to form a gyro-free IMU (GF-IMU) or to aid the GF-IMU by conventional gyros. By improving the performance we mean reducing noise and bias level in the estimated inertial quantity.
In the integrated navigation side, we present different approaches to implement the GPS/INS integration filters and the attitude and heading reference system (AHRS) algorithms. We use direction cosine matrix (DCM) based algorithms which implies estimating the elements of the DCM directly within the filter. The basis for this method is the ground alignment method for attitude and heading determination. The attitude update of the DCM is performed using angle rotation vector. The filter is able to detect the gyro bias vector and follow its variation and hence it fits low-cost sensors as well as high grade sensor. We validated the efficiency of the algorithms using proper simulations and real-time implementations.