Comparison of two different mass-market IMU generations: Bias analyses and real time applications

The inertial positioning is becoming a technique gradually more and more used both in geomatics and in professional world, thanks to the improvement of the Inertial Measurement Unit (IMU) platforms performances combined to the reduction of the electronical element size and cost. A mass-market IMU platform has today costs of about few euros: this kind of instruments are installed in a widely instruments, typically used for integrated positioning such as smartphones, Mobile Mapping System (MMS) or Unmanned Aerial Vehicles (UAV), where the requested accuracy is lower. These kind of devices are traditionally called as "mass market", considering the level of distribution and the cost, but the quality and performance of the internal sensors could be completely different. In the case of low cost IMU is also possible to obtain completely different performance, even considering similar devices. In this work the performance of two IMU mass-market platforms have been analyzed, used in different positioning modes (static and kinematic), integrated also with GNSS single frequency receivers, analyzing the positioning accuracies obtainable in the presence or absence of obstacles that obstruct the satellite visibility. The comparison of these results will be carried out using a tactical grade platform, defined as a reference. Firstly some bias analyses are made, focusing the attention on errors estimation: the achievable accuracies in static mode will analyzed, estimating the noise through the classical tests available in the literature. These test were carried out in laboratory, with purpose to work in "free magnetic field" condition, considering that several IMUs contains an internal magnetometer which is disturbed by eventual artificial magnetic field. The Allan Variance of each sensors were estimated in order to identify eventually the flicker noise, white noise and random noise in the single devices. Each dataset (acceleration, angular velocity, magnetic field) has been also analyzed with wavelet filter, with purpose to remote the noise, using the filtered data for estimating the navigation solution. After that, the precision and accuracies obtainable in real conditions was investigated; these performance have been evaluated analyzing some kinematic trajectories and considering a low-cost Mobile Mapping system (MMS), which belongs to the Geomatics Group of the Politecnico di Torino. Accuracies of positions and attitude will be evaluated considering different methods of data processing in post processing approaches. Both the loosely coupled (LC) and tightly coupled (TC) approaches are followed and the results obtained with all two methods are shown and described. To do this, the Inertial Explorer 8.60 post processing software was used to determine positions and attitudes. In this study, an interesting comparison about two different mass-market IMU generations is realized, in particular for mobile mapping applications.