Performance evaluation of drift-free integration for navigation and velocity feedback applications

Abstract

In many applications, such as navigation or control, the vital localization or feedback information can be obtained via inertial sensors with subsequent time integration. However, the intrinsic near-DC noise existing in virtually all signals would cause signal drifting after integration and eventually disables the associated applications. Various approaches have been proposed to solve this problem. In this study, the existed drift-integration approach is re-examined for evaluating its performance and optimal design criterion through both simulation and experimental characterizations. A drift-free integrator based on a first-order low-pass filter and two second-order Sallen–Keys high-pass filters was realized using commercially available OP 741 and associated circuit elements. An electromagnetically actuated vibrating system was designed and realized for providing vibration excitation, and its acceleration and displacement were measured separately by a PCB accelerometer and a Micro-Epsilon displacement probe. Both signals are subsequently converted to velocities using the drift-free integration and standard signal differentiation for the purpose of the performance evaluation and the optimal design of the integrator. The experimental results indicated that the integrator can effectively suppress the signal drifting due to the near-DC components commonly existing in all inertial signals. However, it also was found that a trade-off exists between drifting suppression and the extent of signal distortion, and there exists an optimal design criterion based on the dynamic parameters of both circuit and mechanical systems. The conclusion drawn from this study would be useful for a designed drift-free integrator for related applications such as an accelerometer-based velocity feedback control and the MEMS accelerometer-based inertial navigation systems.