Validation and Performance Evaluation of a Simulink Inertial Navigation System Simulator

Abstract

Great interest has been generated in low-cost inertial navigation systems (INS) in the last few years. The development of Micro-Electro-Mechanical Systems (MEMS) in the last decade has permitted mass production of devices, thereby reducing the cost of previously expensive sensors. Simulation is part of the design process of an INS. However, if we exclude proprietary simulators, simulation package tools available until now to achieve this goal are Matlab script files. To realize fast prototyping systems, to take advantage of modular design and to allow rapid real-time testing, a Simulink-based INS simulator has been created. This simulator is designed in several modules allowing point-wise improvements or modifications that do not affect the overall modules. This paper addresses the validation and the performance evaluation of the simulator. Given the results of two validation methods, it can be concluded that the Simulink simulator can now be used to design inertial navigation algorithms. Also, performance analysis of built-in Simulink integration schemes, compared with the well-known Savage two-speed integration algorithm, indicates that the automated-code integration algorithm is suitable for inertial navigation applications. In fact, for similar processing time, the position error of the 4th-order Runge–Kutta solution, compared to the Savage one-speed simplified integration algorithm, is improved by 8% for low-sample-rate inertial measurements. The Simulink simulator will be an important tool in the design of innovative algorithms suitable for use in conjunction with MEMS inertial sensors in low-cost navigation systems. Upon completion, this software will permit rapid prototyping and easy-to-use design for low-cost GPS-aided electronic inertial navigation systems.