Accuracy Degradation Of Mems Integrated Navigation Systems Due To Poorly Aligned Inertial Sensors

Zainab Syed,Xiaoji Niu,Priyanka Aggarwal,Naser El-Sheimy

2007-01-01

Abstract:Global Positioning System (GPS) provides reliable long term navigation information but requires a direct line of sight between the receiver and the GPS satellites. On the other hand, an Inertial Measurement Unit (IMU) offers continuous autonomous solution but the accuracy cannot be guaranteed for long period of time as the errors accumulate due to the integration of its signals. The integration of GPS long-term and accurate information with the continuous but short-term accuracy of IMU systems should be able to provide accurate and uninterrupted navigation system for many difficult navigation scenarios. Vehicle navigation is, in fact, one of the challenging navigation modes as the GPS signal quality degrades significantly when the vehicle goes through urban centers or forest canopies. In addition, GPS signals are completely lost when a vehicle goes through a tunnel or passes under a bridge and therefore no positioning information can be obtained in these situations. Consequently, the integration of the GPS with Micro Electro Mechanical System (MEMS) IMU is the best solution for land vehicle navigation applications. The inertial sensors, however, need to be pre-calibrated and aligned with the body frame of the vehicle to properly estimate the position and attitude of the vehicle.With the exception of IMUs consisting of three orthogonal accelerometers and three orthogonal gyroscopes (full IMU) without any additional constraints in the vehicle frame (e. g. non-holonomic), the mounting misalignment of the IMU with respect to the vehicle body frame will be a limiting factor for the obtainable accuracies. It is expected that non-holonomic constraint results will degrade with larger misalignment angles as the constraints will set the two components of the vehicle velocity to zero. In the case when IMU and the vehicle's body frame are not aligned properly, incorrect components will be set to zero which will result in loss of information. In addition, the partial IMU which cannot estimate roll or pitch due to only one gyroscope in the vertical direction, could seriously degrade the results when there is roll and pitch misalignment. The results from this study will provide valuable design aspects for the civilian vehicle navigation units. As both with and without non-holonomic constraints will be included for full and partial IMUs, it will be helpful to choose a compromised method (if any) that will satisfy the needs of vehicle navigation.