Error Modeling of Reduced IMU using Recurrent Neural Network

Although GNSS/IMU integration has been studied for decades, an efficient estimator of their integration has remained a challenge. In the statistical approaches, the observation model of sensors and distribution of data must be known beforehand. This paper proposes a deep learning based approach to integrate GPS and reduced IMU information. In contrast to statistical approaches, our approach learns the observation model and data distribution of sensors. We apply a Recurrent Neural Network (RNN) with one hidden layer and train our network with reduced IMU as input and GPS observations as target values. The Results show our network is properly trained and reduced IMU errors are accurately predicted for a few seconds. Keywords—Inertial Measurement Unit (IMU); Reduced IMU; Global Positioning System (GPS); Recurrent Neural Network (RNN); Long Short Term Memory (LSTM); Reduced Inertial Sensor System (RISS)