Machine Learning Aided Gait Recognition for Inertial Navigation and Orientation - Year 1

This report details the system, test environment, and results used to evaluate a Global Navigation Satellite Systems (GNSS) denied pedestrian inertial navigation system that is aided with velocity estimates from a machine learning algorithm. A machine learning algorithm was developed and trained with data from foot-mounted Inertial Measurement Units (IMU) and GNSS data from a user to estimate the users velocity. After the machine learning algorithm is trained, the algorithm can estimate the users velocity with only the foot-mounted IMU data. The velocity estimates are combined with data from a back-mounted IMU and an Extended Kalman Filter (EKF) to estimate the users position without GNSS data. The system will be evaluated with different terrains and multiple data collects to measure performance across different conditions. The data collections and evaluations described in this report show that the system can estimate the users position with a range of percent error over distance traveled of 5 to less than 1 . It also shows that the system can work with different terrains and gaits including slow walking, walking, and running.