The proposed project aims to pursue an innovative approach of utilizing deep learning algorithms to improve the performance of state-of-the-art MEMS Inertial Measurement Units (IMUs) to push their performance to fullest extent possible as replacements for existing Fiber Optic Gyroscopes (FOG) systems for navigation and guidance in future missile system avionics including Sounding Rockets. During Phase I, the proposed team will evaluate artificial intelligence algorithms to model noisy MEMS sensor data. A deep learning methodology will be employed to effectively define complex, compositional nonlinear functions, to learn distributed and hierarchical feature representations, and to make effective inference of both labeled and unlabeled data. We plan on implementing machine learning algorithm on a FPGA platform after identifying the most promising software algorithms. The Phase I program is ambitious by design but we believe that research plans are realistic and achievable because of our expertise and resources accumulated from on-going research in the areas of inertial sensor design, machine learning, and microelectronics. The prototype hardware is useful for ground testing and to demonstrate that the proposed solution can meet flight environments. At the end of Phase II, a prototype will be ready to be flown on a sounding rocket flight test.Approved for Public Release | 17-MDA-9395(24 Oct 17)