Intelligent Human Motion Classification and Quantification

US soldiers increasingly have the need to operate in GPS-denied environments or in environments without access to GPS, such as caves and indoor buildings. To date, inertial integration and pedometry have been used for position estimation in such environments. The inertial integration process suffers from excessive velocity drift over time and leads to errors in position estimation. Pedometry avoids velocity drift, but depends on an estimated scale factor that converts steps into distance traveled. A fundamental problem with direct pedometry for position estimation in GPS-denied environments is that a scalar step count cannot be translated to a vector displacement for an accurate position update. SSCI's approach for position estimation employs segmentation, classification, and quantification techniques for position estimation. The segmented motion classes will be used as inputs to a transfer function that is aware of different scale factors for different motions. SSCI is teaming with Draper Labs for this effort. Draper Labs is one of the world’s premier sources of guidance, navigation, and control (GN&C) expertise and has the domain experience, advanced sensor technology, comprehensive algorithm and software design and development skills, and testing capabilities needed to support SSCI in the development of an improved and innovative personal navigation system.