Our target is making a low-cost, Wearable telemetry device that can estimate position and orientation accurately. it can store retrieved data on internal memory (8GB microSD) or transmit data in real-time over Wi-Fi network into a server that could be local or on the internet. This device has several sensors such as a gyroscope, accelerometer, magnetometer, thermometer. It could be very useful in sport, rehabilitation, tracking human daily activity and so on. For verifying accuracy of IMU output we have compared it with data acquired simultaneously by Vicon motion capture system.
Madgwicks algorithm is implemented to eliminate gyroscope drift error for more accurate estimation of position and clear it from cumulative error caused by double integration, we want to use regression methods and unsupervised ANN.