dead reckoning from quaternion and accelerometer data

We have hardware and firmware to capture quaternion and accelerometer data from a wearable device. Your job would be to fuse these using a Kalman Filter and apply an algorithm to derive the 6 DOF dead reckoning. We seek high accuracy, which should be possible given the chosen sensor and other problem constraints. The successful candidate will be able to cite examples where they have successfully executed a similar real-time algorithm.