1. Navigation Everywhere Article
   1. Sensors (Updates/Inputs)
      1. GNSS
      2. Accelerometers
      3. Gyroscopes
      4. Magnetometer
      5. Magnetic Compass
      6. Barometers
      7. Pedometers
      8. Electromyography (EMG)
      9. Odometers
   2. Fusion
      1. Kalman Filter
      2. Extended Kalman Filter (EKF)
      3. Particle Filter
      4. Monte Carlo Method
      5. Sequential Monte Carlo
      6. Markov chain process
      7. Gauss Process
      8. Bayesian inference
   3. Constraints
      1. Zero Velocity Update (ZUPT)
      2. Non-holonomic (NHC) same as Zupt but for lateral and vertical velocity
   4. Design decision
   5. Backward Smoothing (BS)
      1. Fixed-Interval
         1. Rauch-Tung-Stried Smoother (RTS)
      2. Fixed-Point
      3. Fixed-Lag
      4. Two-Filter Smoother
2. PDR MS Acadimics and Google Scholar year >= 2008 Just the first two pages of search result
   1. A Robust DR Pedestrian Tracking System with low-cost sensors
      1. Exploits the fact that user nowadays carry multiple DR systems have stable relative displacement w.r.t. the center of motion and to each other
      2. Formulate the robust tracking task as a generalized maximum posteriori sensor fusion problem
      3. Related work in DR-based Pedestrian Tracking
      4. State of the art step-based DR scheme
      5. Proposed MAP Algorithm
   2. A Comparison of PDR Algorithms using low-cost MEMS IMUs
      1. They use a MEMS IMU attached to the foot of a person
      2. Step Detection (sec 2)
      3. Stride Length (sec 3)
      4. Heading and Position Estimation (sec 4)
   3. Improving Pedestrian Dynamics Modeling Using Fuzzy Logic
      1. Different Approaches to PDR Navigation
         1. Pattern Recognition is correlated to Biomechanical Principles
         2. And Combined with Fuzzy Logic
      2. Detection and Classification of walking behaviors in 3D
         1. Forward walking
         2. Stairs climbing
         3. Stairs Decent Forward
         4. Stairs Decent Backward
   4. Dead Reckoning from the pocket (An experimental study)
   5. A Method of Pedestrian DR Using Action Recognition
      1. Localization accuracy can be improved by action recognition with use of a machine learning framework
      2. Action recognition mechanism can be refined by the estimated location and orientation with map information