Dead reckoning represents a class of methods for relative position estimation based on a previously determined absolute reference position. The estimate is formulated from a combination of the known speed, time and heading information with the known reference position. One of the main obstacles to effective positioning of pedestrians via dead reckoning is the lack of accurate speed estimation algorithms. Existing methods are either complex or provide results that are unsatisfactory at the low velocities associated with pedestrians. In contrast, the two algorithms proposed in this paper are relatively simple to implement and provide accurate results at low velocities. In the first algorithm, a one-dimensional and unidirectional two-antenna solution is described where the speed can be easily estimated from a knowledge of the fixed inter-antenna distance and the time it takes for the trailing antenna to sense the same channel conditions (radio frequency (RF) signature) previously observed at the leading antenna. Computer simulations show that, with typical estimation errors of less than 2.67% around average pedestrian speeds, the approach is indeed effective and accurate. A by-product of the algorithm is an environment specific spatial correlation function which is used in the second algorithm to provide even better estimates. With the improvements offered by the latter algorithm, relative errors of merely around 0.15% on average are achievable. This improvement in performance over the first algorithm comes at the cost of slightly higher computational complexity. When subsequently used for user displacement estimation, a relatively small error of 24.5cm is observed after a duration of 60s.
展开▼
