Presentation 11/13: Experiments on local positioning with Bluetooth
Su Kim
Title: "Experiments on local positioning with Bluetooth,"
Published: ITCC2003, 2003
Su Jin Kim
jun shen
by Jun Shen
Overview:
The paper mainly deals with how to achieve local positioning with
Bluetooth (bt) connections. Positioning is based on the power level of
received bt signals. And exact position of an object can be calculated
by adopting extended Kalman filter.
Contribution & Strength:
1. The authors design a way of local positioning with the low-price bt
technology. The advantage of the technology is that no infrastructure
is needed except a bt-enabled PC.
2. Besides, the authors present the implementation of their method.
3. The paper even explores what the uncertain variable is in the
positioning computation.
4. The paper defines a model to evaluate the accuracy of positioning.
Weakness:
1. The error range of the positioning is too high. The effective range
of bt is always around 10 meters and the error can be 3.76 meters.
2. In the computation, some details are omitted. For example, the
paper fails to mention why h(Xk) can be approximated using Taylor's
series expansion with first-order terms. Why not use more terms?
3. This technology is not practical because bt signals interfere with
those of WLAN which has a wide coverage.
4. there can not be obstacles between bt transmitter and receiver.
This requirement will significantly limit the usage of the positioning
technology.
5. The paper claims that the accuracy of bt signal receiver influence
the accuracy of positioning, however, the paper does not prove it.
On Nov 13, 1:29 pm, "Su Jin Kim" <sujin...@gmail.com> wrote:
> Please find out the reference for Kalman Filter herehttp://robotica.itam.mx/espanol/archivos/kalman_filter.pdf
> (Applications of the Kalman Filter Algorithm to Robot Localisation and World
> Modelling)
> <http://robotica.itam.mx/espanol/archivos/kalman_filter.pdf>
weiji...@asu.edu
Experiments on Local Positioning using Blue-Tooth
Critique : Weijia Che
SUMMARY:
The paper explores a method to triangulate the user device position
through
measuring received signal power levels. The RSSI (received signal
strength indication)
is first convert to absolute RX power levels, then an EKF (extended
Kalman filter) is
adopted to computer the positions from the distance measurements.
Finally an EBSK
(Ericsson Bluetooth Starter Kit) is used for implementation and
evaluation.
PROs:
EKF is adopted to diminish the mean Error.
The EBSK is simple and easy to use which indicates that for
both researchers
and normal users, it will be easy to setup without
exhaustive experience and
horrible cost.
CONs:
The BLPA (Bluetooth Local Positioning Application)
implemented in this
paper requires neighbor devices to be stationary and their
positions to be
configured manually.
The position estimate is inaccurate due to the unreliability
of RSSI values.
Absolute RX power level is hard to obtain.
The situation without line of site TX, RX is not explained
though it should be
a case for consideration.
Best
Weijia Che