Documentation : Hardware | Software | Tutorials    Wardrobe : Jackets | Applications
Papers : Notes | Proposals | Research    Administration : Stock | Minutes | Progress

This page details the design of an indoor positioning system for the Computer Science Atrium at Nottingham.

The original indoor positioning system was designed to work in a conventional office of around 4m x 4m with a ceiling height of 3m. With careful selection of  transducers we were able to extend the coverage to 8m x 8m.

23rd March 2001 - Genesis

Following the Playing and Learning Workshop, the feasibility of providing coverage over Nottingham's 10m x 30m Atrium was explored. The Atrium has a high ceiling which should enable a larger area to be covered. An initial test demonstrated that the prototype system worked in the Atrium but was confounded by the planting and glazed screens. These problems should be alleviated by raising the ultrasonic transducers to raduce the shadoews cast by these objects.

An expected coverage using six transducers is shown below:

atrium3.gif (21866 bytes)    

3rd/4th/5th July 2001 - First Installation and Proof of Concept

The first installation was carried out using six transmitting devices in a 9m x 9m grid at a height of 7.23m. By transmitting the chirps 17.4ms early we were able to retain the 50ms repetition rate. This meant that the receiver had to be at least 6m from the transmitters before valid data was received (i.e. less than 1.23m above the ground).

Main problems encountered were:

1) Corruption of RF Ping signal - the first byte of each ping was intermittantly corrupted - many potential source of interference were investigated without identifying the source. The solution was to ignore the first byte.

2) Insufficient range of ultrasonic signals - while satisfactory results (range >14m) had been obtained with the original transducers, the new transducers with improved omnidirectional directivity have a shorter range. We experimented with different drive voltages and the following results were obtained:

Drive Voltage (supply) Range (100us units) Range - metres
     
12 80+174 = 254 8.64m
18 100+174 = 274 9.32m
24 120+174 = 294 10m
30 140+174 = 314 10.68m

3) Signal Obstruction - the furniture, screens and light fittings all provided serious signal obsrtuction. In the longer term the furniture and screens will be removed, however we are left with the problem of the light fittings.  

Revised Design

To address the problems outlined in 2) and 3) above we decided to move to a 6m x 6m grid using eight transducers. This would compensate for the poor range, and also, by not having the transducers directly above the light fittings, the shadowing effect would be reduced. Separately, research will be carried out into extending the range further by increasing the sensitivity of the receiver.

Results

The drawing to the right shows the results obtained with four transducers in a 6m grid. The axes are calibrated in mm.  Two of the transducers are centred on the horizontal '0' x axis with the other two centred on the -6m y axis. The effect of reflections off the LH wall can be seen as giving a spread of x values.

atrium_1.gif (2573 bytes)

(source data)


This page last updated April 23, 2002
The material displayed is provided 'as is'. Contents may not be reused without prior permission.
For problems or questions regarding this web contact Cliff Randell.