Base Station Placement in Multi-Floored

Multi Buildings
Department of Electrical and Computer Engineering Liza Pujji, Kevin Sowerby and Michael Neve

AIM CASE STUDY RESULTS

To find the optimal base station configuration for a Potential Base Station (BS) sites Optimal BSP and of calls connected
multi-floored building from a set of internal and Scenario Traffic Case I Case II
Tower
external potential base station sites considering a Internal BS sites only Internal and External
range of traffic scenarios. BS sites
OPTIMISATION ALGORITHM [1, 2] Floor 8 5
6
I Very 3 50% 3 50%
Floor 7 3
4 Low 5 48.5% 5 48.5%
The RCR algorithm is divided into two stages: 12
1
Library 8 10 2 II Low 3 28.5% 4 41%
1. Reduction Estimation; and 11
7
9 4 32% 6 39.5%
2. Combinatorial Optimisation with or without 6 39% 7 18.5%
Reduction Approximation depending on the size
III Med. 3 28% 4 41%
of the system. 4 32% 6 39%
Input set of Base Stations (bs), 6 38.5% 7 18.5%
mobiles (u) and propagation
conditions IV High 3 23% 4 35.5%
• 12 Potential base station sites. 4 27.5% 5 23%
STAGE I - Reduction Estimation
(Estimate the minimum number of base stations required, ܰ݉݅݊ ) Vertically aligned sites : (1,3, 5) and (2, 4, 6). 5 22% 6 27.5%
6 27% 8 13%
Internal_user floor sites : 3 -6 .
Iteration, ݅ = 0 and call failure rate is 100%
V Very 3 23% 4 35.5%
• 108 potential user locations on floor 7 and 8.
High 4 27% 5 23%
• Measured propagation data [3]. 5 22% 6 27%
Is Combinatorial
NO STAGE II - Reduction Approximation
Optimisation
(Reduce the number of BSs, ܾܰ‫) ݏ‬
• WCDMA system model assumed. 6 26.5% 8 13%
possible?
• Average call failure rate is 2% [4].
YES • Five traffic scenarios (with 501 trials) considered ― CONCLUSIONS
݅ =݅+1
Very low (5 erlangs) to Very high (25 erlangs).
RCR algorithm can be applied to find optimal
STAGE II - Combinatorial Optimisation
(Find Optimal BS Configuration, ‫) ݅ܲܵܤ‬
Dynamic call schedules ― erlang model [4] BSP for multi floor buildings under dynamic
• Call arrival rates ― Poisson distribution. conditions.
• Call duration rates ― Negative exponential
Vertically aligned internal_user floor base station
Is the call failure rate for
ܰ݉݅݊ = ܰ݉݅݊ − 1
YES
both distribution. sites are preferred as they produce minimum call
‫ ݅ܲܵܤ‬. ‫ܦܰܣ‬. ‫݅ܲܵܤ‬−1 ≤ 2%
• Blocked calls are cleared. failure rate.
NO

As traffic increases, an external site is selected
over an internal site because it causes less
REFERENCES
ܰ݉݅݊ = ܰ݉݅݊ + 1
YES
Is the call failure rate for
both [1] L. K. Pujji, K. W. Sowerby, and M. J. Neve, “Base Station Placement in Multi- interference.
‫ ݅ܲܵܤ‬. ‫ܦܰܣ‬. ‫݅ܲܵܤ‬−1 > 2% Floored Buildings ,” 20th Virginia Tech Symp.
Symp on Wireless Pers. Commun., 2010.
If the traffic is too high , another internal_user
[2] L. K. Pujji, K. W. Sowerby, and M. J. Neve, “A hybrid algorithm for efficient
optimisation of indoor base station placement,” submitted to IET Commun., floor site is selected to cope with the load.
NO
If call failure rate for ‫ ≤ ݅ܲܵܤ‬2%
2010.
The internal base station sites serve the greater
[3] K. S. Butterworth, Performance, planning and deployment of DS-CDMA in-
ܵ‫݅ܲܵܤ = ݊݋݅ݐݑ݈݋‬
Else ܵ‫݅ܲܵܤ = ݊݋݅ݐݑ݈݋‬−1 systems PhD thesis, The University of
building wireless communication systems. proportion of calls with maximum calls served by
Auckland, New Zealand, 2000. vertically aligned internal sites.
[4] M. Tolstrup, Indoor Radio Planning: A practical guide for GSM, DCS, UMTS and
HSPA. John Wiley & Sons Ltd, 2008.