ICOM 6505: Wireless Networks Cellular Networks

1
ICOM 6505 Wireless Networks- Cellular Networks -

• By Dr. Kejie Lu
• Department of Electronic and Computer Engineering
• Spring 2008

2
Outline

• Overview
• Channel assignment
• Fixed channel assignment
• Dynamic channel assignment
• Location management
• Handoff

3
Cellular Network Architecture

• Cellular System
• Provide wireless coverage to a geographic area
  with a set of slightly overlapping cells
• Network Components
• Mobile Station (Terminal) handset
• Base Station (cell site) - provides radio
  channels between mobile units and network.
• Base Station Controller (BSC) - manages a cluster
  of BS, channel assignment, handoff, power
  control, some switching, etc.
• Mobile Switching Center (MSC)- provides switching
  functions, coordinates location tracking, call
  delivery, handoff, interfaces to HLR,VLR, AUC,
  etc..
• HLR/VLR/AUC (Home Location Register/Visitor
  Location Register/Authentication Center)
  databases to track, bill and authenticate users

4
Cellular System

• Geographic region is divided into cells
• Channel reuse
• Frequencies/timeslots/codes reused at
  spatially-separated locations
• Co-channel interference between cells that use
  the same channel
• BS/MSC coordinate handoff and control functions
• Shrinking cell size increases capacity, as well
  as networking burden

5
Cellular System
Handoff
6
Cellular Networks

• Propagation models represent cell as a circular
  area
• Approximate cell coverage with a hexagon -
  allows easier analysis
• Cluster of cells K group of adjacent cells
  which use all of the systems frequency assignment

7
Important Design Issues

• Wireless networks provide communication services
  to a large number of mobile users
• The design of such a network is based on cellular
  architecture that
• Allows efficient use of limited available
  spectrum
• Channel assignment
• Tracks mobile users to establish communication
  with any particular user
• Location Management
• Ensures continuous communications when users move
  from one cell to another while the communication
  is active
• Handoff

8
System Capacity

• System Capacity is the number of all users that
  can communicate (use the system) at the same time
• A base station (cell) has a fixed number of
  channels available, hence at a given time a fixed
  number of users can talk simultaneously

9
Channel Assignment

• Fixed channel assignment
• Dynamic channel assignment
• Adaptive channel assignment

10
Fixed Channel Allocation (FCA)

• FCA is the most common channel allocation method
• It assigns frequencies to each cell/sector
  requiring channels beforehand and the frequencies
  assigned stay fixed during network operation
• Cellular providers may chose to adjust their
  fixed plans from time to time
• Frequency reuse factor determination
• Radio interference limits the number of radio
  channels that can be used in a single cell site
  and how close nearby cell sites that use the same
  frequency can be located together
• The main types of interference are co-channel
  interference and adjacent channel interference

11
Radio Interference

• Co-channel interference
• Co-channel interference occurs when two nearby
  cellular radio operating on the same radio
  channel interfere with each other
• Adjacent channel interference
• Adjacent channel interference occurs when one
  radio channel interferes with a channel next to
  it
• For example, channel 412 interferes with channel
  413
• Each radio channel has a limited amount of
  bandwidth (e.g., 30 KHz wide), but some radio
  energy is transmitted at low levels outside this
  band
• A cellular radio operating at full power can
  produce enough low-level radio energy outside the
  channel bandwidth to interfere with cellular
  radios operating on adjacent channels

12
Distance Between Two Hexagonal Cells

• Consider a coordinate system UV as below
• The center of each cell can be represented by (U,
  V), where U and V are both integers

13
Distance Between Two Hexagonal Cells

• Distance D between two cells C1(U1, V1) and
  C2(U2, V2)
• D (U2 U1)2 (COS 300)2 (V2 V1) (U2
  U1) SIN 30021/2
• (1)
• D (U2 U1)2 (V2 V1)2 (V2 V1) (U2
  U1)1/2
• (2)
• Let (U1, V1) (0, 0) and C2(U2, V2) (i, j),
  then D (i2j2ij)1/2 (3)

14
Distance Between Two Hexagonal Cells

• The normalized distance between adjacent cell is
  1
• However, the actual center-to-center distance
  between two adjacent cells is
• 2R COS30o (4)
• where R is the center to vertex distance
• Co-channel interference is a function of q where
  q D/R

15
Distance Between Two Hexagonal Cells

• From (3) and (4), D2 3R2 (i2 j2 ij) (5)
• Area of large hexagonal l
• Al k3R2(i2 j2 ij), where k is a constant
• Area of small hexagon s
• As kR2

16
Distance Between Two Hexagonal Cells

• From the above discussions, we have
• And ... (6)
  
• Now Assume the frequency reuse factor is N
• Then the total number of cells enclosed in Al is
• N 6(1/3)N 3N cells
• Since there is only one cell in As, we have
• ... (7)
• From (6) and (7) we have and

17
Co-channel Interference Ratio

• S/I is received signal strength to interference
  signal strength ratio, where Ik is co-channel
  Interference from a cell and r is terrain factor

18
Co-channel Interference Ratio

• In normal cellular practice, S/I is 18db or
  higher and the terrain factor r 4
• So, the frequency reuse factor shall be N 7

19
An Example

• Example
• A cellular system with 395 voice channel
• Traffic is uniform with average call holding time
  120 seconds
• Call blocking probability 2
• Terrain factor r 4
• Frequency reuse factor N 4
• Calculate
• (a) Number of calls per cell site per hour
• (b) S/I

20
An Example

• Solutions
• (a)
• Number of channels per cell 395/4 99
• With 2 blocking, the total traffic carried by 99
  channels (using Erlang-B formula) 87
  Erlangs/cell, so the number of calls per hour
  (873600)/120 2610 calls/hour/cell
• (b)

or, 10 log 24 101.38 14 dB
21
An Example

• Using the same formula, the results for N 7 and
  N 12 are given in the Table in the following

From the results in the table, by increasing the
reuse factor from N 4 to N 12, the mean S/I
ratio is increased from 14 dB to 23.3 dB (a 66.4
improvement). However, the call capacity of the
cell site is reduced from 2610 to 739 calls per
hour (a 72 reduction).
22
Directional Antennas with Sectors

• In order to reduce co-channel interference, the
  omni cell can be divided into sectors and use
  directional antennas

Mobile station to cell site interference
Cell site to mobile station interference
23
Graph Modelling

• We discussed FCA method using distance D between
  co-channel cells such that
• Transmission quality requirements such as the
  minimum signal-to-interference ratio (S/I) are
  met in all cells
• Even if all cell at a mutual distance of D or
  greater are using the same channel simultaneously
• This distance D is known as the reuse distance
• A cell can use a channel if no other cell within
  distance D is using the channel
• It can be represented by a graph model
• In the graph representation of a cellular system
• Each vertex represents a cell
• An edge exists between two vertices if and only
  if the distance of corresponding cells is less
  than the reuse distance D

24
Graph Modelling

• Example
• Assume the cell radius
• Assume the reuse distance D 2
• In 7 cell system, the distance between adjacent
  cells

Independent set a set of cells which can use a
channel simultaneously. 6, 3, 4, 1, 2,
5 Distance between cells gt reuse distance D
(2)
25
Graph Modelling

• Disadvantages
• The graph model is brought out by studying a
  regular hexagonal system.
• The cell distance between cells in hexagonal
  system
• In this model, there can only be discrete
  distance D, such as 1, , 2,
• The worst case transmission quality in the system
  depends on the reuse distance D only
• There can only be discrete values of the worst
  case transmission quality possible, and these are
  generally quite far apart
• Suppose, the required transmission quality falls
  between any two distance values say T1 and T2 and
  the corresponding reuse distance are D1 and D2.
• Then the system has to settle for the
  transmission quality better than that which is
  required
• Full potential for channel reuse offered by
  system is not realized.

26
Hypergraph Graph Modelling

• Hypergraph H (V, E) where
• V is the set of vertices, each cell corresponds
  to a vertex
• E is the set of edges
• Forbidden set
• A group of cells that all cells in the set can
  not use a channel simultaneously but cells in
  subset can use a channel simultaneously
• Independent set
• A set of cells which can use a channel
  simultaneously. Independent set can not be
  forbidden set
• Hypergraph modeling removes the weakness of graph
  modeling
• Hypergraph modeling of a cellular system offers
  much greater reuse of channels

27
Hypergraph Graph Modelling

• Interference representation
• In hypergraph modeling, the interference produced
  in cell U due to using of the same channel in V
  is d(U, V)-4, where d(U, V) is the
  center-to-center distance between cells U and V
• Total interference produced in cell U
• where C is the set of cells using the same
  channel as U

28
An Example

• Example
• Assume the cell radius , hence the
  distance between adjacent cells is 1.
• Let the required transmission quality be that
  the maximum interference must be lt 1/5.
• In a 7 cell system

In the following, use Graph Modeling Hypergraph
modeling
29
Using Graph Modelling

• Graph Modeling
• Discrete distance 1, , 2,
• Discrete distance produces a maximum
  interference of 2/9
• Interference between 2 and 4
• Interference between 2 and 6
• So total interference
• The total interference 2/9 is greater than the
  required maximum interference 1/5, hence cells 2,
  4, and 6 can not use the same channel
  simultaneously
• So, discrete distance can not be used as
  reuse distance

30
Using Graph Modelling

• Discrete distance 2 produces a maximum
  interference of 1/16
• Interference between 2 and 5
• So total interference
• The total interference 1/16 is less than the
  required maximum interference 1/5, hence cells 2
  and 5 can use the same channel simultaneously
• The interference produced between cell 2 and 4 is
  1/9 which is less than required maximum
  interference 1/5. But the cells 2 and 4 can not
  use the same channel in the graph modeling
• The discrete distance 2 is selected as the reuse
  distance D 2

31
Using Hypergraph Graph Modelling

• Hypergraph Modeling
• The adjacent cells produce interference 1/14 1
  which is greater than required maximum
  interference 1/5.
• The straight lines and ovals represent edges
• The small circles represent vertices
• The interference produced between cells 2 and 4
  is 1/9 (lt 1/5). Hence cells 2 and 4 form an
  independent set. Hence cells 2 and 4 can use the
  same channel at least sometimes.
• But the cells 2 and 4 can not use the same
  channel if cell 6 use the same channel since the
  total interference produced is 2/9 which does not
  meet maximum interference requirement (lt 1/5).
  In this case a forbidden set 2, 4, 6 is
  generated and it is represented as an edge in the
  form of oval.
• In the hypergraph modeling, 2, 4 is independent
  set but in the graph modeling the cells 2 and 4
  can not generate a independent set, so they can
  never use the same channel together.

32
Linear Programming Approach

• The channel assignment problem is formulated into
  a mathematical programming model
• The channel assignment problem is represented by
  compatibility matrix, requirement vector and
  channel service matrix
• Compatibility matrix
• The channel compatibility constraints in N cell
  cellular network described by NxN symmetric
  called the compatibility matrix C
• C11 C12 C13 C1N
• C21 C22 C23 C2N
• C C31 C32 C33 C3N
• CN1 CN2 CN3 CNN

33
Linear Programming Approach

• Compatibility matrix (Cont.)
• Each non-diagonal element Cij in C represents
  the minimum allowable separation distance in the
  frequency domain between a channel assigned to
  cell i and a channel assigned to cell j.
• When Cij 0, there is no restriction on the
  channels i and j used by cell i and j. The
  channel i and j could be the same. In this
  case, the channel i is used by cell j.
• When Cij 1, it is called co-channel
  constraint.
• When Cij 2, it is called adjacent channel
  constraint.
• The required separation among channels used by
  the same cell is expressed by the diagonal
  elements Ckk of matrix C. It is called co-site
  constraint, where the condition Ckk gt 1 is
  always satisfied.
• In order to assign any channels in cells, the
  channel distance between two cells has to be
  greater than or equal to Cij.

34
Linear Programming Approach

• Compatibility matrix (Cont.)
• Example
• Suppose a system has three cells, cell 1, cell
  2, and cell 3. Available channels are 431, 432,
  433.
• And defined compatibility matrix C is
• C11 C12 C13 1 1 0
• C C21 C22 C23 1 1 1
• C31 C32 C33 0 1 1
• 1) Try to assign channel i 431 to both cell 1
  and cell 2.
• 2) Try to assign channel i 431 to cell 1 and
  channel j 432 to cell 2.

35
Linear Programming Approach

• Requirement vector
• The channel requirements for each cell in N-cell
  cellular network are described by an N-element
  vector with non-negative integer elements.
• R r1, r2, , rj, , rN
• Each element R(j) rj in R represents the
  number of required channels to be assigned to
  cell j.
• Channel Service Matrix
• In a cellular system with M available channels
  in N cells, the channel service matrix S is an
  MxN binary matrix. The elements Sij of S are
  binary variables.
• Sij 0, if channel i is not used at cell j
• Sij 1, if channel i is used at cell j

36
Linear Programming Approach

• Problem Statement
• Given - The compatibility matrix C
• - The requirement vector R
• - The number of the available channels M
• Determine the channel service matrix S so that
  it satisfies
• - The compatibility constraints (described by
  C)
• - and the required (traffic) constraints (given
  by R)

37
Mathematical Model

• The problem statement is described in an Integer
  Programming problem. The objective is the channel
  service matrix S needed to satisfy both the
  compatibility and requirement constraints.
• The compatibility constraints ensure that in any
  incompatible pair of cells (Cij ? 0) each channel
  i can be used only once.
• Sij Sij 1, (i, j), (i, j) I
• where I is the set of incompatible assignments.
• where Cjj is the matrix C(j, j) element.
• The resulting channel service matrix S must
  satisfy the requirement for each cell. The
  requirement constraint is written as
• j 1, 2, , N
• where rj is the R(j) element of the requirement
  vector.
• The objective of matrix S is to satisfy the
  maximum of requirement, so the objective function
  can be

38
Linear Programming Approach

• In summary
• Problem (P)
• subject to Sij Sij 1, (i, j), (i, j)
  I
• j 1, 2, , N
• Sij, Sij 0, 1, i, j, i, j

39
Dynamic Channel Allocation

• Dynamic channel allocation assumes all the cells
  (base stations) in a network share all the
  available channels. Each channel is assigned to
  support user traffic dynamically on demand (call)
  based on the current channel usage and
  interference level in the network.
• Channel Segregation
• - Each cell can access any channel by tuning a
  carrier frequency and selecting a time slot.
• - The centralized channel segregation system
  maintain tables for each cell where a priority
  function is stored for every channel.

40
Dynamic Channel Allocation
41
Dynamic Channel Allocation

• Priority P P Ns/Nt
• Ns the number of successful use of the channel
• Nt total number of trails for the channel
• The priority will be different for every channel
  in every cell
• If a channel is successfully used
• Ns ? Ns 1 and Nt ? Nt 1
• Otherwise
• Nt ? Nt 1
• In general Dynamic Channel Allocation
• No channel is fixed to any specific cell
• No channel reuse planning is required
• DCA is adaptive to changes in radio channel
  environment such as the appearance of new
  buildings and new cell sites

42
Adaptive Channel Allocation

• Similar to DCA, but distributed channel
  allocation
• Each cell scans channels and acquire good
  channels independently, interference adaptive
• Distributed algorithm, no load to central
  management
• Channel competition and conflict
• Requires fast tuning radios

43
Issues of FCA/DCA/ACA

• FCA
• Can not adjust to the traffic change
• Low channel efficiency a cell must be
  provisioned with enough channels for peak traffic
• DAC
• Can process real time channel allocation
• Heavy load to mobile switch center (MSC)
• ACA
• Can process real time channel allocation
• Channel competition and conflict
• Need costly radios (fast tuning radios)