Mobility Models in Ad Hoc Networks

1
Mobility Models in Ad Hoc Networks
Abstract

• Mobility management in ad hoc wireless networks
  faces many challenges. Mobility constantly causes
  the network topology to change. In order to keep
  accurate routes, the routing protocols must
  dynamically readjust to such changes. Thus,
  routing update traffic overhead is significantly
  high. Different mobility patterns have in general
  different impact on a specific network protocol
  or application. Consequently, the network
  performance will be strongly influenced by the
  nature of the mobility pattern. In the past,
  mobility models were rather casually used to
  evaluate network performance under different
  routing protocols. In this seminar I have
  discussed some of the common Mobility Models and
  few uncommon ones. Their impact on various
  network parameters and routing parameters have
  been discussed

2
Mobility Models in Ad Hoc Networks

• Deepanshu Shukla
• deepanshu_at_it.itb.ac.in
• KReSIT, IIT Bombay.
• Seminar Guide- Prof. Sridhar Iyer

3
Outline

• Introduction to Ad Hoc Networks
• Routing in Ad Hoc Networks
• DSR Algorithm
• AODV Algorithm
• Mobility Models
• Brownian Motion Model
• Random Walk Model
• Random Waypoint Mobility Model
• Reference Point group Mobility Model
• Mobility Vector Model
• Other less discussed models
• Summary and Conclusions

4
Mobile AD hoc Networks(MANET)
5
Introduction-MANET

• No pre-existing communication infrastructure
• Autonomous system of mobile routers (and
  associated hosts) connected by wireless links
• Routers are free to move randomly and organize
  themselves arbitrarily
• Wireless topology may change rapidly and
  unpredictably

6
Why Ad Hoc Networks ?

• Setting up of fixed access points and backbone
  infrastructure is not always viable
• Infrastructure may not be present in a disaster
  area or war zone
• Infrastructure may not be practical for
  short-range radios Bluetooth (range 10m)
• Ad hoc networks characteristics
• Do not need backbone infrastructure support, Are
  easy to deploy
• Team collaboration of large number of mobile
  units
• Limited Bandwidth
• Low latency access to distributed resources
• Useful when infrastructure is absent, destroyed
  or impractical to construct

7
Routing in MANET
8
Routing Protocols

• Proactive protocols
• Traditional distributed shortest-path protocols
• Maintain routes between every host pair at all
  times
• Based on periodic updates High routing overhead
• Example DSDV (destination sequenced distance
  vector)
• Reactive protocols
• Determine route if and when needed
• Source initiates route discovery
• Example DSR (dynamic source routing)
• Hybrid protocols
• Adaptive Combination of proactive and reactive
• Example ZRP (zone routing protocol)

9
Dynamic Source Routing (DSR)
10
Dynamic Source Routing (DSR) Johnson96

• When node S wants to send a packet to node D, but
  does not know a route to D, node S initiates a
  route discovery
• Intermediate nodes use the source route included
  in a packet to determine to whom a packet should
  be forwarded

11
Dynamic Source Routing (DSR) Johnson96

• Advantages
• Routes maintained only between nodes who need to
  communicate
• Route caching can further reduce route discovery
  overhead
• A single route discovery may yield many routes to
  the destination, due to intermediate nodes
  replying from local caches
• Disadvantages
• Packet header size grows with route length due to
  source routing
• Flood of route requests may potentially reach all
  nodes in the network
• Potential collisions between route requests
  propagated by neighboring nodes
• Increased contention if too many route replies
  come back due to nodes replying using their local
  cache
• Stale caches will lead to increased overhead

12
Ad Hoc On-Demand Distance Vector Routing (AODV)
13
AODV (Ad Hoc On-Demand Distance Vector
Routing)Perkins99Wmcsa

• When a node re-broadcasts a Route Request, it
  sets up a reverse path pointing towards the
  source
• AODV assumes symmetric (bi-directional) links
• AODV attempts to improve on DSR by maintaining
  routing tables at the nodes, so that data packets
  do not have to contain routes
• Route Reply travels along the reverse path set-up
  when Route Request is forwarded

14
Mobility in Ad Hoc Networks
15
Mobility

• Realistic models for motion simulation are needed
  
• Used to derive traffic and mobility prediction
  models in the study of various problems in
  network such as
• Traffic load Traffic control overhead
• Location Management
• Unlike in cellular there is no concept of
  cells
• Researchers validate their algorithms against
  these models
• invalid conclusions may be drawn from overly
  simplistic or unrealistic models
• it is difficult to compare performance results of
  algorithms due to the variety of models used

16
Mobility

• Impact of Mobility on
• Network connectivity
• Routing Protocols
• DSR
• AODV
• HSR (Hierarchical Routing Protocol for Group
  Mobility )
• FSR (Fisheye State Routing)
• Out of scope of this seminar

17
Mobility Models
18
Brownian Motion modelEinstein in 1926

• It is totally random motion pattern
• Not a very realistic model
• Each node moves a certain amount of space after a
  random period.
• Movement is completely isolated.

19
Pursue ModelSanchez

• Nodes chase after a single target that may or may
  not be moving.
• Tracking is usually done with some error and
  randomness
• Nodes are not allowed to change the velocities on
  the fly
• http//www.disca.upv.es/misan/manet/MobApp2.html

20
Column ModelSanchez

• Represents a searching activity
• Nodes are distributed initially more or less like
  a row
• The whole row moves in some direction
• Each node can get close to some other and also
  can abandon the perfect line formation. The whole
  thing can also be moving some direction

21
Random Walk ModelZonoozi Dassanayake

• Memory less movement
• Randomly selected speed v min , v max and
  direction 0 to 2?
• This Model is extended to various specialized
  models as
• Random Way Point Model Jhonson
• Random Gauss-Markov Model Haas
• Random Mobility Model as Total random
• Constant velocity model as zero randomness
• Markovian Model Chiang

22
Random Waypoint ModelJohnson

• Breaks the movement of MH into pause and motion
  periods
• MH selects a random destination on the simulation
  space and moves to that destination at a speed
  uniformly distributed between an upper and lower
  bound.
• Upon reaching the destination, the node pauses
  again and repeats the process for the duration of
  the simulation.

23
Random Gauss Markov ModelHass

• Incremental Model
• Speed and direction of MH randomly diverge from
  the previous speed and direction after each time
  increment
• v (t? t) minmax (v(t) ? v , 0), Vmax
• ?(t ?t) ? (t) ?(?)
• Where ?v and ?? are uniformly picked from
  reasonable data range of -Amax ?t , Amax ?t
  and -a?t , a?t
• Amax is unit acceleration
• a?t is maximum unit angular change

24
Mobility Vector ModelX. Hong, T. J. Kwon, M.
Gerla, G. Pei, D. L. Gu

• In real world the network is heterogeneous in
  nature
• Different type of nodes will have different types
  of motion behavior
• Used to avoid some unrealistic behavior
• Sudden stops, sharp turns, turn backs etc. which
  are impossible in real world
• Natural motions by remembering mobility state
  and partial changes in its current mobility state
• Advantages are
• Simplification of position updates
• Ease of implementation
• Opportunity for mobility prediction

25
Mobility Vector ModelX. Hong, T. J. Kwon, M.
Gerla, G. Pei, D. L. Gu

• Mobility is expressed as a vector (xv, yv)
• Scalar value (norm) of Vector is the speed
• Mobility Vector is sum of Base Vector B? and
  Deviation Vector V?.
• M?(xm, ym) or (rm, ?m)
• B ?(bxv,byv) or (rb, ?b)
• V?(vxv,vyv) or (rv, ?v)
• Model shows that M?B? ? V? ? is
  acceleration factor
• By properly adjusting ? we generate a smoother
  trajectory, eliminating unrealistic MH motion

26
Mobility Vector ModelX. Hong, T. J. Kwon, M.
Gerla, G. Pei, D. L. Gu

• Mobility Vector as Framework
• Gravity Model
• Receivers tend to move towards signal source
• Every MH node is assigned a charge (ve ve or
  none) Base stn is ve
• Mobility Vector is function of distance and
  charges
• Location Dependant Model
• Collective mobility pattern in specific area
• MV has common component which represent the
  direction and speed
• Targeting Model
• Nodes move toward a common target
• Given a target co ordinate it is easy to
  calculate a base vector
• Group Motion Model
• Teams which tend to co ordinate their movements
• Different Group Patterns can be represented using
  a Base Vector and different Deviation Vector

27
Reference Point Group Mobility ModelX. Hong, M.
Gerla, G. Pei, C. C. Chiang

• Collaboration among members of the same team is
  common in Manet.
• Partition the network into several groups each
  with its own mobility behavior
• One of the first examples of group mobility are
• Exponential Correlated Random (ECR) model
• Model reproduces all possible movements including
  individual and group by adjusting the parameters
  of motion function.
• ? adjusts the rate of change from old to new (
  small ? causes large change) ? is a random
  Gaussian variable with variance ?
• ? ? vary from group to group. ECR requires to
  have sets of (?,? ) for all MHs

28
Reference Point Group Mobility ModelX. Hong, M.
Gerla, G. Pei, C. C. Chiang

• In RPGM group trajectory is determined by
  providing path for the center
• Defines the motion of groups explicitly by giving
  a motion path for the Group
• Path is given by defining a sequence of check
  points along the path corresponding to given time
  interval
• Moves from one check point to another.
  Re-computes Motion Vector
• Advantage of providing a general and flexible
  framework for describing mobility patterns which
  are task oriented and time restricted

29
Reference Point Group Mobility ModelX. Hong, M.
Gerla, G. Pei, C. C. Chiang

• Applications of RPGM Model
• In-Place Mobility Model
• Overlap Mobility Model
• Conventional Mobility Model

30
Other Mobility Models

• Flies on a Cake Sanchez Nodes are modeled as
  the flies flying around a cake while the cake is
  moving (depending on the acceleration of this
  movement you'll get different flies density a
  fast move will increment the separation between
  every node, but as the acceleration decreases the
  cloud of flies will be concentrated in a smaller
  volume (higher flies density).
• Nomadic Community Sanchez Similar to the flies
  on cake, but minimum and maximum separation
  between nodes is bounded, and the whole group of
  nodes movement is done in stages after which the
  nodes spend an amount of time moving like the
  Brownian model but only inside its bounded circle
• Smooth Random Mobility Model BettstetterUses
  stochastic principles for direction and speed
  control in which the new values for speed
  direction are correlated to previous values. This
  feature makes movement of nodes more smooth than
  random movement. Speed control is based on
  target speeds changing according to Poisson
  process.

31
Mobility Parameters

• Average Speed Distance Traveled
• Transmission Range Link Changes
• Network Performance

32
Mobility Parameters

• Average Speed and Distance Traveled
• Average Speed is actual distance traveled oer
  simulation time
• Traveled distance is large but Geographical
  displacement is small
• Extra distance is to be traveled to achieve a
  certain geographical displacement

33
Mobility Parameters

• Transmission Range Link Changes
• Effect of node distribution density and
  transmission range of nodes
• Choice of transmission range is related to
  mobility
• We monitor the change of Link State status
  (Up/Down)
• Comparison chart on next slide.
• Rate of change is indicative of topology change
• Choosing 1.5 2 time of mean distance is good
  solution in free space channel environment

34
RW has higher rate at high mobility when trans
range is small. When transmission range eq 2 mean
dist b/w nodes-high change rate(35). Incr to 1.5
timereduces to half of 35increases to 2
timesrate decreases to 1/3rd (12)
35
Mobility Parameters

• Network Performance
• No matter what model is used, increase of
  transmission range increases the delivery ratio
• At high mobility, increased density will increase
  the chance of finding new routes. Also lead to
  more collisions.
• MV and Random Waypoint show improvement whereas
  in RPGM Random Walk throughput drops.
• Increase in transmission range has different
  effects on different routing protocols
• FSR has large degradation from 200-400m
• Transmission range from 1.5 2 times the mean
  distance will produce uniformly the best
  improvements in Delivery Ratio

36
Mobility Parameters
37
Mobility Parameters
38
Mobility Parameters
39
Mobility ParametersExperimental Configuration

• Protocols Used are
• Dynamic Source Routing (DSR), Ad hoc On Demand
  Distance Vector Routing (AODV), Fisheye State
  routing (FSR)
• Uses discrete-event simulation language PARSEC.
• Packet Delivery Ratio is used as performance
  metric
• Simulation area is 1km x 1km with 100 nodes
  uniformly distributed
• 50 Constant Bit Rate(CBR) used.
• 512 bytes packets
• Channel capacity is 2Mbps.

40
References

• http//www.cs.tamu.edu/people/youngbae/
  publications.html
• http//www-2.cs.cmu.edu/desney/15824/
  proposal_1st_draft.htm.
• http//www.ee.surrey.ac.uk/Personal/
  G.Aggelou/Manet20Publications.html
• http//147.46.59.102/imhyo/papers/ papers.html
• http//citeseer.nj.nec.com/hong99group.htm
• The Routing concepts in MANET was primarily taken
  from Prof Sridhars slides on Ad Hoc networks at
  http//www.it.iitb. ac.in/ sri/ talks/manet.ppt

• M. M. Zonoozi and P. Dassanayake. User mobility
  modeling and characterization of mobility
  patterns. IEEE Journal on Selected Areas in
  Communications, 15(7)1239- 1252, September 1997.
• X. Hong, M. Gerla, G. Pei, and C.-C. Chiang. A
  Group Mobility Model for Ad Hoc Wireless
  Networks. In Proceedings of ACM/IEEE MSWiM'99,
  Seattle, WA, Aug. 1999, pp.53-60
• M. Sanchez. Mobility models. http//www.disca.upv.
  es/misan/ mobmodel.htm, 1998.
• C. Bettstetter, Random Mobility Model for
  simulation of Wireless Networks. 4th ACM
  International Workshop on Modeling, Analysis and
  Simulation, Rome, Italy.

41
Thanks . . .