Location aided protocols for wireless communication A survey by broverya76


									     Location aided protocols for wireless communication: A survey

                                        Dat Ta
                                       CS 790G
                                       Fall 2005
                                      Term Paper

Abstract:                                    Systems. In part 2 of the paper, we will
                                             talk about proposed protocols that use
   Wireless communication has become         location-aided information in its routing
a part of our everyday lives during the      schemes. First, we will discuss two
past century and it actually picked up       routing protocols in mobile ad-hoc
some speed in the last few decades.          network that are the original LAR [3]
There are several types of wireless          (Location Aided Routing) and SPAAR
networks including cellular networks,        [5] (Secure Position Aided Ad-hoc
mobile ad-hoc networks, and sensor           Routing). We will focus more in the
networks. In wireless communication, a       LAR schemes because other protocols
node has to figure out how to talk to        are developed based on these schemes or
other nodes in the network using routing     use the design idea from these schemes.
protocols. There are various ways to         Then we will see how a hybrid network,
route packets in a network and it can        LLR       (Link-Connectivity-Prediction-
differ from different types of networks.     Based Location-Aided Routing Protocol)
And in this paper, we will discuss about     in this case, take advantage of the
how to use location-aided information to     information that a node knows about
help improve routing protocols in            locations of other nodes in the network.
wireless communication in general and        After that, we will take a look at a
in mobile ad-hoc networks, hybrid            routing protocol for sensor network
wired-wireless networks, and sensor          SELAR [4] (Scalable Energy-Efficient
networks in particular.                      Location Aided Routing Protocol for
                                             Wireless Sensor Networks).
1. Introduction:
                                             2. Proposed Protocols:
   There are different techniques used in
routing protocols. And there are ways to     2.1. Mobile Ad-hoc Routing Protocols:
help improving routing protocols. In
this paper, we will discuss particularly     Current routing protocols are categorized
about using location-aided information       into three categories according to their
to help routing protocols in mobile ad-      characteristics [9]:
hoc networks, hybrid wired-wireless            a) Table-driven routing protocols: In
networks, and sensor networks. There is           Table-driven routing protocols each
one assumption to be made in location-            node maintains one or more tables
aided protocol and that is each node in a         containing routing information to
network knows its location by using               every other node in the network.
information from the Global Positioning           All nodes update these tables so as
    to maintain a consistent and up-to-     receiving the message from the sender
    date view of the network. When the      will forward the message to its neighbors
    network topology changes the            other than the sender. Node X receives
    nodes propagate update messages         the route request message, it will check
    throughout the network in order to      to see if it is the desired destination. If
    maintain consistent and up-to-date      node X is not the desired destination, it
    routing information about the           will add itself to the routing path and
    whole network. These routing            forward the message to its neighbor. In
    protocols differ in the method by       order to stop infinite message
    which the topology change               forwarding, a node will not forward a
    information is distributed across the   message if it already did by using
    network and the number of               sequence number. In this case, node X
    necessary routing-related tables.       will only receive the message from
    The following sections discuss          either node C or B.
    some of the existing table-driven ad
    hoc routing protocols.

 b) On-demand routing protocols: In
    contrast to table-driven routing
    protocols, a route from a source
    node to a destination node is
    initiated only when the source node
    wants to send to the destination.       Figure 1: Illustration of flooding
    The route will be used during the
    communication period.                           The route request message will
                                            get to the destination node D from node
                                            X. Node D checks and it realizes that it
 c) Hybrid routing protocols: For           is the intended destination. It will then
    routing within a certain distance       send a route reply message to the sender
    from the node concerned, a table        using the path that is included in the
    driven approach is used. For nodes      message. So the route reply message,
    located beyond this, an on-demand       which is included with the path that used
    approach is used.                       to get to the destination, will reverse the
                                            path used to get to the destination to get
                                            to the sender. The sender will wait for a
Using flooding                              time interval since it sends out the route
        Flooding is one of the easiest      request message. In the case that the
ways to send packets in wireless            route request message cannot get to the
communication. Assuming node S,             destination node or the route reply
referred as the sender, wants to send a     message cannot get to the sender node,
data packet to node D, referred as the      the sender will re-send the route request
destination. At this moment, the sender     message.
node does not know a path to the
destination node. Node S sends out a
route request message to all of its
neighbor nodes. A neighbor node upon
2.1.1 Location-Aided            Routing
      (LAR):                                L is the current location of node D. Then
                                            the circle around node D denotes the
        Using location-aided routing will   possible location of node D at time t1, as
make good use of location information       in Figure 2(a). If node S knows that node
to help reduce routing overhead.            D is moving north, then the expected
Location-aided protocols use the            zone can be reduced to a semi-circle.
information provided by the Global
Positioning System (GPS). There are         Request Zone
assumptions in location-aided protocols.    Before sending a route request like in the
One of them is that the calculated          normal flooding algorithm, node S
position gather from GPS is correct         defines a request zone. If a node is not in
which means each host knows its             this request zone, it will not forward a
location precisely.                         route request message to its neighbor. In
                                            order to increase the probability that the
    Expected     Zone    and    Request     route request will reach the destination
Zone:                                       node D, the request zone should include
                                            the expected zone. This request zone
Expected Zone                               will include regions that are around the
         Assume that node S wants to find   expected zone. There are two reasons for
a route to node D and S knows the exact     this:
location of D at time t0. At the time t1,   a) If node S is out of the expected zone
node S will assume a region that will           of node D, then a path from node S
contain node D. This region is called the       to node D must include nodes that
“expected zone”. Node S uses node D’s           are not in the expected zone. So the
average speed to determine this zone.           request zone should be bigger than
Initially, when node S does not have            the expected zone.
node D’s location and speed, it will        b) As in Figure 3(a), the request zone
assume the entire region of the ad-hoc          includes the expected zone. But what
network is the expected zone. So node S         happens if all the paths from S to D
basically floods the network to get to          have intermediate nodes that are not
node D (link 1). This will change once          in the request zone like in Figure
node S knows about node D’s                     3(b)? It is not guaranteed that a path
whereabouts.                                    will be found if the request zone is
                                                the one in Figure 3(a). When node S
                                                does not receive a route reply during
                                                a time period, it will re-send a route
                                                request message but with a larger
                                                request zone. This new request zone
                                                could be the entire network. Though
                                                there is a trade-off that there will be
                                                a lot more overhead as the request
                                                zone gets bigger.

Figure 2: Examples of expected zone
                                               knows that it is not in the request zone,
                                               though it is in the neighborhood of node
                                               S. Node D will reply with a route reply
                                               message when it receives the route
                                               request message from node S. Node D
                                               will include its location and maybe its
                                               current average speed. When node S
                                               receives this route reply message, it will
Figure 3: Request zone                         update the location of node D and
                                               everything else if necessary. This
                                               updated information will be used for the
Determining Membership of Request              next route request message.
        We need to have good schemes
to help determining membership of
request zone. This will help a node to
determine whether it should forward the
route request message or not and this
affects the network traffic. There are
currently two schemes to use to
determine whether a node is in the
request zone.

LAR Scheme 1
         This scheme uses a rectangular
request zone. Assume that the location
of node D at t0 is at (Xd, Yd). Node S
knows about node D’s location and its
average speed. Thus, node S can create
the expected zone of node D at time t1
which is a little circle in Figure 4(a). In
this first scheme, the request zone is the
smallest rectangle that has the current
location of node S and the expected zone
of node D such that the sides of the
rectangle are parallel to the X and Y
axes. As in Figure 4(a), this rectangular
has four corners that are S, A, B and C
and A, B, C and G for Figure 4(b).
The scheme can be done by having node
S including its location in the route
request message. A node upon receiving
this message will discard it if it is not in
the rectangle request zone. Thus, node I
will forward the route request message
while node J will discard it because it        LAR Scheme 2
                                            flooding algorithm and LAR scheme 1
In this second scheme of LAR, node S,       and 2 vs. the increasing speed.
which has the location of (Xs, Ys) at t0
knows the location (Xd, Yd) of node D
at time t0. When S sends a route request
message, it includes its distance to node
D with the location of node D at t0. If a
node that is in the request zone will
check its distance to node D and
compare it with the distance that node S
included. It will forward the route
request message if this distance is
smaller than the computed one from
node S.

                                            Figure 7
                                            Figure 7 shows the number of Routing
                                            packet per Data packet vs. the number of
Figure 5                                    nodes in the network at two different
                                            average speeds.

                                            It   shows      that     LAR      schemes
Performance of LAR schemes over             outperformed the flooding algorithm. As
flooding:                                   speed of nodes changes, LAR can help
                                            make better prediction of where each
                                            node is in order to help routing. Also, As
                                            the number of node increases, LAR
                                            schemes give better consistency in the
                                            number of Routing packet per Data

                                            2.1.2   SPAAR (Secure Position Aided
                                                    Ad-hoc Routing):
Figure 6
                                                  One primary application of
Figure 6 shows the number of Routing
                                            mobile ad-hoc networks is in military
packets per Data packets between the
use including tactical operations. In             distance to D, D’s coordinates,
these environments, security is often the         and TUSN (a time stamped
primary concern. Secure routing                   sequence number that is
protocols protect routing messages                incremented each time N
against malicious nodes and attacks that          broadcasts a table update
one     could     expect     in    hostile        message).
environments. Though a lot of routing            Step 2: RREQ recipients decrypt
protocols are vulnerable and lack of              it with the appropriate group
security features even the very standard          decryption key.
ones like DSR and AODV. These                    Step 3: An intermediate node
protocols were not designed to handle             checks to see if it, or any of its
those kinds of tasks. That is why there is        neighbors, is closer to the
SPAAR.                                            destination node D. The RREQ
                                                  is dropped if neither the
Secure      Position     Aided       Ad-hoc       intermediate node nor its
Routing:                                          neighbors are closer to the
        SPAAR          uses         position      destination. If either is closer,
information to improve performance and            the node forwards the RREQ
security     while    keeping       position      with its identifier and distance to
information protected from unauthorized           S.
nodes. Each node only allows messages          o Route Reply:
from one-hop neighbors. This can be              Step 1: When the destination
done using the position information that          node receives a RREQ message,
each node has. Also, the source node              it will reply with a RREP
must know the approximate geographic              message that includes the RREQ
location of the destination.                      sequence         number,         its
- Setup: Each node requires a                     coordinates, its velocity, and a
public/private key pair, a certificate            TUSN. It then signs the RREP
binding its identity to its public key            with its private key and encrypts
(signed by a trusted certificate server),         it with the public key of the
and the public key of the trusted                 neighbor it received the RREQ
certificate server.                               from. The RREP propagates
- The Neighbor Table: Each node                   along the reverse path of the
maintain a neighbor table that contains           RREQ.
the identity and position information of         Step 2: When an intermediate
each verified neighbor, along with the            node receive a RREP message,
cryptographic keys required for secure            it decrypts it with its private key
communication with each neighbor. A               and verifies the signature with
node will only accept routing messages            the public key of the neighbor
from a node in its routing table.                 node it received it from. Then it
- Route Discovery:                                setup forward entries in its route
    o Route Request:                              table that points to the node
        Step 1: Node N broadcasts an              from which the RREP came.
         encrypted RREQ with the                  Then it signs and encrypts the
         RREQ sequence number, the                RREP before passing it to the
         destination     identifier,    N’s       next node in the reverse route.
        Step 3: Once the RREP gets to        processing abilities to locally carry out
         the source node, the source node    simple computations and transmit only
         will decrypt and verify the         the required and partially processed data
         signature. Then it updates its      (Survey Sensor networks).
         destination table with the new
         destination position information    The SELAR protocol:
         which includes its location, its       SELAR combines energy and
         velocity vector and a TUSN.         location information to make routing
      o Route Error messages: Nodes in       decisions locally while achieving the
        the network keep track of active     global goals. The sink node will flood its
        routes in the route table. If a      location information to all its
        route remains unused for a           neighboring nodes. Then all the sensor
        certain timeout period, it is de-    nodes flood their location and energy
        activated. Nodes will remove         information to their neighboring nodes.
        routes that contain a neighbor       Because nodes in a wireless sensor
        node that has been removed. A        networks rarely move, only energy
        route error message is sent to the   information needs to be updated. When a
        source node if data is received      node needs to forward a packet, it
        for a de-activated route. The        creates a forwarding zone, which is the
        source node will then try to look    area formed by the angle α in the
        for another route.                   direction of the sink node and the area
                                             coverage of the sending node. It only
                                             forwards the packet to its neighbors that
2.2      SELAR (Scalable Energy-             are in this forwarding zone. The node
         Efficient  Location   Aided         starts with the angle α which is set to be
         Routing Protocol for Wireless       15o initially. This node will only forward
         Sensor Networks)                    the packet until it has transmitted to a
                                             certain number of its neighbor nodes or
Overview of wireless sensor networks:        the alpha angle reaches 90o. If the node
   A wireless sensor network is a            cannot find a neighbor node with in the
network that consists of a very large        maximum 90o, it will use gossiping to
number of tiny nodes with very               transmit the packet.
constrained capabilities in terms of
energy,      computing      power,     and
communication        possibilities.   The
position of sensor nodes need not be
engineered or predetermined. This
allows      random     deployment       in
inaccessible terrains or disaster relief.
On the other hand, this is also means that
sensor network protocols and algorithms
                                             2.3    LLR
must         possess       self-organizing
capabilities. Sensor nodes are filtered
                                             Overview of wired-wireless networks:
with an onboard processor. So instead of
                                                A hybrid wired-wireless network is
sending the raw data to the nodes
                                             defined to be a two layer hierarchical
responsible for the fusion, they use their
                                             network that contains both mobile hosts
(MHs) and access point (APs). MHs, or         - Route Discovery: Whenever a
mobile nodes (MNs) can communicate           source MN has data packets to send, it
with other MNs, which can be multi-          first checks its routing table to
hops away. APs, or gateways (GWs), are       determine whether it has a current route
nodes with both wireless and wired           to that destination MN. If none exists, it
interface. GWs give MNs access to other      initiates the route discovery process
MNs or fixed hosts (FHs) of wired            similar to that of AODV. Unlike
network.                                     AODV, the RREQ is broadcasted only
    Routing in hybrid networks is quite a    to nodes in the region within a few hops
challenging task because of the existence    away from the source MN instead of the
of MNs in the networks. A simple way         whole network. This region would
to route in this kind of network is to use   include the current registered GW of the
GWs as the default route. All the            source MN. This can be done by
communications between MNs must go           specifying the TTL of the RREQ to be
through the GWs.                             the number of hops away from the
                                             current registered GW of the source
                                             MN. This means the RREQ can
                                             propagate at most K hops away, since
                                             MN is inside the K-hop subnet of the
                                             GW. The reverse route is setup by
                                             RREQ, same as AODV.
                                                   A RREP can be generated by the
                                             destination MN, or intermediate
                                             neighbors with up-to-date route to the
                                             destination. The WR path from source
                                             to destination is then setup as RREP
                                             travels back to the source. Upon
The LLR protocol:                            receiving RREP, source MN starts
                                             sending data packets along the WR
The LLR protocol consists of three           path. If the source MN receives no
separate phases: Wireless Routing path       RREP, the WWR path is used. WWR
(WR) route discovery phase, Route            path is always available since each MN
maintenance phase, and route soft-           establishes and maintains a route
handoff phase. A Wireless Routing path       towards it current registered GW during
is a wireless multi-hop path directly        the gateway discovery process. When
from source to destination. When a           source MN receives no RREP, it sends
source MN wants to find the local            data packets towards its current
routing path, it initiates a local route     registered GW and sets a flag for that
discovery, which is called WR route          destination MN in the routing table to
discovery process. If both the MN node       indicate it is using WWR path. Each
and the destination node are in the same     data packet is embedded with the ID of
subnet, a path will be found. If no WR       the destination. After the data packets
path is found, the source MN uses the        reach the GW, the GW checks its
WWR by forwarding the data packet            routing table for the next-hop node
towards its currently registered GW.         towards the destination MN and sends
                                             out the data packet accordingly. If both
paths exist, source MN always prefers        3. OPEN     ISSUES                   AND
WR over WWR.                                    CONCLUSIONS

  - Route Maintenance: There are two             The general problem of all of the
 possible paths that a node can use that     discussed protocols is the usage of the
 are the WR path and the WWR path.           GPS to determine locations [1, 2, 6, 8].
    o WWR Maintenance: LLR uses              Right now, there are many different GPS
   the gateway discovery algorithm to        systems out there and they do not work
   maintain WWR paths. This algorithm        well together. Thus the location
   provides frequently routes updates        information that a node gets using this
   between MNs and GWs.                      system will be different if it uses another
    o WR Maintenance: intermediate           system.
   nodes update the RET (Route                   Also, most tests were done by
   Expiration Time) in each data packet      comparing the new protocols with the
   based on the Link Expiration Time         flooding algorithm which is a very
   (LET). The “critical time” period is      limited and simple algorithm, as in
   when a destination MN determines a        Figure 6 and 7. There should be done
   route is about to expire. The node will   more test cases between the new
   compute both the expected zone and        schemes and some more advanced
   the request zone like in LAR. It then     schemes other than the flooding
   attaches the information to a specific    algorithm.
   RREQ                                          As in mobile ad-hoc networks, there
(SRREQ) and broadcasts it. Only the          are still a lot of challenges. Though the
source MN an reply to this SRREQ             security issues seem to be solved by the
which also contains the current RET.         SPAAR protocol but there are
Intermediate nodes first checks whether      drawbacks. Such complex protocol as
it is inside the request zone and only       SPAAR will produce a lot of overhead
nodes within the request zone can            because nodes have to check for
forward the SRREQ. An intermediate           signature verification and encryption key
node also checks the LET of last link        between every transmission. SPAAR
that SRREQ was received from and if          does provide very high security.
the LET is less than or equal to RET of          There are issues with the SELAR
SRREQ, the SSREQ is dropped instead          protocol as well. It has not been tested
forwarded. After the source receives the     very thoroughly. There is a problem that
SRREQ, it chooses the best route on          nodes which are closer to the sink node
which to reroute the data packets based      die faster because it has to handle more
on the information contained in the          traffic. One approach for solving this
SRREQ (e.g. number of hops,                  problem is to have a moving sink. That
destination sequence number, etc). After     is the sink node will move away from
that, source starts sending data packets     the previous location if nodes surround it
along the new route.                         have low power.
                                                 Those problems above should be
                                             solved not only to improve location-
                                             aided protocols but also help making
                                             wireless networks to be better and more
[1] “Iowa State University GPS page.”
Web site at
[2] “NAVSTAR GPS operations.” Web
site at
[3] Y. B. Ko and N.H. Vaidya,
“Location-Aided Routing in Mobile Ad
Hoc Networks”, In Procceedings of
ACM/IEEE Mobicom, Dallas Texas,
October 25-30, 1998.
[4] “Scalable Energy-Efficient Location
Aided Protocol for Wireless Sensor
[5] “Secure Position Aided Ad-hoc
[6] G. Dommety and R. Jain, “Potential
networking applications of global
positioning systems (GPS),” Tech. Rep.
TR-24, CS Dept., The Ohio State
University, April 1996.
[7] “Link-Connectivity-Prediction-Based
Location-Aided Routing Protocol”
[8] B. Parkinson and S. Gilbert,
“NAVSTAR: global positioning system
- ten years later,” in Proceeding of IEEE,
pp. 1177–1186, 1983.
[9] “A Review of Current Routing
Protocols for Ad Hoc Mobile Wireless

To top