Professional Documents
Culture Documents
FRESNEL ZONES
If unobstructed, radio waves will travel in a straight line from the
transmitter to the receiver. But if there are obstacles near the path, the
radio waves reflecting off those objects may arrive out of phase with
the signals that travel directly and reduce the power of the received
signal. On the other hand, the reflection can enhance the power of the
received signal if the reflection and the direct signals arrive in phase.
Sometimes this results in the counterintuitive finding that reducing the
height of an antenna increases the S+N/N ratio.
Fresnel provided a means to calculate where the zones are, where
obstacles will cause mostly in phase and mostly out of phase
reflections between the transmitter and the receiver. Obstacles in the
first Fresnel zone will create signals that will be 0 to 90 degrees out of
phase, in the second zone they will be 90 to 270 degrees out of phase,
in third zone, they will be 270 to 450 degrees out of phase and so on.
Odd numbered zones and even numbered zones are both constructive,
but their phase varies for Pi.
The general equation for calculating the Fresnel zone radius at any
point P in between the endpoints of the link is the following:
where,
The cross sectional radius of each Fresnel zone is the highest in the
center of the RF LoS, shrinking to a point at the antenna on each end.
For practical applications, it is often useful to know the maximum
radius of the first Fresnel zone. From the above formula, the
following formulas can be derived, using d1 = d2, D = d1 + d2, and
In SI:
where
r = radius in metres
Or in imperial units:
where
r = radius in feet
The area where the waves interfere with each other constructively is
our area of concern, called the “First Fresnel Zone.” The anomaly will
be seen throughout this region, and this has caused dry holes to be
drilled on anomalies that were off to the side of the seismic line.
Figure1 – Within a Fresnel Zone reflection contributions arrive coherently and thus reinforce.
Outside peaks and troughs tend to cancel each other and thus make little net contribution.
Figure 2 – The Pythagorean theorem allows one to calculate the radius of the Fresnel Zone.
Figure 3 – Nomogram for determining Fresnel Zone radius. A straight line connecting the
two-way time and frequency intersects the central line at the same point as a line connecting
the average velocity and the Fresnel Zone radius. For example, a 20-Hz reflection at 2.0
seconds and velocity of 3.0 km/s has a Fresnel Zone radius of 470 m.
Figure 4 – Three-dimensional migration collapses the Fresnel Zone to a small circle, but 2-D
migration collapses it in only one direction.
Figure 5 – A given point on a reflector affects a surface region by an area equal to the Fresnel
Zone. In migration the entire Fresnel Zone must therefore be summed over to obtain the
correct amplitude.
Figure 6 – Model demonstrating out-of-the-plane imaging and Fresnel Zone effects on data
over a hypothetical reef with the specified offsets. The false image in this example is clearly
seen 1,500 feet away (after Waldo Jackson and Fred Hilterman).
The reflected waves will interfere constructively where their travel paths differ by less than a
half wavelength (see Figure 1), and the portion of the reflecting surface involved in these
reflections is called the First Fresnel Zone.
Beyond this First Fresnel Zone region interference will be alternatively destructive and
constructive. Fresnel showed that the destructive contribution of some of these zones beyond
the First Fresnel Zone will be offset by the constructive contribution of other zones – and thus
the reaction of the reflector responsible for a reflection will be only that of the First Fresnel
Zone.
In other words, a reflection that we think of as coming back to the surface from a point is
actually being reflected from an area with the dimension of the First Fresnel Zone. The
adjective “first” is often dropped.
The dimensions of the Fresnel Zone can be calculated easily by simple geometry. This is
shown in Figure 2 for a plane reflector in the constant velocity case, allowing for two-way
travel time.
Note that the Fresnel Zone radius depends on wavelength (itself a function of frequency and
velocity). For seismic frequencies and the depths of interest to oil finders, the resulting
dimensions are quite large (Figure 3).
The effect of migration can be thought of as lowering geophones through the earth until they
are coincident with a reflector, at which time the Fresnel Zone will have shrunk to a small
circle. If the data and migration are two-dimensional, then the Fresnel Zone will have only
shrunk in one dimension and will still extend its full width perpendicular to the line (Figure
4).
Figure 2 can be turned upside down to show the portion of the surface affected by the
reflectivity at a print on the reflector (Figure 5). If we wish to preserve amplitudes so that we
can interpret amplitude variations as changes in acoustic impedance, porosity, hydrocarbon
accumulations, lithology, porosity-thickness, etc., we must integrate over all of the affected
surface in the migration process in order to get the correct relative value.
Thus, if we are to compute porosity-thickness correctly from a 3-D seismic survey, the survey
must extend for the full Fresnel-zone radius beyond the field.
Because of data coming out of the plane on the 2-D profiles shown in the model (Figure 6),
the algal mound is seen on all profiles that are within a 1,000-foot window. A profile only
800 feet away looks identical to one over the center of the feature. Hence, any survey must
extend beyond the area over which one intends to interpret amplitude changes by a fringe
distance required by the migration process.
FRESNEL ZONE PLATES
Diffraction optics - Diffraction gratings - Fresnel zone plates - Gabor Hologram - Order -
Contact us
Let us place the transparent plate with a round hole between the point source of light and the
screen of observation S. The intensity of light in the centre of the screen will depend upon the
number of Fresnel zones in the hole. Every Fresnel zone is a ring area of the hole. The light
from two adjacent zones comes to point of observation in anti-phase, so the corresponding
waves vanish each other because of destructive interference. On the contrary, waves from
only odd or only even zones will amplify each other because of constructive interference.
Number of the Fresnel zones in the hole of diameter D equals m = D 2(1/a+1/b)/4λ, where à and
b are the distances from the hole to the source and to the image respectively. The outer
radius of m-th Fresnel zone equals Rm = (abmλ/(a+b))1/2. If the source of light is situated in
infinity and the Fresnel zone plate is illuminated by parallel light beam then behind zone plate
the light will be focused in a point similar to lens. Moreover, length à from the plate to source
and distance b from the plate to the image are interrelated by the same equation as for lens:
1/a + 1/b = 1/F, where focal length F is defined as F=Rm2/mλ. If the center of the zone plate is
transparent, then m is odd and we have take for formula outer radius R of the transparent ring
of the plate. If the center of
the plate is black, then R is
the outer radius of the dark
(black) ring. With the aid of
Fresnel zone plate we can
even take an optical
images, but quality of such
an image will be rather low
(as compared to lens). in
contrast to lens Fresnel
zone plate has several focal
distances: Fn=F/(2n+1),
where n is integer.
Row 2 - plates with the number of Fresnel zones equal to N printed above the row. The even
zones from 1 to N are opened (transparent). So, in the left plate (N=1) all Fresnel zones are
closed, in the 2-nd left one (N=3) only 2-nd Fresnel zone is opened, in the 3-rd one (N=5) 2-nd
and 4-th zones are opened, etc. In the very right plate (N=19) all even Fresnel zones from 1 to
19 are opened. The intensity of the light in the focus of the plate equals approximately I = 4MI0,
where M = (N-1)/2 is number of the opened Fresnel zones.
Row 3 - plates with the number of Fresnel zones equal to N printed above the row. All the
zones from 1 to N are opened (transparent). So, in the left plate (N=1) only the first Fresnel
zone is opened, in the 2-nd left plate (N=3) 1-st, 2-nd and 3-rd Fresnel zones are opened, in the
3-rd one (N=5) all the zones from 1-st to 5-th are opened and etc. In the very right plate (N=19)
all the Fresnel zones from 1 to 19 are opened. The intensity of the light in the focus of the plate
equals approximately I = 4I0.
Row 4 - plates with the number of Fresnel zones equal to N printed above the row.
Intermediate plates correspond to intermediate numbers of the Fresnel zones N. All the zones
from 1 to N are opened (transparent). So, in the very left plate (N=1) only the 1-st Fresnel zone
is opened, in the 2-nd one first two zones are opened, in the 3-rd plate first three zones are
opened etc. In the very right plate (N=19) all the Fresnel zones from 1 to 19 are opened. The
intensity of the light in the focus of the plates with the odd N is the same and equals
approximately I = 4I0, while in the focus of the plates with the even N the intensity equals
approximately 0.
Row 5 - plates with the number of Fresnel zones equal to N printed above the row. "W" means
that all the odd Fresnel zones are transparent (center of the plate is transparent), while "B"
means that all the even Fresnel zones are transparent (center of the plate is black). The
intensity of the light in the focus of the plate equals approximately I ~ 4(N/2)I0=2NI0.
For all the plates in the rows 1-5 the focus distance F equals 10 cm for the wavelength 632,8
nm. Source of the light is considered to be in infinity.
Row 6 - Fresnel zone plates with the focal length equal in centimeters to F above the row.
Wavelength is 632,8 nm (if illuminated be the light with different wavelength, then the focal
length must be recalculated). All the odd Fresnel zones from 1 to 200 are opened (center of the
plate is transparent). Source of the light is considered to be in infinity.
Row 7 - Fresnel zone plates with the distance to image equal in centimeters to F above the
row. Source of the light is considered to be located symmetrically the plate at the same
distance a=F from it. Wavelength is 632,8 nm (if illuminated be the light with different
wavelength, then the focal length must be recalculated). All the odd Fresnel zones from 1 to
200 are opened (center of the plate is transparent).
3. Scheme of experiment.
The source of the
coherent optical radiation
is required for our
experiments. For example,
we can use
semiconductor laser
pointer with the
wavelength 650 nm or any
other laser. We used He-
Ne laser with the
wavelength 632,8 nm. If
we use the laser with
different wavelength, then
the focal distances must
be recalculated in
accordance with the new
wavelength. Diameter of
the laser beam equals
about 1 mm. It must be
extended to diameter > 1
cm. For such a purposes
two lenses are used:
short focus lens L1 with
focal distance F1 < 2mm
and long focus lens L2
with focal distance F2 ~ 8
cm. System assembled of
these lenses produce a
parallel and coherent light
beam of diameter 4-5 cm,
which is incident on
Fesnel zone plate FZP.
Changing the distance
between FZP and the
screen S we can observe
the focusing of light in the
moment when the
distance to screen equals
the focal length F of FZP.
Moving the screen towards the zone plate we can observe several points in which image is
focusing. If a photodetector with a small hole is situated in the place of screen, then intensity
of the radiation behind the zone plate can be measured.
Design Objective
The main objective in this research is to design an efficient data transmission and routing scheme
for ad hoc networks to maximize the transport layer throughput for a given system capacity. We
define the throughput metrics as follows. Suppose K transmission sessions are active between
source-destination pairs {S 1D1, S 2D2, ..., S KDK} and measured throughput at the destinations is
{T0, T1, ..., TK}, then the aggregate transport throughput of the network, TA, is defined as the sum
TA =
XK
i=1
Ti
If we measure the traffic transmitted or forwarded by node Nn for all K session as {T0,n, T1,n,...,
TK,n},
the aggregate network throughput, TN, can be expressed as the sum of the throughput Pn (Pn =
TN =
XN
n=1
Pn =
XN
n=1
XK
i=1
Ti,n
We define the system capacity of an ad hoc network as the maximum achievable aggregate
network throughput. Suppose each node in the network is transmitting at the maximum possible
rate and each node shares the medium equally with its neighbors, then the sum of the transmission
rate of all nodes gives an estimate of the upper bound of the system capacity.
In reality, it is not feasible to directly compute or measure the system capacity of an ad hoc network,
since the system capacity is a function of many factors in multiple protocol layers. Several factors
may affect the maximum transmit and receive rate of a node. From the physical layer perspective,
radio coverage range of a node, the use of a directional antenna pattern, and frequency channel
allocation schemes may contribute to the effective data rate supported by the wireless links. In
the wireless multiple access layer, the effective goodput may be influenced by traffic patterns from
higher layers. Here, goodput is defined to be the maximum effective data rate measured at the
17
receiver, excluding error packets and retransmissions. Some other limiting factors include the
choice of un-slotted or slotted transmission, use of a collision avoidance mechanism, and random
Researchers have defined several theoretical capacity bounds for ad hoc networks [24, 45], but
they are not applicable to our problem. In this research, we do not attempt to determine the system
capacity theoretically. Instead, we make some general assumptions concerning the capacity of the
underlying physical and medium access control (MAC) layer rather than rely on simulation results
of specific wireless physical and MAC protocols. Assuming each node participating in the network
has equal share of the wireless medium, a simple ideal capacity distribution among ad hoc nodes
is that the transmit capacity of each node is inversely proportional to its number of neighbors.
This implies that the MAC layer can effectively minimize collisions between neighbor nodes and
allocate the shared channel bandwidth fairly among the neighboring nodes. It is also assumed
that each node can receive all data transmitted by neighbor nodes. Under these assumptions, the
system capacity can be estimated as the sum of the maximum transmit rate of all nodes. These
assumptions are considered for the optimization problem in this research.
If we consider the system capacity of an ad hoc network as constant for a given topology, the
problem of maximizing the aggregate transport throughput includes two aspects. First, for a given
traffic pattern and maximum efficiency, we maximize the aggregate network throughput. In other
words, we should use the network capacity as much as possible. Second, we maximize the efficiency
of the ad hoc routing, defined as _ = TA/TN. If packet drops occur, the efficiency will be
less than one. Also, an increase in path lengths will reduce the efficiency. To increase the efficiency,
the routing schemes should try to avoid congestion and packet drops on forwarding nodes
and reduce path lengths. Link capacity and congestion condition should be considered in routing
The routing approach is the optimization variable in this problem. We assume that the physical
and MAC layer performance will not be affected by higher layers. A natural choice for the routing
approach for the problem is multipath routing. Multiple paths tend to use more available links
from the source to destination, thus maximizing the aggregate network throughput. Using multiple
paths may achieve load balancing among different paths and avoid congestion.
We use discrete event simulation to evaluate the performance of the proposed approach and
to compare performance to existing multipath routing protocols. The simulation modules include
abstract physical and MAC layer modules, routing modules for different routing approaches, IP
18
The traditional single shortest path (unipath) routing approach faces many challenges in wireless
ad hoc networks. Due to mobility, power control, and frequent link failures, routes in wireless ad
hoc networks are much more dynamic than in wired networks. To reduce control overhead while
maintaining up-to-date routing information, various techniques have been adopted. For example,
OLSR broadcasts only partial topology information and uses multipoint relay nodes to reduce
flooding of link state messages [11]. On-demand protocols such as AODV only try to find routes to
a destination when required [61], thus avoiding periodic exchange of topology information across
the whole network. Nevertheless, the costs of finding alternative routes after route disruption may
still be high and extra delay in packet delivery is introduced. In [18], experimental results from
wireless test beds have demonstrated that shortest single-path routing often chooses routes that
Unipath routing schemes sometimes do not respond well to network congestion. Packet queuing
and packet drops introduce extra costs in the transmission path and waste system capacity. For
multiple simultaneous connections sharing some links along their transmission paths, available
bandwidth to each connection is reduced and congestion may occur. Regardless of the actions
taken, e.g., queuing, random packet dropping, or packet prioritizing, the transmission costs will
increase.
Sending data along a single shortest path often results in creating congestions in some links
while keeping some other potential links idle. Shortest path routing may not be the best choice in
that case.
To illustrate the difference between unipath and multipath routing, we can make analogies
resemble interstate highways, access networks resemble in-state country roads, and ad hoc
networks resemble streets in some metropolitan cities. Usually highways have higher capacity and
utilization, while country roads have lower capacity and utilization. Now consider the traffic and
routes in highways and country roads. Since there are no or only a few traffic lights on highways,
there is usually no congestion leading to queuing delays. Thus, the variance of expected travel time
is small. Finding routes between two remote towns in many cases is trivial. One first identifies the
shortest or fastest interstate highways, and then selects the shortest in-state routes to the highway.
The traffic and routes in metropolitan cities are quite different. The capacities of roads are
usually of the same order of magnitude and all roads have high utilization, especially during rush
hours. Many intersections of roads (routers) have traffic lights (queues), and expected travel delays,
therefore, have high variance. Usually there is more than one route between two places with
approximately
equal distance. However, determining the optimum (fastest) routes from the topology
is non-trivial. Many drivers rely on real-time radio broadcast of traffic information to select routes.
19
For For the same source and destination, routes are chosen dynamically and tend to use all
possible
combinations.
The above comparison gives us some hints on how to use multiple routes in ad-hoc networks.
First, since there is no coordination between hosts in selecting routes, optimum routes cannot
be solely inferred from topology. The actual costs of routes will be a function of the choice of
routes of all connections. Information about the status and potential congestion of routes can
help
in deciding routes. Second, routes between end hosts should be selected dynamically and it is
desirable to utilize all available routes dynamically to spread out the traffic to improve
capacity
utilization. Third, queuing increases the variance in the estimated cost of routes and makes
the
network more dynamic, thus making it more difficult to predict good routes prior to
transmission.
Indefinite queuing delays also affect end-to-end performance.
The obstruction-intolerable space has ellipsoid shape and is thickest in the middle between
receiver and transmitter. The radius of the ellipsoid in the middle can be calculated according
to the formula r = 5.196 × sqrt(d / f) where d is distance between receiver and
transmitter in kilometers, f frequency in gigahertz and sqrt is square root. The resulting radius
is in meters.
Fresnel zone can be disrupted by various stationary objects - ground, terrain features, houses
and trees. Such disruption manifests as permanently weak or even absent reception. Moving
objects can disrupt Fresnel zone as well - buses, trams, trains, lorries, cars or even
pedestrians. These cause temporary losses of function.
Ronja has a Fresnel Zone as well, but it is very small. If we use the above formula for Ronja's
range 1.4km and frequency 476 THz (630 nm), we get 9mm. The beam width is 130mm.
After adding the 9mm on each side, we get 148mm thick sausage that we need to keep clear.
That's practically equivalent with a line of sight.
With such small Fresnel zone, Ronja is a solution for cases where Fresnel zone for
microwave would be obstructed - when the line of sight goes over roofs of houses, tips of tree
or above a road where the traffic can cause dropouts in the signal.
Below you can see 4 other installations where the beam is close to an obstacle. Click pictures
to zoom if you cannot see the red light on the other side.