Monthly Archives: Май 2014

Ground — and Foliage-Penetrating Radar

In the recent years, an attempt has been made to use radar to detect and map “targets” buried under the Earth’s surface or obscured by foliage. Primarily, interest arises from a number of potential applications, such as detecting and locating unexploded ordnance in a battlefield, manmade objects in landfills, buried hazardous waste, subsurface plumes of refined hydrocarbons, or military equipment hidden in forest vegetation. The radar is either spaceborne, airborne, or ground-towed and possibly operates in SAR mode. The radar needs to utilize ultra-wide-band (UWB) signals, containing both low — (for deeper penetration) and high — (for higher resolution) frequency components. This can be achieved in two ways: (1) by emission of pulses that are very short in duration (and, subsequently, ultra-wide in bandwidth) or (2) by sequential emission of narrow-band signals of carrier frequency increases in steps, covering a wide frequency band.

Even though the perspective of ground — or foliage-penetrating radar from initial tests has been encourag­ing, a number of difficulties have delayed the development of this technology. These include:

• High electromagnetic wave absorption, especially under moist soil conditions

• Random distributions of soil particles, such as rocks, that tend to scatter the electromagnetic energy, increase propagation losses, and reduce the image contrast

• High clay content to which water binds, and thus dipolar relaxation loss mechanisms are encouraged

• Roughness of the air-soil interface that tends to increase backscattering that interferes with the penetrating radar signature

Similar factors affect the development of foliage penetrating radar systems.

To date, the success of ground — and foliage-penetrating radar surveys seems to be absolutely site depen­dent. A thorough understanding of a site’s geology, hydrology, and topography is of paramount importance. Before undertaking a radar survey, it is necessary to obtain as much information as possible about the physical characteristics of the specific site. If boring log or monitor well data are available, they should be analyzed to determine soil stratigraphy and hydrology. If such data are not available, it is prudent to gather representative soil samples.

The applications for radar in subsurface target detection seem to fall into two broad categories, depending on the scale of the system, target, terrain structures, and search volumes. The case of large scales is made if the targets sought are large relatively to the average wavelength and the soil inhomogeneities. In this case, imaging would play a (secondary) role in reducing the number of false alarms of the detection procedure. If small targets, such as mines or weapons, are of interest, they would be hard to distinguish from clutter and the role of imaging would be enhanced. Thus, it is difficult or perhaps pointless to develop a single radar system for the detection of both large or deep and small or shallow targets. The wide-frequency-range requirement imposes stringent requirements in the range of both the electronics and the size of the relevant antennae and contributes to the delay of development of this significant radar application. However, ground — or foliage-penetrating radar technologies are presently an area of significant research investigation.

Coupling-Slot Characterization

The characteristics of these elements are determined by a method similar to that for the radiating elements. The for­ward and backward scattering components are the same as those for the radiating slot given in Eq. (14). Now they are applied in both the input and coupled guides to give the back — scattered, forward-scattered and coupled fields. The coupling coefficient between the waveguides is a function of the slot characteristics in the two waveguides and the characteristic impedances of the two guides. The power balance require­ment is applied to the junction to show that the coupling coef­ficient can be calculated from the ratio of the scattered fields and characteristic impedances in the two guides. The scat — tered-field terms are the same as those derived for the radiat­ing slot except for the value of the coefficient of the field in the slot. For the coupling slot, the ratio of the scattered fields in the two guides is independent of the value of the coefficient, so it need not be determined. The coupling ratio for the slot is given by (22)

where Y1 and Y2 are the characteristic impedances in the in­put and coupling guides, respectively, and B1 and B2 are the backscattered fields in the two guides. The value of 1/vM gives the turns ratio for the transformer in the lumped-cir — cuit model.

Coupling-Slot Geometries

The series-series angle slot is a convenient element for broadwall coupling between transverse waveguides because the coupling is controlled by adjusting the slot angle without changing the relative locations of the waveguides and a wide range of coupling values is attainable. This slot and its equiv­alent circuit model are shown in Fig. 6. The coupling value for this element at resonance, assuming a narrow rectangular slot with zero thickness, is given by

«АУ1ІД1І2

а2^2-^2І-®2І2

l2^2^S2 іфіХ gi

X

2

(24)

M1/2

H 1-

Bs

(b)

Figure 6. Series-series angle slot coupling element and its lumped — element circuit model.

where the expressions in the equation are the same as those defined in Eq. (20) for the angled radiating slot. The trans­verse series-series slot, which is the limiting case for the angled series-series slot, provides large coupling values be­tween the guides.

The offset shunt-series coupling slot provides coupling be­tween transverse waveguides. The slot is oriented as an offset shunt slot in one waveguide and a transverse series slot in the other. The coupling values are adjusted by changing the slot offset which is equivalent to moving the location of the shunt waveguide relative to the slot. The coupling value for this slot at resonance is given by

1 a2^2^g1^g2

~ 4 ЦЬг

1 — (l/a2)2 1 — (2l/Xgi )2

where waveguide 1 contains the shunt-slot orientation.

These approximate values for the slot coupling depend on the accuracy of the assumption made for the field in the slot and neglect effects, such as slot thickness and the reactive component of the slot, as the frequency departs form the reso­nant frequency. More accurate models for the slot are derived by the method of moments (23). Computer simulation tools based on the finite-element method (24) are also suitable for modeling coupling slots because they operate in a closed structure. This differs from the radiating slot case which does not lend itself well to this type of model because the problem is unbounded.

Queuing Models

To compute the queuing delay for a packet, we have to un­derstand the nature of the packet arrival process to a link, the kind of service time it needs (amount of transmission time), and the number of links we have from the source to the destination. In most queuing systems (35, 36), we assume that the arrival process is a Poisson process. We also assume that the holding time (the amount of time a request requires to service) follows an exponential distri­bution with parameter /г. The mean service time is then given by Vx. If two nodes i and j are connected by m links, then m packets can be transmitted from node i to node j at the same time. Generally m = 1 and therefore packets are transmitted one at a time. In case of circuit switching, it can be observed as one request being established at a time.

M/M/m Queue. A queuing system with m servers, Pois — son arrival process, and exponentially distributed service times is denoted by the M/M/m queuing system. The first letter M stands for memoryless. It can also be G for general distribution of interarrival times or D for deterministic in­terarrival times. The second letter stands for the type of probability distribution of the service times and can again be M, G, or D. The last number indicates the number of servers.

In a M/M/l queuing system, the average number of re­quests in the system in steady state is given by and

г — X

the average delay per request (waiting time plus service

time) is given by. Utilization of the system is denoted

г — X

by p — X/г, and the average time for a request in a system is given by average service time/(l — p). The average wait­ing time Tw is given by the difference of the average time in system and the average service time. This time is equal to 1/(г — X) — 1/г. The average number of requests in the queue is given by X * Tw. Also, the probability that exactly k requests are waiting is given by Pk = (1 — p)pk. These results for a queuing system will be used later on.

Performance Metrics. When a request for service arrives, the server (link) may be busy or free. If the server is free, the request is serviced. If the server is busy, then there are two possibilities: 1) The request is queued and serviced when the server becomes available. In this case, we are interested in finding out how long, on average, a request may have to wait before it is serviced. In other words, we need to find out how many requests are pending in a queue or the average length of the queue. This has implications in designing queues to store requests. 2) The incoming re­quest is denied service, which is called blocking. In this we are interested in determining the blocking probability for an incoming request. Again, this has implications in net­work design. We would like the blocking probability to be

Figure 17. A fiber divided among multiple wavelengths.

as small as possible.

EXAMPLE 1: DESIGN OF A NETWORK USING WDM FIBER OPTICS Wavelength Division Multiplexing-Based Optical Networking Technology

With the advent of optical transmission technology over optical fibers, the communication networks have attained orders of magnitude increase in the network capacity. The bandwidth available on a fiber is approximately 50 THz (terahertz). Hence, wavelength division multiplex­ing (WDM) was introduced that divided the available fiber bandwidth into multiple smaller bandwidth units called wavelengths. Figure 17 depicts the WDM view of a fiber link. Different connections, each between a single source/destination pair, can share the available bandwidth on a link using different wavelength channels. Advanced features such as optical channel routing and switching sup­ports flexible, scalable, and reliable transport of a wide va­riety of client signals at ultra-high speed.

Early optical networks employed broadcast and select technology. In such networks, each node that needs to transmit data broadcasts it using a single wavelength and the receiving node selects the information it wants to re­ceive by tuning its receiver to that wavelength. To avoid unnecessary transmission of signals to nodes that do not require them, wavelength routing mechanisms were devel­oped and deployed. The use of wavelength to route data is referred to as wavelength routing, and networks that em­ploy this technique are known as wavelength-routed net­works. In such networks, each connection between a pair of nodes is assigned a path and a unique wavelength through the network. A connection from one node to another node established on a particular wavelength is referred to as a lightpath. A wavelength-routed WDM network is shown in Fig. 18. The figure shows connections established between nodes A and C, B and C, H to G, B to F, and D and E. The con­nections from nodes A to C and B to F share a link. Hence, they have to use different wavelengths on the fiber.

One alternative to circuit switching, described above, is to use optical packet switching (OPS) or optical burst switching (OBS) (44-46)technology in the backbone. The major advantages of OPS/OBS are the flexible and efficient bandwidth usage, which enables the support of diverse ser­vices. However, implementation technologies are not yet there for successful deployment of them in an all-optical domain.

Multipliers

One difficulty in an SC multiplication technique is that con­tinuous programmability or multiplication of two signals is not available. A digitally programmable coefficient is realized with a capacitor bank, as shown in Fig. 5. The resolution of this technique is limited because the capacitor size increases by 2k where k is the number of programming bits.

Multipliers

Vin

Figure 3. Double-sampling S/H archi­tecture.

Multipliers

Multipliers

(a) (b)

Figure 4. (a) SC S/H single-ended. (b) Double-sampling S/H.

When continuous programmability is required, a continu­ous multiplier is used. Despite many reported multiplier cir­cuits, only two cancellation methods for four-quadrant multi­plication are known. Because a single-ended configuration does not completely cancel nonlinearity and has poor PSRR, a fully differential configuration is often necessary in a sound multiplier topology. The multiplier has two inputs. Therefore there are four combinations of two differential signals, that is (x, y), (—x, y), (—x, —y), and (x, —y). The multiplication and cancellation of an unwanted component are achieved by ei­ther of the following two equalities:

These two approaches are depicted in Fig. 6. The topology of Fig. 6(a) is based on two-quadrant multipliers. Fig. 6(b) is based on square law devices. X and Y are arbitrary constant terms and are not shown in Fig. 6.

ION-SELECTIVE GLASS MICROELECTRODES

By adding a membrane material to a glass micropipette, the activity of ions in sample solutions can be measured. The selectivity depends mainly on the type of membrane used. The pH of solutions can be measured with a glass membrane, and recent developments in polymer mem­branes have made it possible to measure the activities of Na+, K+, Ca2+, Cl-, and other ions. These ion-selective microelectrodes are inexpensive and are small enough to measure the activity of ions in both extracellular and in­tracellular spaces. Abnormal cellular states often result from an imbalance of ionic concentration. Therefore, the fact the activity of ions (not just the concentration) can be measured with very small electrodes compatible with glass micropipette technology has provided researchers with a powerful method to analyze both the normal and diseased physiological states.

A membrane is chosen to allow the diffusion of certain ions selectively and then is inserted into the glass mi­cropipette. The potential difference generated between an external reference electrode (Ag/AgCl, for example) and the ion-selective microelectrode immersed in the sample solu­tion is given by the Nicolsky equation (13):

where Vr is a constant and takes into account the voltages within the electrochemical cell, ai is the activity of the ion in the sample, aj is the activity of the interferention j, kij is the selectivity constant of the electrode for ion j rela­tive to ion i, x the charge of ion i, R is the gas constant, T is the absolute temperature, г is the charge of the ion, and F is the Faraday constant. Typically, electrochemical cells have very large impedances, and the voltage differ­ence must be measured with high input resistance ampli­fiers (-1015 Q), such as electrometers. The selective mem­branes can be grouped into three classes: glass, liquid, and solid state. Glass membranes are very sensitive to the con­centration of H+ ions. Liquid membranes are commonly used in biological preparations for other ions. For exam­ple, potassium ion activity can be measured using a liq­uid membrane containing the antibiotic Valinomycin com­bined with a resin. When applied to the extracellular space of brain tissue undergoing epileptiform activity, the potas­sium concentration observed rises from its resting value of 7 mM to 11 mM. The rise in potassium activity is ac­companied by increased neuronal activity and is similar to neural activity observed during epilepsy. The last class of membranes is made of solid-state material, such as crys­tals or insoluble salts. For example, a pellet of silver sulfide can be used to detect Ag+ with very high sensitivity.