Author Guidelines for 8 - Ryerson University



DIgital Pulse INterVAL MODULATION AS AN

ALTERnative Modulation Scheme for FRee Space Optics

Mohammed Huroy

RYERSON UNIVERSITY

Abstract

This paper discusses the solutions and limitations for the Last Mile Challenge in high-speed fiber optic communications. This paper examines Digital Pulse Interval Modulation (DPIM) as an alternative to other modulation scheme such as On-Off Keying (widely used) and Digital Pulse Position Modulation, as way to increase the transmitted (laser) power for free space optical communications. This increase of power will also be analyzed to combat; transmitter and receiver misalignments, and attenuation due to weather conditions. The characteristics such as bandwidth efficiency, power efficiency, bit-rate, and complexity will be compared between the modulation schemes. Finally, the use of a Q-switched laser with DPIM will be also be analyzed.

1. Introduction

Most Metropolitan Area Networks (MAN) of today are inter-connected through high speed fiber optic communication links, offering bit rates in the high Mbps and Gbps range (see Figure 4.1). However, most end-consumers within the MAN are connected to this fiber optic backbone through a low speed connection such as the T1, xDSL, and Cable or Microwave links. It is also interesting to note that only 5 percent of the buildings in the United States are connected to the fiber optic backbone, yet 75 percent of the other buildings are within one mile of fiber[1] (see Figure 4.2). A very interesting and promising solution to this stumbling block is the use of free space optics (FSO). The optical links (a laser and photodetector) are installed on building rooftops where access is required. Through this configuration, the end-consumer is connected to the fiber optic backbone through this high speed link that provides bit rates in hundred Mbps range and also in the Gbps range in certain cases. Most importantly a license is not required for its operation. One of its major advantages is that the setup time is very small and the cost is relatively cheaper. A fiber optic link would require digging very busy districts and high associated costs for permits, etc.

One of the most important aspects in a free space optical link is the precise point-to-point alignment of the transmitter (laser) and the receiver (photodetector). Any minor side drifts of the laser or the photodetector will result in transient broken links due to very low optical power being received at the photodetector. These drifts may occur due to the building’s motion or the pole if the laser or the photodetector is tethered to it. Tracking systems have also been used to correct the alignment of the laser-photodetector arrangement. These systems utilize some sort of feed back link where by the receiver would tell the transmitter that it is drifting away and transmitter would alter its position using some sort of precise servo motors, and vice versa. These tracking systems will, however, increase the over all cost of the optical link.

Another important aspect affecting the performance and reliability of free space optics is the attenuation, which is directly related to the weather condition. Severe weather conditions (heavy fog, snow, or rain) may hinder the link. Ideally, it would be desired to increase the transmitted (laser) power, however there many factors limiting the output power. Some links offer an RF link as a backup incase the optical link is not reliable any more (see Figure 4.3). The RF backup would employ a very directional (high gain) antenna to provide a point-to-point link. Unfortunately, the RF link does not offer bit rates close to the one offered by the optical counterpart. Moreover, including an RF link, once again, increases the setup and maintenance cost.

A solution is desired to somehow aid the free space optical link in severe weather conditions and also become insensitive to alignment drifts between the laser transmitter and the photodetector receiver. An alternative modulation scheme to OOK (On-Off Keying) such as DPIM and will be discussed as a way to increase the transmit power. Finally, the implementation of transmitting high laser power through a Q-switched laser will be discussed.

2. THEORY

2.1. Free Space Optics (FSO)

Free Space Optics is exactly the same as regular fiber optic communication except that the medium in which the light travels is not the fiber but free space (air). The transmitter (laser) and the photodetector (photodiode) are placed on building rooftops instead of burying fiber between the buildings, which is a substantial cost. The major difference between fiber optic and free space optic communication is the attenuation. The attenuation in fiber is constant and predictable while the attenuation in free space (air) is dependent on weather conditions and is unpredictable in certain cases. Free Space optics can offer similar bit rates as fiber optics. Since a fiber optic link can be replaced by a free space optic link, all other building can now be connect to the fiber optic backbone through FSO providing bit rates from 100Mbps to 1.25Mbps (see Figure 4.4).

2.2. Power Link Budget

The Power Link Budget is the concept used to determine the received power at the photodetector for a certain amount of transmitted power based on certain parameters such as attenuation, beam divergence, etc. The link equation is relating all of these parameters are shown below:

[pic]

(1)

where,

PTX – Power Transmitted

PRX – Power Received

dTXA – Transmit Aperture Diameter (m)

dRXA – Receive Aperture Diameter (m)

D – Beam Divergence (mrad)

R – Range (km)

( – atmospheric attenuation factor (dB/km)

Looking at the equation, the atmospheric attenuation is uncontrollable in an outdoor environment and is almost independent of the operating wavelength for high levels of attenuation. Also, for high levels of attenuation, (the atmospheric attenuation is the dominant term by many orders of magnitude due to the logarithmic relationship.

2.3. Lasers

A laser is coherent and a monochromatic source of light. Lasers are being used in all kinds of application such as medicine, industrial, communication, etc. In terms of communication, lasers are primarily used as a transmitter of incoming data. The data is transmitted by modulating the laser’s intensity (power per unit area). There two different medium in which light travels to transfer information. One is through fiber, in which the intensity is attenuated as the light travels through the fiber. The second is free space (air), which is the medium used for FSO.

2.3.1. Intensity, Beam Radius and Divergence

The laser intensity is not only attenuated but it also spreads as the light moves through the air since it is not confined, as in the case in a fiber. The intensity of a laser for the lowest TEM mode (TEM00), which results in a Gaussian output beam as the light beam leaves the laser transmit optics is given by:

[pic]

(2)

where,

I(r) – Intensity of the beam

IO – Peak intensity

r – Distance away from IO

RB – Beam Radius

The intensity of a laser follows a Gaussian profile (see Figure 4.5). The beam radius is defined such that the intensity equals Io/e2. In other words, the beam radius is the region of the laser spot where 85% of the intensity (power) is contained in. As light propagates through the air, the size of the Gaussian profile decreases and the beam radius/diameter increases (see Figure 4.5 and Figure 4.6). The equation relating the beam radius and the distance away from the laser is shown below:

[pic]

(3)

where,

R(z) – Beam radius at z

z – Horizontal distance away from the laser

( – Wavelength of the light

RB – Beam Radius

Incorporating equation (3) into equation (2) to reflect the dependence of the intensity on the distance is shown below:

[pic]

(5)

In terms of the received power, the area of the photodetector is the most important factor. Ideally, it would be desired to have the area of the photodetector to be the exact size or even bigger than the laser spot to maximize the received power. However, due to the limitations by manufacturability and most importantly the capacitance (which is directly proportional to the area of the photodetector), the photodetectors are extremely small in comparison to the expanded beam. Therefore, the received power will be small. In an ideal case with no attenuation, if a circular photodetector has with a radius Rpd, the power received at a distance zo away from the laser and perfectly placed at the center, where the intensity is at its peak, will be:

[pic]

(5)

2.4. Atmospheric Attenuation

Atmospheric attenuation is primarily due to atmospheric absorption and atmospheric scattering. In both cases, the attenuation levels can vary anywhere between 30dB to 350dB. Attenuation in FSO can be as good as fiber optics (for a clear or hazy day) but it can also be extremely poor (extreme fog) such that the link may not be reliable anymore. Attenuation is dependent on the weather conditions (which is dependent on the city). Some Cities are very horrendous for FSO (San Diego, CA) and some cities are excellent for FSO (Tuscan, AZ). Table 1 shows the different attenuation numbers for different weather conditions for two primary wavelengths used in optical communications. Atmospheric Attenuation dependence on wavelength is extremely miniscule and the due to the logarithmic relationship, the effect on the received power is the same. There are other benefits for operating at 1550nm instead of 785nm, which will be discussed later.

Table 1.1 Attenuation Levels as a function of

visibility and wavelength

[pic][2]

2.5. Pulse Time Modulation (PTM)

Many different modulation schemes are used for different application in communications. The standard modulation schemes modulate the amplitude, phase or frequency of a carrier signal. However, PTM involves modulation of the timing of a carrier signal. There are different kinds of PTM (see Figure 4.7). The one of interest is DPIM as an alternative to On-Off Keying (OOK), also know as Amplitude Shift Keying (ASK). Another modulation scheme, PPM, will also be discussed for comparison as well. In general, PTM schemes have better Average Power efficiency (APE) and can also provide even higher APE at the expense of higher BW requirement. Hence, it is possible to increase the Peak-to-APE. If there are M-bits, PPM/DPIM will use L = 2M symbols or for L symbols, there will be M = log2L bits being mapped.

2.5.1 Digital Pulse Position Modulation (DPPM)

In a PPM scheme, depending on the number bits (M), there will be 2M symbols. Each symbol period will be divided in to L intervals in which a single pulse will only occupy a single interval. The pulse’s position with in the symbol period will vary and will depend on the incoming bits (see Figure 4.8). The symbol period TS is given by number of bits (M) divided by the bit-rate (RB) and each interval (TI) has duration of TS/L. Unfortunately, at the receiving end, symbol synchronization is required which increases the complexity for implementation.

2.5.2 Digital Pulse Interval Modulation (DPIM)

DPIM is almost similar to DPPM in the sense that the incoming bits are modulating the distance between successive pulses (see Figure 4.8). However, in DPIM, the pulse’s position is always in the same interval for every symbol period. The number of intervals between successive pulses is numerically determined by the group of M-bits (digital word). As it can be seen from Figure 4.8, as the decimal equivalent of the digital word increase so does the number of intervals following the pulse. If a decimal zero is being transmitted then one interval will be followed by the pulse. An extra interval is added to avoid zero intervals between pulses for decimal zero. The symbol period for DPIM is not constant as is the case for DPPM. DPIM can have symbol period as small as 2TI and as large as (L + 1)TI. The average symbol period would simply be (L + 3)TI/2. Fortunately, at the receiver end, symbol synchronization is not required which makes its implementation very simple. The symbol length is determined by simply counting the number of interval between the received pulse or by simply counting the total time interval between the pulses and dividing it by TI, which is known for the overall system.

3. ANALYSIS

The main issue today with FSO is either reliability or range. The objective of this paper is to discuss of ways to increase reliability for a fixed range. The most important obstacle, as mentioned earlier, is the atmospheric attenuation. For instance, if the range selected for a dedicated link between two buildings is 300m and the link is only 90% reliable for a whole year, then reliability can be increased to 99% simply by significantly increasing the laser output power. It is important to note that on clear or hazy days where the attenuation levels are small (see Table 1.1), the link can be 100% reliable but as the weather deteriorates so does the reliability. Hence, the 90% reliability is for the worst case scenario (300+ dB/km). It is always desired to increase the power of any kind of transmitter to combat attenuation. However, there are always limitations in terms of how much power a laser can output. First, the laser has its maximum average and peak power rating. Second, for safety reasons the average power is capped. In terms of safety, low power levels are allowed for wavelengths under 1400nm, since light is focused on the retina. On the other hand, any wavelengths above 1400nm are absorbed by the lens and the cornea of our eyes, allowing higher power levels. This is a reason why 1550nm is desired vs. 785nm.Usually, the lasers currently being used are already operating at their limits in very bad weather conditions where the attenuation reaches around 350dm/km. The highest peak power seen in the market today is around 2 – 5 Watts. The laser systems used in FSO have power adjustment mechanisms incorporated. Hence, the lasers are not transmitting at those power levels all time unless it is extremely foggy all time which is not case. Therefore, the lasers are actually transmitting smaller power levels most of time.

Unlike fiber optics, the precise alignment between the transmitter (laser) and the receiver (photodetector) is another important obstacle for FSO. The laser beam spot increases as the distance between the laser and the receiver increases. A standard laser with a laser beam diameter of 2mm can increase to 30cm at a distance of 300m away. As mentioned earlier, it would be desired for the photodetector diameter to equal the expanded beam diameter to capture all the transmitted power. However, the photodetector is extremely small (100mm2 area, max) in comparison to the expanded beam diameter. The photodetector is aligned in such a way to maximize the received power. The receiver is placed at the center of the incoming laser beam and focused onto the receiver (see Figure 4.10). Obviously, the intensity at the center of the laser beam spot is at its peak and decreases as u move further away from the center. Unfortunately, since the laser/photodetector system is placed in building rooftops, structural vibrations, wind gusts and building sway will alter the alignment. If there is any misalignment, the power being received from the laser will not be from the center which will severely reduce the focused power/intensity. However, by increasing the laser output power, the overall intensity will increase; therefore the focused power onto the receiver will be also increase. The amount of power increase, however, should be substantial.

In order to increase the output power of a laser, other techniques or other lasers have to be used. A Q-switched laser is an interesting laser such that the average power is moderate but the peak powers are very high. A Q-switched is a type of pulsed laser which shortens its output pulse width and boosts peak output power. The short laser pulse duration of the Q-switched laser is in nanoseconds or picoseconds, and peak power in the hundreds of Watts can be achieved by small sized lasers. However, for FSO applications it is very sufficient transmit peak powers of 5 to 20 Watts. However, it is inefficient to use a Q-switched laser with OOK modulation scheme. If the incoming data of 0’s or 1’s are equiprobable, then for alternating 1’s and 0’s or consecutive 1’s, the output peak power will be used frequently, especially for consecutive 1’s. This is where DPIM comes in. As it can be seen in Figure 4.9, less power is used for PPM and DPIM since there are few on-pulses because the incoming data is stored in the time interval between 2 pulses. DPIM and PPM also use pulses with a much shorter pulse width. If M is increased, even less power is used and the shorter the pulse width. Therefore, a Q-switched laser can be used more efficiently with DPIM and PPM. This high APE is achieved, however, at the expense of higher bandwidth requirement. As it can be seen from Table 1.2, the power requirements for DPIM and PPM shown, is in order to achieve a bit-error rate similar to OOK However, the bandwidth requirement is increased for DPIM and even more for so PPM. Even though DPIM/PPM schemes have a higher BW requirement, in optical communications there is enough BW to allow for this increase.

Table 1.2 Bandwidth and Power Efficiency

for DPIM and PPM vs. OOK

[pic]

It is clearly shown that it would be preferred to use DPIM instead of PPM, due to lower BW requirement and simpler receiver structure. Hence, a Q-switched laser can be used efficiently with DPIM due to higher APE than OOK, allowing for high peak output power but with a moderate average power. Therefore, when higher power levels are being transmitted, this can help withstand the high attenuation. Plus any minor misalignment between the TX/RX would not be affected since the intensity of the laser beam away from the center will be significantly higher than usual, hence increase the power received by the photodetector.

4. Illustrations

[pic]

Figure 4.1 Map Example of inter-connected MANs.

[pic]

Figure 4.2 The Last Mile Problem[3]

[pic][4]

Figure 4.3 A TX/RX FSO system with a RF backup

[pic][5]

Figure 4.4 Buildings connected in a mesh configuration

[pic]

Figure 4.5 Gaussian Profiles

[pic] [pic]

(a) (b)

Figure 4.6 (a) Laser intensity confined in a small spot

(b) Laser intensity expanded at a far distance

[pic][6]

Figure 4.7 Pulse Modulation Tree

[pic]

Figure 4.8 Simple definition of PPM and DPIM, M = 2

[pic]

Figure 4.9 Waveform comparisons for the

different modulation schemes

[pic]

Figure 4.10 Perfect Transmitter and Receiver alignment

5. Conclusion

This paper has discussed the issues facing free space optics and how to contend with it. This paper discussed the used of Digital Pulse Interval Modulation (DPIM) as an alternative modulation scheme to On-Off Keying (OOK) to be used efficiently with Q-switched laser. In which, the Q-switched laser offers higher peak output power in comparison to other standard lasers. Moreover, the high Average Power Efficiency (APE) is achieved at the expense of higher bandwidth (BW) requirement which is tolerable under optical communication, where the BW is almost abundant. Finally, the analysis was based for a fixed range, with objection of increasing reliability. However, for certain cities where attenuation levels are low (< 10dB/km) all year round, the use of DPIM and Q-switched lasers can not only provide 100% reliability but also provide extended range.

6. BIBILIOGRAPHY

[1] S. Bloom, J. Schuster, and H. A. Willebrand, “Understanding the Performance of Free-Space Optics”, WCA Technical Symposium, San Jose, CA, January 14, 2003

[2] M.T.T. Lynn, "Modulation Techniques for Optical Communications: Infrared", May 9, 2003

[3] J. Zhang, "Modulation Analysis for Outdoor Applications of Optical Wireless Communications", IEEE, Finland

[4] E.D. Kaluarachi, Z. Ghassemlooy and B.Wilson, "Digital Pulse interval modulation for optical free space communication links", IEEE, London, UK, 1996

[5] S. Bloom, "The Physics of Free-Space Optics", AirFiber, Inc., May 2, 2002

[6] D. A. Rockwell and G.S. Mecherle, "Optical Wireless: Low-Cost, Broadband, Optical Access", fSONA Communications Corporation, Richmond, BC

[7] I. I. Kim, B. McArthur, and E. Korevaar, “Comparison of laser beam propagation @ 785nm and 1550nm in fog and haze for optical wireless communications”, Optical Access Incorporated, San Diego

[8] I.I. Kim, and E. Korevaar, “Availability of Free Space Optics (FSO) and hybrid FSO/RF systems”, San Diego, CA

[9] I. I. Kim, B. McArthur, and E. Korevaar, “Comparison of laser beam propagation @ 785nm and 1550nm in fog and haze for optical wireless communications”, Optical Access Incorporated, San Diego

[10] D. A. Rockwell and G.S. Mecherle, "Wavelength Selection for Optical Wireless Communication Systems", fSONA Communications Corporation, Richmond, BC, February 2001

[11] E.D. Kaluarachi, A.R. Hayes, Z. Ghassemlooy and N. L. Seed, "Digital Pulse interval modulation for optical free space communication links", IEEE, London, UK, 1996

-----------------------

[1]

[2] I. I. Kim, B. McArthur, and E. Korevaar, Comparison of laser beam propagation @ 785nm and 1550nm in fog and haze for optical wireless communications, Optical Access Incorporated, San Diego

[3] See 2

[4] I.I. Kim, and E. Korevaar, Availability of Free Space Optics (FSO) and hybrid FSO/RF systems, San Diego, CA

[5] See 2

[6] M.T.T. Lynn, "Modulation Techniques for Optical Communications: Infrared", May 09, 2003

-----------------------

Fiber Optic Cable

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download