Directive antennas offer improvements for cell sites In the beginning, service providers placed their cell sites as high as they could, and added high-gain, omnidirectional antennas and blasted out as much power as they could to reach as far and as wide a
Cellular providers are faced with the vexing problem of boosting capacity in their systems, which in turn increases interference at a time when call quality is an increasingly important issue. Metawave has attacked these problems by equipping cell sites with highly directive antennas that “track” desired mobiles and block out undesired ones. The result can be 50% better coverage, markedly better call quality, double battery life, and double system capacity.
Creation of cellular In the beginning, service providers went to the mountaintop, placed their cell sites as high as they could, and added high-gain, omnidirectional antennas and blasted out as much power as they could to reach as far and wide as possible to serve their fledgling flock. Since there were lots of channels and not many customers, frequency re-use was hardly a consideration, and co-channel interference was not an issue. But, the deluge of customers, although a blessing for the marketing and sales departments, brought the curse of implementation for the engineers. The engineers had studied the BSTJ Bible (remember the old Bell System Technical Journal) well, however, and knew exactly what to do.
Mythical 7-cell re-use never happened With six potential interferers around a cell site, odds are at least one is going to be strong enough to cause trouble often enough to be annoying and occasionally downright disrupting. Long before this happened, operators, as taught by their BSTJ-trained engineers, switched over to sectors. With sectors, the coverage pie is divided into three 120-degree-wide pieces. By doing this, at most, two co-channel interferers are visible, compared to six before. Besides, the interferers are a little farther away so that, on average, the interference power they create is somewhat lower. Now, the theory says that co-channel interference should be a rare event. An example of this is when a customer is at the edge of a cell and is shaded from the cell site a little bit, and at the same time a co-channel mobile a few cells away is in a similar situation. So over there is an interfering mobile at the edge of its cell trying to get a signal back to its cell, but spews energy across the adjacent cells and causes interference back at the home cell. Again, the theory says the propagation laws are such that the interference will always be too weak to cause much of a problem. However, everyone knows Murphy’s laws are stronger than physical laws, and Murphy plays a mean game of statistics. So it seems that whenever the mobile in our cell drops into a little propagation shadow, the interferer pops up into the clear causing havoc. Still, sectorization has helped a lot, because the directive antenna blocks out the interference from behind quite well, and from the sides fairly well.
Playing statistics with Murphy The logical question is, if three sectors are better than omni, why not use six, 12, or 24 sectors and be even better? The theory says yes, it should work, but to be convincing, real-life propagation measurements were called for to find out what other secrets Murphy had lurking out there. Our company has performed extensive field tests at several locations in urban, suburban and rural areas to quantify the actual “what if” situation of 12 or 24 antenna beams being available at a cell site. The results showed that 12 beams provided a remarkable improvement in interference reduction close to the theoretical 6dB (i.e.: the average interference is reduced by 75%). (See Figure 1 on page 10.) With these test results in hand, Metawave settled on a 12-beam solution. Even though 24 beams provided more improvement than 12 beams, the added cost and complexity was not cost-effective. As we consider implementation in the next section, this will become clearer.
More sectors alone will not help capacity We can prove experimentally that a 12-sector system significantly reduces interference; however, sectorization alone actually reduces capacity. For example, an omni cell and a tri-sectorized cell each have a total of 57 channels, yet any one sector has only 19. Thus, for 1% blocking, the sector cell carries 11.2 erlangs* [ *An “erlang” is the equivalent to 1 voice circuit in use at all times.] of traffic per sector, or a total of 33.6 erlangs, where the omni cell, having the benefit of a larger trunk group, can carry 44.2 erlangs for the same blocking probability. Carrying this example to a 12-sector system leaves on average slightly less than five channels per sector, which carries a pitiful 1.3 erlangs at 1% blocking. Of course, if the reuse cells can be brought closer together, capacity can be increased. A seldom-used reuse of four (N = 4) six-sector systems still has smaller trunk groups than a reuse of seven (N = 7) three-sector systems. Now, imagine if one could have the trunking efficiency of an N = 7 omni system combined with the interference rejection of a 12-sector antenna or, better yet, an N = 4 tri-sector grid with a 12-beam overlay. (See Figure 2 on page 12.) The respective capacity increases compared to a traditional N = 7 tri-sector system would be 31% for the N = 7 system and a whopping 100% for the reuse of four system. This is precisely what our system offers.
The antennas are dumb, the smarts are in the algorithms The logic is simple–just connect the narrowbeam antenna to the radio in the cell site that best serves each mobile. The best signal may not necessarily mean the strongest signal because it could be corrupted by interference. Also, there is a need to address the question of the diversity port on the radio, and what about the transmitter? The answers to these and a host of other questions lie in algorithms based on extensive field measurements, propagation models, and self-learning. At the same time, it is extremely advantageous to be transparent to the existing hardware and software operating the cell site. Our solution provides 12 antenna beams with an active switch matrix as an overlay unit that replaces the sector antennas. For transmission, the current Class C amplifiers or linear amplifiers are replaced with a set of 12 amplifiers, one for each beam. For starters, there is a significant benefit available just from a receive-only system (See Figure 3 at the left.)
Receive-only system benefits portables Cellular systems were designed for 3W mobile transmitters mounted on a ground plane (car roof). Among the many operating conditions, this determined the spacing between cell sites. Now along comes the three-fifths (0.6W) portable used inside the car, not outside. This often leaves coverage gaps between cell sites because there is not sufficient energy to reach back to the old cell site or forward to the new one. In rural areas it is not worth the cost of adding another cell site just to stop a few dropped calls; however, high-gain directive antennas can receive a ten times weaker signal compared to an omni antenna and roughly a four times weaker signal compared to a sector antenna. The greater sensitivity easily fills in the coverage gaps. Because the cell transmitter always has much more power than a portable, this configuration can ensure a link between cell sites when portables are used.
In urban areas, traffic demand can justify another cell site; however, the need for a narrowbeam antenna solution in this case is perhaps more important. In this environment we find portable users everywhere, especially in buildings. Again, since the narrowbeam directive antennas can detect a much weaker signal, better building penetration is obtained. This obviously applies to any area around the cell where coverage was marginal, so quality improves everywhere. A side benefit of the narrowbeam directive antenna solution is that if a mobile user stands by a window in a tall building, the mobile’s signal can reach out too far. If the cell using the same channel has a narrowbeam antenna system, the chance of that mobile causing interference is only 1/12 of what an omni cell site would experience, since there is only a one-in-twelve chance that the directive antenna is pointed toward that interfering mobile. Finally, in stronger signal areas, less power is needed. By reducing the power needed, battery-life can be extended, which equates into increased talk time.
Tx/Rx system adds capacity Adding transmission capability through a narrowbeam antenna provides the same reduced power benefit for the cell as it does for the mobile. However, the real benefit for the service provider is that the same cell can provide 30% to 100% increase in capacity. Because a narrow antenna beam is directed precisely at the mobile user, the amount of unnecessary and definitely unwanted signal energy that falls outside of the cell is reduced. The hardware and software are configured so that the best antenna signal for reception is also used for transmission. The second best signal on reception is routed to the radio diversity port. Test results have shown that this is an excellent diversity solution, eliminating the need for a second space diversity antenna. As mobiles move about, the directive antenna system constantly monitors the environment with independent scanning receivers that can be tuned to any channel in the system and connected to any antenna beam in a fraction of a second.
Simple antenna structure Our system uses a panel antenna about 2.5 feet wide that actually produces four side-by-side, 30-degree antenna beams. Therefore, with just three antenna panels, a full 360-degree circular coverage is obtained. Each panel can be aligned to match a 120-degree sector for sectored cell sites, so the same antenna can be used for sector or omni. In fact, if propagation permits (or dictates) that a sector be larger or smaller than 120 degrees, it is just a matter of adding or subtracting more 30-degree elements to define a sector.
Implementation The most cost-effective method of implementing a narrowbeam directive antenna system is as an appliqu‚ on a cell site. The existing radio hardware and software doesn’t know about the smart antenna system, it just works better. Figure 4 above shows how easy it is to implement for reception.
Transmission similarly routes the outbound signal through a spectrum management unit (SMU) and a linear power amplifier (LPA) to the proper antenna. The platform is designed to be modular because not all cell sites carry the same amount of traffic. Each plug-in SMU handles six radios. Each card cage holds six SMUs, and there is no real limit to the number of card cages. There is no requirement that a SMU be dedicated to a sector. In fact, by splitting up the SMU outputs, a worst-case failure of a SMU card would be a loss of only two radio channels in any sector. With a design objective of 100,000 hours mean time between failure for a SMU, there is no need for redundancy, except in extreme cases.
What comes next? With the narrowbeam technology each cell site is designed to operate independently and to optimize its antenna beam system to best serve the mobile customers in its own cell. However, there are situations beyond the control of the local cell site that provide opportunity to either increase capacity or reduce interference on a dynamic basis. Thus, each cell site’s capability to report in real time to a host provides an opportunity to optimize the entire network. Actually, a close approximation to an optimum network can be obtained by looking at the activity of a cell and its nearest neighbors. A cell that has a problem with a particular channel on a particular beam can ask for help by sending this information to a host computer. The host can relay the information to the site causing problems for action. This relaying of information will lead to dynamic beam and channel assignments, which allows even tighter packing of cellular customers while retaining high cell quality. We have developed a networked product, MetaNet, that provides a central store for trouble reports, frequency planning, growth planning and records time data in minute increments and can log a map grid down to 0.1 square miles, if desired. The networked product is a platform for optimizing network performance while providing the capability for future enhanced services based on its distributed information routing architecture.
Where Does It End? If and when analog dies (remember, there are 30 million analog customers today and still growing), will the narrowbeam technology work with a digital system? The answer is yes, for both time-division multiple access (TDMA) and code-division multiple access (CDMA). Although the solutions are different, the directive antenna allows you to pick up the reverse link advantages of narrowbeam antennas for either system. For TDMA, the switch matrix keeps in step with the time hops and switches over to the right antenna for the next mobile during the channel slot guard time. For CDMA, a pre-combiner configuration collapses the 12 beams into the usual CDMA sector configuration. Both analog and digital service providers will be able to get the most out of their networks through the use of narrowbeam technology. The benefits of the narrowbeam technology will allow service providers to reduce cell site interference, to extend coverage and to increase capacity to serve the staggering growth of wireless customers.