Using Smart Antennas to Improve Mobile Ad Hoc and Mesh Networks

RF_Doc.

As wireless networks have moved from coffee shops and college campuses into the homes of soccer moms and blue-collar dads, some problems have arisen, most often from the fact that the range or performance of the wireless access point is not as advertised, leaving large dead spots within the home or small business. While device manufacturers or engineers understand that this is a fact of wireless life, consumers will often blame the type of device and seek to return it. With this in mind, it is even more important for device manufacturers to understand how range, performance, power drain, or interference problems in any wireless system can be solved.

There are several options for overcoming these problems, but the two we'll discuss here today are smart antennas and ad hoc networks-and how both can be correctly used together to increase the range and performance of wireless systems.
Smart Antennas: Two Types, One Clear Choice

A smart antenna is commonly defined as a multi-element antenna where the signals received at each element are

1a – Multibeam

intelligently and adaptively combined to improve the overall performance of the wireless system, with the reverse performed on transmit. The benefit of smart antennas is that they can increase range and capacity of systems while helping to eliminate both interference and fading. There are two types of smart antenna-multibeam (or directional), where the beam with the strongest signal is selected for reception and transmission; and adaptive array, where the signals from several antenna elements are combined to maximize the output signal.
There are pros and cons to each type of antenna-to start with, the multibeam antenna is less complex than the adaptive array antenna, using simple beam tracking instead of the complex calculations required by an adaptive array antenna to combine and maximize the signal power. In addition, while both types of antenna can provide again increase in range and performance, they each have different situations where they can best make this happen.

1b - Adaptive Array

With multibeam antennas, the full increase in range and performance only occurs in environments where there is a line-of-sight and/or limited interference. It's very unlikely that signals will arrive in a single beam, in a single direction. In business and home environments there is a great deal of opportunity for interference, meaning signals will arrive from multiple directions, limiting the opportunity for performance gain in multibeam antennas.
In comparison, an adaptive array antenna can provide the gain in range/performance in whatever environment it is set up in, as long as the antenna elements are spaced at least a half wavelength apart from each other and can suppress interfering signals. In an environment with a great deal of interference (a multipath environment), adaptive arrays can suppress this interference no matter the point of origin, and diversity gain can also be achieved (as long as the signals have independent fading at the antennas).

Figure 2: An ad hoc network

Most wireless systems-Wi-Fi, WiMAX, cellular, UWB, GPS, RFID or satellite audio and video-can benefit from the addition of an adaptive array antenna. Since Wi-Fi systems are time-division-duplex (TDD) systems, the received weights can be used for transmission to obtain the same gains in both directions with the use of smart antennas on one side (client or AP) only. As examples, a four-antenna array can provide up to a 13 dB signal-to-noise ratio gain (6 dB array gain plus a 7 dB diversity gain), or the possibility of data rates as high as 500 Mbps (as considered for IEEE 802.11n). Similar gains can be achieved in WiMAX systems (particularly those using TDD), and gains on the order of 6 dB can be achieved in cellular systems which have frequency-division-duplex operation.

In fact, because of its ability to provide gains in any environment-line-of-sight or multipath-while at the same time blocking interference and providing substantial capacity increases, the adaptive array is the main focus for smart antennas in wireless systems, including the IEEE 802.11n standard currently in development.
Wireless Ad Hoc Networks: Advantages and Concerns

The definition of a wireless ad hoc network is a network of mobile hosts with no pre-existing infrastructure. Multiple hops are used for routing, and the neighbors and routing changes with time (with user mobility and changes in the environment).

A type of wireless ad hoc network is a mesh network, which is made up of an infrastructure of fixed, regular hosts. Wireless ad hoc or mesh networks are ideal for home networking, temporary networking (as seen for meetings and conventions), industrial settings, and military or emergency networking.

Ad hoc networks have several advantages-they use a lower amount of power in transmission, resulting in a longer battery life; they're easy to deploy, meaning they can be quickly set up; they have performance that is not critically dependent on the infrastructure, increasing the locations where they can be ideally deployed; and they have a higher frequency reuse, meaning a higher possible capacity than other types of networks.

As with any wireless system, the technology is not entirely perfect-as there are several issues with wireless ad hoc networks that can limit their effectiveness. These issues include power limitations, range issues between nodes, signal routing changes due to movement or failure of nodes, and fading or packet loss. Further complicating wireless ad hoc networks is the mixture of different types of users or equipment within the network, or frequency reuse limits resulting from interference, and a situation known as "the hidden node problem."

The hidden node problem is defined in the below diagram (Figure 3). In a wireless ad hoc network, nodes 1 and 2 and nodes 2 and 3 are close enough to communicate, but nodes 1 and 3 are too far apart to hear each other. This means that if node 1 is transmitting to node 2, node 3 is too far away to "hear" the transmission-and will think the channel is clear. Node 3 will then transmit to node 2, resulting in a devastating collision at node 2 of separate packets from both node 1 and node 3-losing both packets of data, and effectively eliminating the usefulness of the network.

Figure 3: Illustration of the hidden node problem

To avoid the hidden node problem, wireless ad hoc networks can be set up to use request-to-send packets (RTS) and clear-to-send packets (CTS), as is done in 802.11. This would result in a situation where if node 1 has a packet to send to node B, it sends an RTS to node 2, node 2 responds with a CTS, and the data is then sent.
Adding Smart Antennas to Ad Hoc Networks: a Right and a Wrong Way

When added properly to wireless ad hoc networks, smart antennas can produce a range increase and capacity gain. However, they must be added properly-simply adding an antenna is not enough, as it could decrease the network capacity when incorrectly implemented.

The majority of wireless ad hoc networks use only omni-directional antennas, which reserve the spectrum over a large area-wasting limited network resources. However, by using smart antennas in wireless ad hoc networks, this problem is reduced, as the signal energy is focused where it is needed, and gains in range and capacity can be seen. But these gains are dependent on what type of antenna is used, and how the antennas are implemented.

Typically, the type of smart antenna considered for wireless ad hoc networks thus far has been the multibeam (or directional) antenna, mainly because it is seen as easier and less costly to implement and to study-but that is not necessarily true.
Multibeam Antennas: Multipath Problems

It has been shown that multibeam (directional) antennas can increase the range and capacity of wireless ad hoc networks and provide greater frequency reuse for the network through the use of a media access control (MAC) that incorporates the use of these antennas. For example, if node 1 (the transmitter) knows the location of node 2 (the receiver), then the request-to-send (RTS) can be sent with a directional beam, although it would be received with an omni-directional beam at node 2 since node 2 would not know that the RTS was sent. Node 2 would then send the clear-to-send (CTS) with a directional beam, which would be repeated when the data is sent from node 1 to node 2 and the acknowledgement from node 2 to node 1. This process results in increased frequency reuse, better connectivity, increased range and lower battery usage, but does not answer the problem of the hidden node. This is worsened by the asymmetry in the gain at node 2, and by the loss of receive gain for the RTS packet, which means that node 2 doesn't know which beam to use when node 1 transmits first.

In addition, another common problem with directional antennas can result in a large reduction of the network's overall capacity. To illustrate this situation, consider two access points, where access point AP1 uses a directional beam, and access point AP2 uses an omni-directional beam, as shown in Figure 4. With beacons transmitted omni-directionally from both AP1 and AP2 (as they are intended for all clients), a client will associate with the closer access point-AP2-rather than AP1, with which it should be associating, because its directional beam will provide a stronger signal for the data. In addition, once a client associates with an access point with a directional beam, it may continue to associate with that access point-even after it has moved beyond other access points (because the received data signal is still strong enough), resulting in a large capacity reduction in the network.

Figure 4: A mixed mode network with association problems

Most importantly, however, is the fact that directional or multibeam antennas, as discussed earlier, do not work well in multipath situations. When implemented in such environments, the direction-of-arrival may not be a good indicator of the user's location, and the potential problems can eliminate all gains. Add to this the fact that a variety of factors, such as the beamwidth, location of the client relative to the access point, the number of beams, and the location and beamwidth/pointing direction of adjacent access points, must be taken into account in order to do radio resource management in multibeam networks, and it becomes clear that the use of multibeam antennas can greatly complicate the radio resource management process.
Adaptive Arrays: Multipath Successes

The difference between the success or failure of directional or adaptive antennas in wireless ad hoc networks comes down to one main issue: multipath environments. Adaptive arrays have been proven to work well in multipath environments, as they can provide the full gain and suppress interference in any environment. Unlike multibeam antennas, adaptive antennas can "listen" omni-directionally, but beamform when the packet is received-achieving a smart antenna gain, even when the packet arrives from an unknown source. This increases the range for the RTS packet, unlike directional beam systems.

Although the hidden node problem still exists when using an adaptive array antenna, the adaptive array's increased ability to suppress the interference results in only the loss of the interfering packet - the other packet is still correctly received. Even the association problem is reduced somewhat since adaptive array can beamform when receiving the beacon to provide multipath mitigation that is not present in the directional beam system. Furthermore, the only information needed for radio resource management is the number of antennas at the clients and access points (as this determines the gain and interference suppression capability for each device) and the signal strengths at each device, which results in much less complicated radio resource management techniques.
Effectively Combining Smart Antennas and Wireless Ad Hoc Networks

With adaptive array antennas being used in many of the latest generation of WLAN clients and access points today, it makes sense to assume that most future ad hoc networks will have smart antennas in them. The ability of adaptive arrays to suppress interference, allowing for spatial reuse in the adjacent cell and to use spatial multiplexing with MIMO (as being proposed for the 802.11n standard) that permits multiple channels in the same frequency, can provide substantial benefits. Theoretically, adaptive arrays in ad hoc networks can increase the network capacity by a factor of 2, 4, or even 8 times that without smart antennas, with only 2 or 4 antennas at the access point. But achieving that takes careful planning, as discussed below.

It is easy to simply add a smart antenna to an existing ad hoc network-but there is no guarantee that this will be successful. For instance, take a network that uses MIMO with spatial multiplexing. Using smart antennas for both transmission and reception can increase the range and capacity of a given link, but not necessarily of the network as a whole.

In Figure 5, an 802.11 b/g WLAN network is illustrated which has three non-overlapping channels-C1, C2 and C3, and all four nodes fully connected. If MIMO with spatial multiplexing is added to nodes 1 and 2 (as is being considered by the IEEE for the forthcoming 802.11n standard), the goal would be to double the data rate, but in reality, interference issues could arise, making the capacity actually lower as described below.

In particular, with MIMO with spatial multiplexing the tolerance of adjacent channel interference can be significantly reduced. Thus, with MIMO with spatial multiplexing being used for the link from node 1 to node 2, the links between nodes 2 and 4, and nodes 2 and 3 may be unable to use channels C2 and C3 without the creation of too high a level of adjacent channel interference, making all of the links share one channel, and therefore decreasing the network capacity to less than that without MIMO.

Figure 5: The effect of smart antennas on network capacity

Thus, the MAC/routing techniques need to accommodate smart antenna capabilities to achieve their full potential. To start two features are needed in the standards: 1) the appropriate information needs to be available in the network, and 2) hooks need to be present for smart antennas.

Specifically, the information needed for adaptive arrays are the number of antennas at each node, the number of interferers that can be suppressed by each node, and the number of spatially multiplexed channels that can be supported by each node (to be obtained when the access points and clients join the network), as well as measurements of the signal strengths at each node. Hooks are needed for the frequency assignment techniques to include reusing a frequency. Hooks are also needed to allow the system to include multiple radios in the same channel.

By ensuring that the MAC/routing algorithms accommodate smart antennas as described above, wireless ad hoc networks can still have most of the improvements of smart antennas without a significant increase in complexity of the MAC/routing algorithms and without significant changes to the algorithms.
Conclusion

Smart antennas, if used correctly, can increase the range and performance of wireless ad hoc networks like no other technique. In order to realize these benefits, however, the MAC/routing algorithms in wireless ad hoc networks need to be modified to allow adaptive array smart antennas to provide increased capacity, and should do so without significantly increasing these algorithms' complexity.

As consumer and business adoption of wireless technologies continues to grow, soon too will the usage of wireless ad hoc networks in home and business networking. By taking advantage of the improvements that can come from combining smart antennas with these networks, soon both the consumer and business user will not have to worry about range, performance or interference issues when setting up a wireless network.