Automatic channel selection in a radio access network

[0001] This invention relates to a technique for automatically selecting channels in a Radio Access Network, such as a wireless Local Area Network (LAN).

[0002] Advances in the field of wireless LAN technology has led to the availability of relatively inexpensive wireless LAN equipment, which, in turn, has resulted in the emergence of publicly accessible wireless LANs (e.g., “hot spots”) at rest stops, cafes, libraries and similar public facilities. Presently, wireless LANs offer users access to a private data network, such as a Corporate Intranet, or a public data network such as the Internet. The relatively low cost to implement and operate a wireless LAN, as well as the available high bandwidth (usually in excess of 10 Megabits/second) makes the wireless LAN an ideal access mechanism through which a mobile terminal user can exchange packets with an external source.

[0003] Within the wireless LAN, there exists one or more Access Points (APs) each typically comprising a transceiver for exchanging radio frequency signals with a mobile terminal user. Each AP communicates with the mobile terminal user on a particular channel. In most radio technologies, such as the radio technology embodied in the IEEE 802.11b standard, adjacent channels tend to overlap. For this reason, geographically adjacent APs within the wireless LAN try to assign non-overlapping channels to avoid interference. Often such channel assignments require manual intervention or specific protocol modifications. Unfortunately, this approach doesn't work well when geographically adjacent APs belong to wireless LANs operated by separate entities that lack the incentive to cooperate to reduce radio frequency interference.

[0004] Thus, there is need for a technique for automatically selecting channels in a wireless LAN to reduce interference without any modifications to the underlying communication protocol.

[0005] Briefly, in accordance with present principles, a method is provided for selecting a radio channel by an access point (AP) within a radio access network. The method commences upon the AP initially choosing a channel randomly from among a plurality of available radio channels. The AP then monitors the selected channel for a random interval to determine if any traffic is present. If the channel is presently free (i.e., no use by another AP), then the AP claims the selected channel to carry traffic. Otherwise, if the channel already carries traffic, the AP chooses another channel. Thereafter, the AP monitors the newly selected channel during a random interval to determine the availability of the newly chosen channel. In practice, the AP repeats the process until it finds an available channel found, or it scans all the available channels.

[0008] FIG. 1 depicts a block schematic diagram of a communications system 10 that includes at least one, and preferably, a plurality of radio access networks, illustratively depicted by radio access networks 11₁ and 11₂. The radio access networks 11₁ and 11₂ each enable at least one user, and preferably a plurality of users (e.g., users 12₁, 12₂, and 12₃) to access an external data network 14 such as the Internet or the like. In a preferred embodiment, the user 12₁ utilizes a lap top computer while the user 12₂ utilizes a Personal Data Assistant and the user 12₃ utilizes a wired communications appliance. Other users (not shown) could employ other types of wired or wireless communication appliances.

[0009] Each of the radio access networks 11₁ and 11₂ includes at least one, and preferably, a plurality of access points (APs), illustratively illustrated by APs 18₁-18₄, via which each of the users 12₁, 12₂ and 12₃ accesses a wireless Local Area Network (LAN) 20 within each access network. In the illustrated embodiment, each AP, such as AP 18₁, includes a wireless transceiver (not shown) for exchanging radio frequency signals with a radio transceiver (not shown) within a communications appliance employed by a user, such one of users 12₁ and 12₂. One or more of the APs 18₁-18₄ could also include a wired access mechanism by which a user, such as user 12₃, can access the network via a wired communications appliance. Each of the APs 18₁-18₄ in each of the radio access networks 11₁ and 11₂ employs one or more well-known wireless or wired data exchange protocols, such as the “HiperLan 2” or IEEE 802.11 protocol. Indeed, different APs can employ different wireless protocols to accommodate users whose communications appliances use different protocols.

[0010] Each of the access points 18₁-18₃ communicates with a corresponding one of users 12₁ and 12₂ over a radio frequency link by choosing a particular radio channel over which to send and receive data. With most radio technologies, such as the IEEE 802.11b standard, adjacent channels exhibit some degree of overlap. Thus, radio interference will occur when geographically adjacent APs transmit on adjacent channels. In the past, the problem of adjacent channel interference has been overcome by manual intervention, or the adoption of specific protocol modifications to force the selection of non-adjacent channels. These approaches incur the disadvantage of requiring either full-time administrative control over geographically adjacent APs or requiring them to implement specific protocol modifications. Typically, the ability to exert control over adjacent APs does not exist when such adjacent APs belong to different network entities.

[0011] FIG. 2 illustrates in flow chart form the steps of a method in accordance with the present principles for selecting a radio frequency channel for an AP, such one of the APs 18₁-18₄ within one of the access networks 11₁ and 11₂. The channel selection method of FIG. 2 commences upon execution of step 100 during which the AP sets the following parameters:

[0012]

	PARAMETER	DESCRIPTION
	current_channel	The current channel selected by the AP
	channel_min	The lowest channel number available
	channel_max	The highest channel number available
	channel_step	The separation between selected channels
	time_min	The minimum time for channel monitoring
	time_max	The maximum time for channel monitoring

[0013] Following step 100, the AP makes a check during step 102 to determine whether the last channel used by the AP still remains available. If so, then AP selects the last channel used as the current channel during step 104. Checking the availability of the previously used channel increases efficiency. Often times, the AP can re-use the last channel before another AP claims the channel. Should the AP find the last channel unavailable, then the AP will select a channel at random from the available channel list during step 106.

[0014] Following either of steps 104 and 106, the AP executes step 108 during which the AP sets the channel-monitoring interval scan_time to a random value lying between time_min and time_max. Thereafter, the AP monitors the channel during step 110. The AP monitors the channel by listening during the interval scan_time for the presence of traffic (if any) carried by another access point (AP). During step 112, the AP checks whether the monitoring performed during step 110 uncovered use of the channel by another AP. Upon finding the channel free of traffic during step 112, the AP claims the channel for use during step 114 and the channel selection process ends (step 116).

[0015] If the AP finds the channel in use during step 112, the AP then makes a check during step 118 to determine whether other channels remain available for selection, taking into account the separation between selected channels prescribed by the current value of channel_step. Thus, for example, if the channel_step=2, then after selecting channel #1, the channel #3 becomes the next channel available for selection. Depending on the value of the current_channel and the number of channels available, the AP could find that additional channels still remain available. If so, the AP executes step 120 selects a next new channel in accordance with the relationship:
new channel=current_channel+channel_step
before proceeding to re-execute step 108. Otherwise, upon detecting exhaustion of the available channels during step 118, the AP resets the value of current_channel during step 122 in accordance with the relationship:
channel_step=floor(channel_step/2)
In other words, during step 122, the AP resets the value of channel_step to the lowest integer value of one-half the previous value of channel_step. Thus, if the value of channel_step prior to step 122 was 3 for example, then following step 122, the new value would be one. After step 122, program execution branches to step 120.

[0016] To better understand the above-described channel selection method, consider the following three examples.

[0017] Assume that the APs 18₁ and 18₂ in the radio access network 11₁ of FIG. 1 use channels #1 and #6, respectively, while AP 18₄ remains off line. Further assume that AP 18₃ just commenced operation and is the only AP implementing the channel selection method of the present principles. In implementing the channel selection method, assume that AP 18₃ employs the following values: channel_min=1, channel_max=11, channel_step=5, time_min=500, and time_max=1000, with the time_min and time_max measured in milliseconds. Additionally, assume that AP 18₃ did not store any information about the previously used channel.

[0018] In accordance with the channel selection method of the present principles, the AP 18₃ will first select a current channel in accordance with the relationship
current_channel=random(1, 6, 11)

[0019] For present purposes, assume that AP 18₃ selected channel #6. Next, the AP 18₃ selects the value for scan_time in accordance with the relationship:
scan_time=random(500, 1000)

[0020] For present purposes, assume that AP 18₃ selected the value of scan_time as 600 milliseconds. Accordingly, the AP 18₃ will listen to channel #6 for that interval. Since the AP 18₂ is currently using channel #6 in this example, the AP 18₃ will detect use of this channel. Therefore, the AP 18₃ will select a new channel in accordance with the relationship:
new channel=current_channel+channel_step

[0021] In the present example, with current_channel=6 and channel_step=5, the AP 18₃ will select channel #11. The AP 18₃ now establishes a new value for the parameter scan_time, say 750 ms and then listens to channel #11 for that interval. Assuming that this channel carries no other traffic, the AP 18₃ will now claim channel #11 for use.

[0022] Assume that the APs 18₁, 18₂, and 18₃ within the radio access network 11₁ use channels #1, #6, and #11, respectively. Further assume that AP 18₄ in the same radio access network now commences operation and implements the channel selection method according to present principles. In implementing the channel selection method, assume that the AP 18₄ selects the following parameter values: channel_min=1, channel_max=11, channel_step=5, time_min=500, time_min=1000. Additionally assume that AP 18₄ did not store information about the previously used channel.

[0023] Using the channel selection method of the present principles, the AP 18₄ will first select a channel in accordance with the relationship
current_channel=random(1, 6, 11)

[0024] Assume that AP 18₄ selects channel #6 as in the previous example. Next, the AP 18₄ selects the value for the parameter scan_time in accordance with the relationship:
scan_time=random(500,1000)

[0025] Assume that AP 18₄ selected a value of 660 milliseconds for scan_time so AP 18₄ will listen to channel #6 for 660 ms for that interval. With channel #6 in use by AP 18₂ in this example, the AP 18₄ will find channel #6 occupied. Thus, the AP 18₄ selects a new channel in accordance with the relationship:

 new channel=current_channel+channel_step

[0026] In the current example, the AP 18₄ will now select channel #11.

[0027] The AP 18₄ next establishes a new value for scan_time, say 550 ms and then listens to channel #11 for that interval. Since channel #11 remains in use by AP 18₃ in this example, the AP 18₄ will find channel #11 occupied as well.

[0028] Having found channel #11 busy, the AP 18₄ selects another channel, and in this example, will now pick channel #1 due to wrap around. Having selected channel #1, the AP 18₄ selects a new value for scan_time, say to 800 milliseconds and will listen to channel #1 for that interval. Since channel #1 remains in use by AP 18₁ in this example, the AP 18₄ will find channel #1occupied as well.

[0029] All possible channels for selection have been exhausted in this example when channel_step=5. Therefore the AP 18₄ will reduce the value of the parameter channel_step in accordance with the relationship:
channel_step=└channel_step/2┘

[0030] In the current example, the new value of channel_step becomes 2. After reducing the value of channel_step to 2, the AP 18₄ now selects another channel using the previously described relationship. In this example the AP 18₄ now selects channel #3. Thereafter, the AP 18₄ selects a new value for scan_time, say 730 milliseconds and now listens to channel 3 during that interval. Upon finding channel #3 free of traffic, the AP 18₄ now claims this channel for use. Note that in this case, interference cannot be completely avoided because all interference-free channels are already in use.

[0031] This example describes scenario when two APs, such as APs 18₂ and 18₃ within the radio access network 11₁, both startup simultaneously, while another AP, such as AP 18₁ in the same network, currently uses channel #1. Assume that both APs 18₂ and AP 18₃ utilize the following parameter values: channel_min=1, channel_max=11, channel_step=5, time_min=500, time_min=1000. Also, assume that neither AP 18₂ nor AP 18₃ store information about the channel each previously used.

[0032] Using the method of the present principles, both of the APs 18₂ and 18₃ will each select a current channel as follows in accordance with the relationships:

[0033] For AP 18₂
current_channel=random(1, 6, 11)

[0034] Assume AP 18₂ has randomly selected channel #6.

[0035] For AP 18₃
current_channel=random(1,6,11)

[0036] Also assume AP 18₃ selects channel #6. If both AP 18₂ and AP 18₃ were to actually claim the same channel, severe interference would result. However, as will become apparent from the discussion hereinafter, the channel selection method of the present principles will prevent both of the APs 18₂ and 18₃ from claiming the same channel.

[0037] After each of the APs 18₂ and 18₃ initially selects a channel, each AP establishes a value for the parameter scan_time as follows:

For AP 18₂
scan_time=random(500,1000) (say 660)
For AP 18₃
scan_time=random(500,1000) (say 820)

[0038] Next, the AP 18₂ listens to channel #6 for 660 ms while AP 18₃ also listens to the same channel for 820 milliseconds. Assuming that no traffic otherwise exists, then AP 18₂ will find the channel free during the period it listened and thus claim the channel for use. Since the parameter scan_time established by AP 18₃ is longer than the parameter scan_time established by the AP 18₂, the AP 18₃ will listen longer and will ultimately detect use of channel #6 by the AP 18₂ during the time interval between 660 and 820 ms.

[0039] Having found channel #6 occupied, the AP 18₃ now must choose another channel and does so in the manner previously described. In the current example, assume AP 18₃ selects channel #11. The AP 18₃ now selects a new value for scan_time, say 530 milliseconds The AP 18₃ now listens to newly selected channel #11 for 530 ms, and upon finding it free, now claims the channel for use.

[0040] The foregoing describes a technique for selecting a radio channel by an access point in a radio access network to minimize adjacent channel interference.