It is well known that various factors reduce the quality of signals transmitted between an access point and a client device in a network. One such performance-reducing factor is noise. Another is contention, ie the existence in the locality of access points of other networks which use the same frequency band for communication and therefore contend with the access point at issue for airtime.

One way of minimising the effect of such performance-reducing factors is through judicious selection of the operating channel that the network uses. This task is made more complicated by the fact that the magnitude of the performance-reducing factors can vary with time. It is desirable to find an improved method of channel selection for minimising reductions in signal quality.

According to a first aspect of the invention there is provided a method of selecting a channel for wireless communication in a WLAN, the method comprising

Communicating data over a first communication channel between an access point and a client device for a first time period,

During the first time period, obtaining a first performance measurement indicative of the performance of the first channel,

Communicating data over a second communication channel between the access point and the client device for a second time period,

During the second time period, obtaining a second performance measurement indicative of the performance of the second channel,

Ranking the first and second channels in order of performance taking the first and second performance measurements into account when doing so,

Selecting an operating channel using the outcome of the ranking step.

Some known WLANs having an access point with a single radio, check the performance of candidate operating channels by pausing communication on the current operating channel and performing a very brief performance measurement of each of the other channels. Our invention is advantageous over such systems because, for example, there is no need to pause communication in order to measure the performance of a channel. Furthermore, the invention enables a method to be performed in which performance data can collected over a longer period than in known systems, enabling a greater quantity of performance data to be collected.

The step of communicating data over the first channel may comprise sending and receiving data over the first channel. The step of communicating data over the second channel may comprise sending and receiving data over the second channel.

The step of sending and receiving data over the first channel and the step of sending and receiving data over the second channel may both be performed using a radio and are preferably performed using the same radio. The first time period may be equal to the second time period. The first time period may be longer than one minute and may be longer than one day. The first time period may be one week.

The second time period may begin immediately after the first time period ends. The method may comprise obtaining a plurality of first performance measurements during the first time period. The method may comprise obtaining a plurality of second performance measurements during the second time period. The step of ranking the first and second channels in order of performance may include using the plurality of first performance measurements to generate a first performance value for the first channel and may further include using the plurality of second performance measurements to generate a second performance value for the second channel. The ranking step may further include comparing the generated first performance value with the generated second performance value.

The method may further comprise dividing the first time period into sub-periods and, for each sub-period, calculating the average value of two or more first performance measurements which were obtained during that sub-period. The first performance value may be generated taking into account the calculated average values. The method may further comprise dividing the second time period into sub-periods and, for each sub period, calculating the average value of two or more second performance measurements which were obtained during that sub-period. The second performance value may be generated taking into account the calculated average values.

The calculated average values of the first and/or second performance measurements may be recorded in a data store. The generated first and/or second performance values may be generated using the most recently stored average first and/or second performance measurements. Therefore a new value of the first and/or second performance values may be generated every time a new average of the first and/or second performance measurements is stored in the data store. The generated first and/or second performance values may be stored in the data store. The data store may be located in the access point. The sub-periods may be of uniform duration and may be greater than or equal to 30 seconds and less than or equal 120 seconds and are preferably 60 seconds.

If the generated first performance value indicates that the performance of the first channel has fallen below an unworkability threshold, the first channel may be replaced as the operating channel before the end of the first time period and preferably immediately.

The one or more first and/or second performance measurements may include one or more measurements of the noise on the channel. Noise on the channel may include interference from sources including but not limited to microwave ovens and analogue TV senders. The one or more first and/or second performance measurements may include one or more measurements of the contention level on the channel. The contention level may be a measure of the amount of contending neighbouring Wi-Fi traffic that is operating on the channel. Preferably the one or more first and/or second performance measurements include one or more measurements of the noise on the channel and one or more measurements of the contention level on the channel.

The first performance value may be generated using a noise score which may be the percentage of averaged noise measurement values that are above a noise threshold in the first time period. Alternatively or in addition, the first performance value may be generated using a contention score which may be the percentage of averaged contention measurement values that are above a contention threshold in the first time period. Alternatively or in addition, the first performance value may be generated using a no measurement score which may be the percentage of the first time period for which measurements have not been made. The step of generating the first performance value may include determining which is the largest of the noise score, the contention score and the no measurement score. The first performance value may be calculated as“100 minus the largest of the noise score, the contention score and the no measurement score”.

The averaged first and/or second performance measurements may be recorded in a data store. The generated first and/or second performance values may be recorded in the data store. The data store may be located in the access point.

In preferred embodiments, in addition to the first and second channels there are one or more further communication channels. In these embodiments, the method may include sending and receiving data over the one or more further communication channels between the access point and the client device for a further time period. The method may further comprise obtaining, from the one or more further channels, one or more further performance measurements indicative of the performance of the one or more further channels. The method may further include ranking the first, second and one or more further channels in order of performance taking the first and second and one or more further performance measurements into account when doing so. Any of the above method steps relating to the first and/or second channels may also apply to the one or more further channels, mutatis mutandis. In some embodiments there is one further channel. In other embodiments there are 1 1 further channels. Other numbers of further channels are possible, depending on factors including the regulatory domain and operating band used. The first, second and one or more further channels may each use a range of frequencies. The frequencies may or may not overlap.

The step of selecting an operating channel using the outcome of the ranking step may involve excluding one or more of the lowest ranked channels from use as the operating channel. Preferably it is the single lowest ranked channel that is excluded from use as the operating channel. If the lowest ranked channel is the only channel available for use as the operating channel, it may not be excluded for use as the operating channel.

The method may include constructing a list of channels upon which the method is to be performed. The list of channels may be recorded in the data store. The performance value corresponding to each channel may be recorded in association with that channel in the list of channels. The ranking step may further comprise comparing the performance value of one or more channels with the performance value of a channel that has been excluded from use as an operating channel. This comparison may use the last recorded performance value for the channel that has been excluded from use as an operating channel. A channel that has been excluded from use as an operating channel may be made available for use again as an operating channel if, after the ranking step it is ranked higher than some, or preferably all of the other channels.

According to a second aspect of the invention there is provided an access point for a wireless communication network, the access point comprising

A transceiver adapted to communicate data between the access point and a client device over a first communication channel for a first time period,

The transceiver being further adapted to communicate data between the access point and the client device over a second communication channel for a second time period,

A performance measurer adapted to measure the performance of the first channel during the first time period and further adapted to measure the performance of the second channel during the second time period,

Channel ranking means adapted to rank the first and second channel using the measurements made by the performance measurer,

A channel selector adapted to use an output of the channel ranking means to select an operating channel for the access point.

The transceiver may be adapted to transmit and receive data between the access point and a client device over the first communication channel for the first time period. The transceiver may be adapted to transmit and receive data between the access point and a client device over the second communication channel for the second time period.

The transceiver may be a radio. The access point may be provided with a single radio. The performance measurer may be adapted to measure the noise level and contention level on the channel.

The features defined above in relation to the first aspect of the invention are also applicable to the second aspect of the invention.

A specific embodiment of the invention will now be described, for illustration only, with reference to the appended drawings, in which Figure 1 is a schematic representation of a wireless LAN for use in accordance with the invention;

Figure 2 is a schematic representation of wireless access point for use in accordance with the invention;

Figure 3 is a flow chart showing method steps in accordance with embodiments of the invention.

Figure 1 shows a wireless LAN which is indicated generally at 1 . The LAN 1 contains an access point 2 and several clients 3. The access point 2 contains a single radio 4 for communicating with the clients 3. The radio 4 is capable of operating on several different frequencies within the frequency band of operation of the access point 2. For example, if the 2.4GHz band is the band of operation, the radio 4 can operate on any of channels 1 -13 within that band of operation.

Figure 2 shows the access point in more detail. A channel selector 5 is provided which assembles a channel cycle list (not shown). The channel cycle list is a list of candidate channels which can be used by the access point 2. The channel cycle list contains channels 1 -13 within the 2.4GHz band of operation. A measurement period timer is provided (not shown). The channel selector 5 sets this timer to a measurement period of one week and starts the timer running. The channel selector 5 then causes the radio 4 to use the first channel in the channel cycle list as the working channel for communication with the clients 3, for a period of one week.

There is also provided a noise level collector 6 and a contention level collector 7.

During the period of one week that the first channel on the channel cycle list is used as the working channel, the noise level collector 6 and the contention level collector 7 measure the levels of noise and neighbour contention on the channel, respectively. By noise level it is meant the amount of interference (e.g. from microwave ovens, analogue TV senders etc) experienced by the working channel. Measuring the noise level involves measuring the signal to noise ratio associated with the channel. By contention level it is meant the amount of contending neighbouring Wi-Fi traffic that is operating on the working channel. Measuring the noise level involves measuring the number of other access points located in the vicinity of the access point 2 that are operating on the channel. The noise 6 and contention level 7 collectors determine the average noise and contention levels over the course of each minute and record the averaged value in a channel performance store 8. Therefore, for example, the noise level collector 6 records 60 different values of the noise level in the channel

performance store 8 each hour. Similarly, the contention level collector 7 records 60 different values of the contention level in the channel performance store each hour.

A channel score generator 9 is provided which generates a performance score for the working channel. It does this by determining the percentage of one minute intervals (where measurements exist) where noise level exceeds a predetermined noise threshold. This is referred to as the“noise score”. It also determines the percentage of one minute intervals (where measurements exist) where contention level exceeds a predetermined contention threshold. This is referred to as the“contention score”. It also determines the percentage of one minute intervals of the channel measurement period for which no measurements have yet been made. This is referred to as the“no measurement score”. The channel score generator 9 calculates the channel performance score using the formula:

Channel Performance Score = 100 - max (noise score, contention score, no measurement score) where“max (noise score, contention score, no measurement score)” means whichever is largest of the noise score, the contention score and the no measurement score.

The channel score generator 9 then stores the generated performance score in the performance store. Each time a new value for the noise and contention level is stored (ie every minute), the performance score generator 9 generates a new performance score for the working channel which takes into account the new values for the noise and contention levels. Therefore an updated performance score is generated and stored every minute.

When the measurement period timer indicates that the channel measurement period has elapsed, the channel selector 5 causes the working channel of the access point 2 to change to the next channel in the channel cycle list. As before, the noise and contention level collectors 6, 7 collect noise and contention levels on the new working channel, average them over the course of one minute and store the averaged values in the channel performance store 8. As before, the performance score generator 9 uses the noise and contention values to generate a performance score for the new working channel and updates the performance score each time new noise and contention values are stored. Once the measurement period expires, the working channel changes to the next channel in the channel cycle list and the process is repeated. This continues until all of the channels in the channel cycle list have been used as the working channel.

If, at any time, the channel performance score for the working channel falls below a predetermined unworkability threshold (ie if the performance of the channel is so poor as to be deemed unworkable), then the channel selector 5 causes the working channel to change to the next channel in the channel cycle list regardless of whether the channel measurement period has elapsed.

Once all of the channels have been used as the working channel, the channel selector 5 compares the stored performance scores for all of the channels. The channel with the lowest performance score is removed from the channel cycle list, unless it is the last remaining channel in the channel cycle list, it which case it is not removed from the channel cycle list. The channel selector 5 then causes the access point 2 to operate on the first channel remaining in the channel cycle list and the whole process restarts. Please note that when the channel selector 5 compares the stored performance scores for all of the channels, this includes the stored performance scores of channels that have previously been removed from the channel cycle list for having the lowest performance score on a particular cycle. If, when the performance scores of all the channels are compared, it transpires that a previously removed channel has a higher performance score than any channel presently on the channel cycle list, that previously removed channel is brought back into the channel cycle list.