ADAPTIVE CHANNEL ASSIGNMENT FOR ISM AND UNLIZENZIERTE FREQUENCY BANDS

The present invention relates generally to a so-called Bluetooth communications system operating at radio frequencies around 2.45GHz and, more particularly, to the allocation of an adaptive transmission channel in a piconet operating in the Bluetooth radio frequency band.

A Bluetooth system provides a communication channel between two electronic devices via a short-range radio link. In particular, the Bluetooth system operates in the radio frequency range around 2.4GHz in the unlicensed Industrial-Scientific-Medical (ISM) band. The Bluetooth radio link is intended to be a cable replacement between portable and/or fixed electronic devices. The portable devices include mobile phones, communicators, audio headsets, laptop computers, other GEOS-base or palm OS-based devices and devices with different operating systems.

The Bluetooth operating frequency is globally available, but the permissible bandwidth of the Bluetooth band and the available RF channels may be different from one country to another. Globally, the Bluetooth operating frequency falls within the 2400MHz to 2497MHz range. In the U.S. and in Europe, a band of 83.7MHz bandwidth is available and the band is divided into 79 RF channels spaced 1 MHz apart. Bluetooth network arrangements can be either point-to-point or point-to-multipoint to provide connection links among a plurality of electronic devices. Two to eight devices can be operatively connected into a piconet, wherein, at a given period, one of the devices serves as the master while the others are the slaves. Several piconets may form a larger communications network known as a scattemet, with each piconet maintaining its independence. The baseband protocol for a Bluetooth system combines circuit and packet switching. Circuit switching can be either asynchronous or synchronous. Up to three synchronous data (logical) channels, or one synchronous and one asynchronous data channel, can be supported on one physical channel. Each synchronous channel can support a 64 Kb/s transfer rate while an asynchronous channel can transmit up to 721 Kb/s in one direction and 57.6 Kb/s in the opposite direction. If the link is symmetric, the transfer rate in the asynchronous channel can support 432.6 Kb/s. A typical Bluetooth system consists of a radio link, a link control unit and a support unit for link management and host terminal interface functions. The Bluetooth link controller carries out the baseband protocols and other low-level routines. Link layer messages for link set-up and control are defined in the Link Manager Protocol (LMP). In order to overcome the problems of radio noise interference and signal fading, frequency hopping is currently used to make the connections robust.

Currently, each of the 79 RF channels is utilized by a pseudo-random hopping sequence through the Bluetooth bandwidth. The hopping sequence is unique for each piconet and is determined by the Bluetooth device address of the master whose clock is used to determine the phase of the hopping sequence. The channel is divided into time slots of 625µs in length and numbered according to the master clock, wherein each time slot corresponds to an RF hop frequency and wherein each consecutive hop corresponds to a different RF hop frequency. The nominal hop rate is 1600 hops/s. All Bluetooth devices participating in the piconet are time and hop synchronized to the channel. The slot numbering ranges from 0 to 2²⁷ -1 and is cyclic with a cycle length of 2²⁷. In the time slots, master and slave devices can transmit packets. Packets transmitted by the master or the slave device may extend up to five time slots. The RF hop frequency remains fixed for the duration of packet transmission.

The ISM frequency bands can be used by many different devices which include wireless local area networks (WLANs), microwave ovens, and lighting equipment. The interference caused by these multiple different applications is inherent to almost any device which is connected to the piconet. Currently, the usage of ISM frequency bands is growing very fast. In order to survive in these frequency bands, new wireless communication systems must utilize a robust modulation scheme with a certain method of channel allocation. For example, WLAN systems are using a Frequency Hopping Spread Spectrum (FHSS) method, in which transmission takes place only a short time in each channel, and Direct Sequence Spread Spectrum (DSSS) modulation, which overcomes narrow-band interference by spreading. However, in these systems the allocation of channels, or channelization, is organized by using either a carrier sensing (CS) method or a Code Division Multiple Access (CDMA) method. In the CS method, each of the channels which are to be used is measured in order to determine whether a transmission is taking place in that channel. If the channel under measurement does not have an ongoing transmission, then the channel can be used for hopping. The major problem with the carrier sensing method is that the measurement is ineffective for the traffic type that uses a different modulation method. In the CDMA method, while the narrow-band interferer is spread in the receiver, the received noise is actually increased, thereby reducing the noise margin of the system. Optionally, it is also possible to establish virtual traffic channels by using different hopping frequencies. However, this does not avoid the parts of the spectrum where the interference occurs.

It is advantageous and desirable to provide a method and system for making connections between devices operating in the ISM bands by effectively avoiding the parts of the spectrum where channel conditions such as interference and noise levels may adversely affect the channel connection.

The primary objective of the present invention is to provide a method and system to ensure the backward compatibility of a piconet device which is capable of operating in the non-frequency-hopping fashion (BT 2.0) in an environment where the frequency-hopping fashion (BT 1.0) is also used. The backward compatibility ensures that a BT 2.0 device is compatible with a BT 1.0 device.

Accordingly, the present invention provides a method for establishing a connection link between a master device and a plurality of slave devices in a communications network having a plurality of frequency channels within a radio frequency band, wherein the connection links between the master device and the slave devices are capable of being carried out in a frequency-hopping fashion. The method comprises the steps of:

• sending a link request to the master device requesting establishment of a non-frequency-hopping connection link between the master device and a slave device;
• establishing the non-frequency-hopping connection link as requested if the master device is able to select a communication channel for said non-frequency-hopping connection link; and
• establishing or maintaining the connection link in the frequency-hopping fashion if the master device is unable to select the communication channel for said non-frequency-hopping connection link.

Preferably, the method further comprises the step of measuring channel conditions including the carrier power of the channel and the interference and noise levels affecting the connection link in order for the master device to select the communication channel for the non-frequency-hopping connection link. The measurement of channel conditions is carried out by the master device or the requesting slave device.

Preferably, the method also includes the step of sending to the requesting slave devices a plurality of measurement parameters including measurement time and frequencies to be measured in order for the slave device to measure the channel conditions based on the measurement parameters.

Preferably, the method also includes the step of sending a measurement report to the master device by the slave device reporting results of the channel condition measurements.

Upon establishing the non-frequency-hopping connection link with the slave device, the master device can give up or retain its role as a master device to the non-requesting slave devices.

The present invention also provides a system for the adaptive allocation of transmission channels in order to establishing a connection link between a master device and at least one slave device in a communications network having a plurality of frequency channels within a radio frequency band, wherein the connection link between the master device and the slave device is capable of being carried out in a frequency-hopping fashion. The system comprises:

• a mechanism for the slave device to request the master device to allocate a channel for a connection link in a non-frequency hopping fashion;
• a mechanism for the master to determine whether it is able to allocate the requested channel;
• a mechanism to establish the non-frequency-hopping connection link between the master device and the requesting slave device on the allocated channel if the master is able to allocate the requested channel; and
• a mechanism to establish or maintain a frequency-hopping connection link between the master device and the requesting slave device if the master is unable to allocate the requested channel.

Preferably, the slave device maintains the frequency-hopping connection link if the slave device fails to receive a response from the master device responding to the request.

The present invention will become apparent taken in conjunction with Figures 1 a to 15.

• Figure la is a diagrammatic representation illustrating the establishment procedure of a connection link in a piconet wherein a slave device sends a request to the master device requesting a BT 2.0 connection link.
• Figure 1b is a diagrammatic representation illustrating that the master device responds to the requesting slave device, asking the slave device to connect channel measurements.
• Figure 1c is a diagrammatic representation illustrating that the slave device sends a measurement report to the master device.
• Figure 1 d is a diagrammatic representation illustrating that the master device sends a plurality of channel parameters to the slave device.
• Figure 1 e is a diagrammatic representation illustrating that the slave device acknowledges receipt of the channel parameters.
• Figure 1f is a diagrammatic representation illustrating that the master device stops being the master device of the non-requesting slave devices.
• Figure 1g is a diagrammatic representation illustrating the establishment of a BT 2.0 connection link between the former master device and the requesting slave device.
• Figure 2 is a frame structure illustrating an exemplary PDU for a slave device to request a BT 2.0 connection link with a master device.
• Figure 3 is a frame structure illustrating an exemplary PDU format used as an LMP_not_accepted response.
• Figure 4 is a frame structure illustrating an exemplary PDU format used as an LMP_accepted _start response.
• Figure 5 is a frame structure illustrating an exemplary PDU format used as an LMP_accepted_establish response.
• Figure 6 is a frame structure illustrating an exemplary PDU format used as an LMP_measurement_report response.
• Figure 7a illustrates a possible signaling sequence in establishing a BT 2.0 connection link.
• Figure 7b illustrates another possible signal sequence in establishing a BT 2.0 connection link.
• Figures 8a and 8b are flow charts illustrating an exemplary state diagram of a slave device requesting a BT 2.0 connection link.
• Figures 9a and 9b are flow charts illustrating an exemplary state diagram of a master device responding to a request for establishing a BT 2.0 connection link.
• Figure 10 is a diagrammatic representation illustrating the selection of channel measurement frequencies.
• Figures 1 la and 11 b are diagrammatic representations illustrating a hopping sequence example for packets that occupy 5 time slots.
• Figures 12a and 12b are diagrammatic representations illustrating a hopping sequence example for packets that occupy 3 time slots.
• Figure 13 is a diagrammatic representation illustrating an example of an RSSI dynamic range.
• Figure 14 shows an example of channel windowing.
• Figure 15 is a block diagram illustrating a system for the adaptive allocation of transmission channels.

Figures 1a through 1g are diagrammatic representations illustrating the establishment procedure of a connection link in a piconet 10 having a plurality of devices M, S1, S2 and S3 which are capable of being connected in a frequency-hopping fashion. The frequency-hopping connection links are well known in the art, and such a connection is referred to herein as a BT 1.0 connection link, associated with the Bluetooth Specification Version 1.0 (BT 1.0). As shown, M is currently a master device and S1, S2 and S3 are slave devices. The procedure described here is limited to the case where a slave device wishes to establish a connection link with the master device M in a non-frequency-hopping fashion. The non-frequency-hopping fashion is herein referred to as BT 2.0. As shown in Figure 1a, the connection links 102, 104 and 106 between the master device M and the slave devices S1, S2 and S3 are initially established according to the BT 1.0 fashion. At any time, either one of the slave devices S1, S2 and S3 can send a request to the master device M requesting a BT 2.0 link setup. For illustrative purposes, in the initialization phase the slave device S2 is the initiating unit which wishes to set up a BT 2.0 connection link with the master device M. As shown in Figure 1a, the slave device S2 sends a request 200 to the master device M requesting a BT 2.0 connection link. For example, the request can be sent in the form of an LMP PDU, as shown in Figure 2. Upon receiving the request, the master device M may respond to the request with three different PDUs, as listed in Table 1.

1. a) an LMP_not_accepted PDU (see Figure 3), if the master is unable to support this non-frequency-hopping connection link; or
2. b) an LMP_accepted_start PDU (see Figure 4) or an LMP_accepted_establish PDU (see Figure 5), if the master is able to support this frequency-hopping connection link.

If the master device M responds with an LMP_accepted_start PDU 202, as shown in Figure 1b, the master device provides a plurality of measurement parameters to the requesting slave device S2 for channel measurements. The LMP_accepted_start PDU 202 contains, for example, the measurement time and frequencies to be measured. During channel measurements, the master device M and the slave device S2 measure the carrier power C and/or the interference and noise levels I+N (denoted as I hereafter). During this time, the master M can support other BT 1.0 traffic in the piconet 10. The measurement of the system frequency band is carried out by scanning through the band at each channel following the hopping pattern of the master device M. The C measurement is carried out during a master-to-slave time slot. Preferably, the carrier power C at each channel is determined by the slave device using the Received Signal Strength Indication (RSSI) functionality of the receiver of the measuring slave device. The I level measurement is carried out during a slave-to-master time slot which is transmitted by another slave device (i.e., not the requesting slave device S2). In order to avoid measuring the slave-to-master transmission itself or its spectral leakages, an appropriate frequency offset between the slave-to-master frequency channel and the frequency to be measured has to be used. The frequency offset is described below in conjunction with Figures 10 through 12b in more detail. After the scanning time as defined by the master device M is over, the slave device S2 conveys a measurement report 204 to the master device M, as shown in Figure 1c. For example, the slave device S2 returns the measurement results in an LMP_measurement_report PDU, as shown in Figure 6.

It should be noted that it is also possible for the master device M to conduct channel measurements. In that case, the procedural steps as described in Figures 1b and 1c can be omitted.

Based on the measurement results, the master device M-selects a non-hopping channel for the BT 2.0 connection link and sends the channel parameters in an LMP_accepted_establish PDU 206 (see Figure 5) to the slave device S2, as shown in Figure 1d. It can be followed by the slave acknowledging receipt of the LMP_accepted_establish with an ACK signal 208, as shown in Figure 1e. At this point, the master device M establishes a master-slave switch operation 118 by delegating one of the non-requesting slave devices, S3 for example, as the new master device in order to maintain the BT 1.0 connection link between the non-requesting slave devices S1 and S3, as shown in Figure 1f. At the same time, the master device M starts a BT 2.0 transmission with the slave device S2 by sending certain data frames 208 at fixed intervals until the slave device S2 acknowledges that frame, for example. Finally, the master device M gives up its master role to become a BT 2.0 terminal T2 in order to establish the communication link in the non-frequency-hopping fashion with the slave device S2 which is now a BT 2.0 terminal T1, as shown in Figure 1g. The BT 2.0 communication link is denoted by numeral 212. The new connection link between the slave devices S1 and S3 is a BT 1.0 link, as denoted by numeral 120. Thus, the backward compatibility of the master device M and the slave device S2 makes it possible for these devices to operate in either BT 2.0 fashion or BT 1.0 fashion.

It should be noted that it is also possible that the master device M still maintains the role of the master device for the non-requesting slave devices S1 and S3 in the BT 1.0 link while simultaneously having the BT 2.0 link with the slave device S2.

It is likely that the channel conditions regarding carrier power C and/or interference and noise I conditions change during the data transfer between terminals T1 and T2. Thus, the selected frequency used for the current non-hopping channel may no longer be the best frequency for data transmission in the BT 2.0 connection link. To monitor the change in channel conditions, terminals T1 and T2 can be adapted to monitor propagation characteristics and data flow quality in the used frequency channel. For example, the monitoring may include continuous averaging of RSSI, transmission power, average packet error rate, average bit error rate, used modulation/coding and data packet memory monitoring. These values are compared to radio quality of service (QoS) parameters which are used as thresholds. If a threshold is not met, another frequency is selected for the new non-hopping channel. Among the BT 2.0 terminals (T1 and T2 in this illustrative example) some are empowered to make a decision regarding the frequency to be used in the new BT 2.0 connection link while some are not. Thus, the non-decision-making terminals must report the threshold failure to the empowered terminals. In particular, a specific PDU, LMP_radioQoS_failure, can be used to report the threshold failure. This PDU may indicate which radio QoS criterion or criteria are not met and the current RSSI value, packet error rate, etc. The PDU can be used to report:

1. a) whether the mean RSSI is above or below a certain threshold;
2. b) whether the packet error rate exceeds a certain threshold;
3. c) whether the transmission power exceeds a certain threshold; and
4. d) whether the used modulation/code belongs to a feasible set of modulation/coding schemes.

When it is required to use another frequency for maintaining the BT 2.0 connection link, the terminal empowered to make the decision regarding the frequency to be used in BT 2.0 connection links has three options:

1. 1) it may decide to stay on the selected frequency that is currently used for the BT 2.0 connection link, and use link adaptation and/or power control to improve the data flow quality. If transmissions are not continuous but repeated periodically, re-timing may be considered;
2. 2) it may start a new measurement process in order to select a new frequency for the new non-hopping channel; or
3. 3) it may allocate a new frequency for the new non-hopping channel based on the previous channel measurement results. For example, it could pick the second best frequency in terms of low interference and noise level in the previous channel measurement results (take Figure 14 for example, where f₂ is the best frequency, and f₁ is the second best frequency).

Selection of the proper action in terms of the above alternatives may include two phases. In the first phase it is determined whether degradation in the radio QoS is caused by insufficient RSSI or due to interference. This can be carried out by comparing RSSI values, packet error rates and used modulation/coding methods. If the cause is interference (i.e., RSSI is sufficient for the used modulation/coding but packet error is high), then a new channel measurement process or a new frequency allocation based on the previous measurement can be carried out. If the cause is insufficient RSSI, then Option 1, as described above, should be selected. The second phase is necessary only if the interference is the cause for the radio QoS degradation. In the second phase, Option 2 should be selected if the involved devices are non-delay sensitive, while Option 3 should be selected if the involved devices are delay sensitive.

Figures 2 to 6 are examples of LMP PDU formats. Figure 2 represents a bit level description of LMP_BT2. 0_req PDU prior to cyclic redundancy check (CRC) and encoding. As shown in Figure 2, Opcode 56 in the payload area is used to indicate that the requested connection link is in accordance with the BT 2.0 fashion.

As shown in Figure 3, the LMP_not_accepted PDU contains the Opcode 56 in the payload area to indicate that the response is related to the requested BT 2.0 connection link. The payload area may contain a reason why the master is unable to support the BT 2.0 link (Utisupport_LMP_feature).

As shown in Figure 4, the LMP_accepted_start PDU contains the Opcode 56 in the payload area to indicate that the response is related to the requested BT 2.0 connection link. The payload area also contains measurement parameters for channel measurements. As shown in Figure 4, the measurement parameters include the scanning time for the slave device to measure the channel conditions at each channel (Measurement_time).

As shown in Figure 5, the LMP_accepted_establish PDU may include link establishment parameters such as the frequency (Used_frequency) to be used for the BT 2.0 connection link, Modular Code Rate (MCR) and QoS parameters. The QoS parameter set also includes radio QoS parameter thresholds. The QoS parameters may include min_mean_RSSI, max_mean_RSSI, max_ packet_error_rate, max_Tx_power, min_ Tx_power, and set_of Jeasible_modulation/coding rates.

As shown in Figure 6, the LMP_measurement_report PDU may include the measured carrier power C value (C_Value) and the interference and noise I levels (I_Value) in a plurality of measured channels (Measurement_freq).

In the course of establishing a BT 2.0 connection link at the request of the slave device, the possible signaling sequences between a requesting slave device and the master device are shown in Figures 7a and 7b. In Figure 7a, originally the slave device and the master device are linked according to the BT 1.0 fashion, as denoted by numeral 100. In the initialization phase, the slave sends an LMP_BT2.0_ req PDU 200 to the master device, requesting the establishment of a BT 2.0 link. If the master is unable to support the BT 2.0 link for any reason, it responds to the request by sending an LMP_not_accepted PDU 201 to the requesting slave, stating the reason for not supporting the BT 2.0 link. For example, the reason for not supporting the BT 2.0 link may include that the data flow quality is currently below the radio QoS requirements. Accordingly, the BT 1.0 link between the slave device and the master device is maintained, as denoted by numeral 100'. It is possible that when the master device does not even know anything about the BT 2.0 connection link and fails to respond to the request 200, the slave device should not wait indefinitely for a response from the master device but maintain the BT 1.0 connection link after a set waiting period (see Figure 8a, step 317). At a later time, the slave device sends another LMP_BT2.0_ req PDU 200' to the master device, again requesting the establishment of a BT 2.0 link. If the master is able to support the BT 2.0 link and it has selected a frequency for the BT 2.0 link, it responds to the request by sending an LMP_accepted_establish PDU 206 to the requesting slave device, including the selected frequency, MCR and the required QoS parameters. Subsequently, a BT 2.0 link is established between the master device and the requesting slave as indicated by numeral 220. However, the master must give up its master role and become a BT 2.0 terminal, as shown in Figure 1g.

Another possible signal sequence is shown in Figure 7b. As shown in Figure 7b, upon receiving a request 200" from the slave device requesting the establishment of a BT 2.0 link, the master device sends the requesting slave device an LMP_accepted_start PDU 202 including the frequencies to be measured in order to establish a non-frequency-hopping link. The slave device measures the carrier power C and/or the interference and noise conditions I as indicated by numeral 190 and reports to the master the measurement results in an LMP_measurement_report PDU 204. Based on the measurement results, the master selects a frequency for the BT 2.0 link. The master sends an LMP_accepted_establish PDU 206' to the requesting slave device, including the selected frequency, MCR and the required QoS parameters. Subsequently, a BT 2.0 link is established between the master device and the requesting slave as indicated by numeral 220'. Because LMP PDUs are sent over an asynchronous connection-less (ACL) link, all packets are acknowledged in the Link Control level. Hence, a separate acknowledge signal ACK in the Link Management level is not required.

Figures 8a and 8b are flow charts illustrating a sequence of steps executed by a requesting slave device. As shown in Figure 8a, initially the slave device is connected with a master device in a BT 1.0 fashion, as indicated by numeral 310. As the slave device wishes to establish a BT 2.0 link with the master, it starts out by initializing a BT 2.0 link setup message from its upper layer at step 312 and sends an LMP_BT2. 0_req PDU to the master device at step 314. It waits for a response from the master at step 316. It is possible that the master device fails to respond to the request for a certain reason and the slave device will not receive a response from the master. Preferably, the slave device sets a time for receiving such a response. As shown at step 317, if the slave device does not receive the response from the master device after the set time has expired, it indicates the request failure to the upper level at step 320. If the set time has not expired, the slave device keeps waiting until it receives a response at step 318. There are three possibilities regarding the response from the master device: a) the response is an LMP_not_accepted PDU; b) the response is an LMP_accepted_establish PDU; or c) the response is an LMP_accepted_start PDU. If possibility (a) occurs, the slave device indicates the request failure to the upper level at step 320. The BT 1.0 link between the slave and the master is maintained or re-established, as indicated by numeral 322. If possibility (b) occurs, the slave device establishes the BT 2.0 connection link according to the frequency selected by the master device at step 324 and indicates the BT 2.0 connection link to the upper layer at step 326. The BT 2.0 link between the slave device and the master device is maintained as long as it is required, as indicated by numeral 328. If possibility (c) occurs, the slave device carries out the channel measurement procedure, as shown in Figure 8b.

As shown in Figure 8b, the slave device measures channel conditions at step 330 and sends measurement results to the master channel at step 332. The slave device must wait for a response from the master device at step 334 in order to take the next course of action. There are two possibilities regarding the response from the master device: a) the response is an LMP_not_accepted PDU; or b) the response is an LMP_accepted_establisch PDU. If possibility (a) occurs, the slave device indicates the request failure to the upper level at step 340. The BT 1.0 link between the slave and the master is maintained or re-established, as indicated by numeral 342. If possibility (b) occurs, the slave device establishes the BT 2.0 connection link according to the frequency selected by the master device at step 344 and indicates the BT 2.0 connection link to the upper layer at step 346. The BT 2.0 link between the slave device and the master device is maintained as long as it is feasible, as indicated by numeral 348.

Figures 9a and 9b are flow charts illustrating a sequence of steps executed by a master device. As shown in Figure 9a, initially the master device is connected with a slave device in a BT 1.0 fashion, as indicated by numeral 360. Upon receiving an LMP_BT2.0_req PDU from a slave channel requesting to establish a BT 2.0 connection link at step 362, the master device determines whether it can support the BT 2.0 connection link and how to respond to the slave device at step 364. There are three possibilities regarding the response to be sent to the requesting slave device at step 366: a) the response is an LMP_not_accepted PDU indicating that the master device is unable to support a BT 2.0 connection link, at least for the time being ; b) the response is an LMP_accepted_establish PDU; and c) the response is an LMP_accepted_start PDU. If possibility (a) occurs, the BT 1.0 link between the slave and the master is maintained or re-established, as indicated by numeral 368. If possibility (b) occurs, the master device provides link establishment parameters to the requesting slave device at step 370 and indicates the BT 2.0 connection link to the upper layer at step 372. The BT 2.0 link between the slave device and the master device is maintained as long as it is feasible, as indicated by numeral 374. If possibility (c) occurs, the master device provides the requesting slave device with measurement parameters for carrying out the channel measurement procedure, and the process continues in Figure 9b.

As shown in Figure 9b, after sending out the LMP_accepted_start PDU to the requesting slave channel, the master device waits for the measurement results as contained in an LMP_measurement_report PDU from the requesting slave device at step 380. Based on the measurement results, the master must decide the next course of action at step 382. There are two possibilities regarding the decision made by the master device at step 384: a) the master sends an LMP_not_accepted PDU to the slave device to indicate that it is unable to support the requested BT 2.0 connection link, based on the channel conditions measured by the requesting slave device; or b) the master sends an LMP_accepted_establish PDU to provide link establishment parameters to the requesting slave device. If possibility (a) occurs, the BT 1.0 link between the slave and the master is maintained or re-established, as indicated by numeral 386. If possibility (b) occurs, the BT 2.0 connection link is established at step 388 and the upper level is notified of the BT 2.0 connection link at step 390. The BT 2.0 link between the slave device and the master device is maintained as long as it is feasible, as indicated by numeral 392.

It should be noted that Figures 8a through 9b illustrate the flow charts of the a slave device and a master device when the establishment of the BT 2.0 connection link is requested by a slave device. In a similar manner, the master device can initiate a BT 2.0 connection link with any slave device in the piconet.

As described in conjunction with Figure 1b, when the requesting slave device S2 carries out the I measurement, it avoids measuring the slave-to-master transmission itself and/or its spectral leakage. Accordingly, an appropriate frequency offset between the slave-to-master frequency channel and the frequency to be measured is used. Preferably, the frequency offset value is high enough so that the transmitted power leakage over the adjacent channels does not significantly affect the measurement results. The exemplary channel measurement frequencies are shown in Figure 10. As shown, the odd-numbered time slots are master-to-slave slots in which the carrier power C measurements are made, and the even-numbered time slots are slave-to-master slots in which the interference and noise I levels are measured. It should be noted that the channel that is used for I measurement in each slave-to-master slot is offset by 4 channels from the slave-to-master frequency in the current hopping sequence. Figure 10 illustrates a possible way to select the I measurement frequency during a slave-to-master slot for packet transmission over one-slot frames.

In multi-slot packet transmission, a special offset calculation is used to prevent measuring slave-to-master slots as an I measurement channel. Figures 11a and 11b illustrate a hopping sequence for packets that occupy 5 time slots. In Figure 11a, the frequency of the master-to-slave slots is f₁, while the frequency of the slave-to-master slot is f₆. It is possible, for example, to use f_b = f₆ ± 4 as the measurement frequency which is different from both f₆ and f₁. Likewise, in Figure 11b, the frequency of the master-to-slave slot is f₁ while the frequency of the slave-to-master slots is f₂. It is possible, for example, to use f_b = f₂ ± 4 as the measurement frequency which is different from both f₂ and f₁.

Figures 12a and 12b illustrate a hopping sequence for packets that occupy 3 time slots. In Figure 12a, the frequency of the first master-to-slave slots is f₁ while the frequency of the subsequent slave-to-master slot is f₄. It is possible, for example, to use f_b = f₄ ± 4 as the measurement frequency which is different from both f₄ and f₁. Likewise, in Figure 12b, the frequency of the first master-to-slave slot is f₁ while the frequency of the subsequent slave-to-master slots is f₂. It is possible, for example, to use f_b = f₂ ± 4 as the measurement frequency which is different from both f₂ and f₁. However, the situation can be more complex. Let f_a be the first possible frequency of a multi-slot packet and f_c be the current hopping frequency, and the frequency of the I measurement channel be f_b which is 10MHz from the current hopping frequency. The 10MHz frequency offset is to ensure that the image frequency of the receiver does not coincide with the actual frequency, because the limited rejection at the image frequency may affect the measurement results.

Within the 79 available frequency channels of the ISM band, if 10<|f_b - f_a|<69, then we can use f_b = f_c + 10. Otherwise, the possible value for f_b is determined from the following equation: fb=gfc⁢fa⁢fb where gfc⁢fa⁢fb=fc-10-79⌊(fc-10)/79⌋,∀i⁢|fbi-fai|<10∨fbi-fai>69

As described earlier, the preferred measurement resolution is 1 MHz. After the channel measurements are completed, there are 79 C values and 79 I values, with one C and one I value for each frequency channel. These values are normally averaged over a certain amount of measured C and I values because the same channel might be measured a number of times. The averaging of the measurement results can be carried out during the measurement (continuous averaging) or after the measurement. The averaging procedure for the C value is shown below: Cf⁢79ave=1/N∑k=NN-1Cf⁢79k, where N is the number of measurements and the averaging is carried out over each of the 79 channels. If the averaging is carried out over the whole band, then Cfave=1/79∑I=0791/N∑k=0N-1Cfik, where N is the number of measurements on each of the 79 channels.

The I measurement results are averaged in a similar way. However, averaging over the whole band is not used. Averaging of the carrier power C over the whole band means that the whole band is not used. Averaging of the carrier power C over the whole band means that the selection of a best channel placement is based on the I measurement only. In this case C measurements are not required. This approach ignores fast fading which is actually desirable. Notches caused by fast fading are changing their locations quite swiftly if there are even slight changes in the propagation environment and, therefore, their locations should not be relied upon when the optimum channel placement is considered. Alternatively, it is possible to measure the I conditions because they probably give satisfactory results in a channel placement.

As a typical procedure, a number of measured C and I values from the same channels are parameterized, as this amount depends on the available measurement time and the connection initialization time requirements. For example, if it is required to make 10 measurements per channel, then the required time for measurement is given by 10x79x0.001250s = 0.98s. The accuracy of the measured C and I values is dependent on the receiver RSSI measurement accuracy. An example of a 64dB dynamic range of an RSSI measurement is illustrated in Figure 13.

Depending on the RSSI measurement resolution, the required amount of bits needed to present C and I values can be estimated. For example, if there is a 3dB resolution, the whole dynamic range of the RSSI measurement can be divided into 22 levels. Thus, a minimum of 5 bits is used so that all the levels can be presented. With the measured I values, it is possible to use only 4 bits of data because the I values above a certain level may not be worthy of being addressed. At those high levels, the interfering source may be too strong and make the C/I ratio too small for channel selection regardless of what the C value would normally be. The possible values for C and I measurement are given in Table 2.

Accordingly, the needed data packet size would be 9x79=711 bits. This packet size indicates that a DM3/DH3 ACL packet type is required. However, it is possible to organize measurement data such that one-slot packet types can be used in transmission. In practice, this signifies a data packet of 136-216 bits (DM1/DH1). In this case, the measurement data has to be sorted, for example, so that only the 9-12 lowest I values and the corresponding C values are reported, instead of all the measured C and I values. It should be noted that when the C and I information is assigned only to certain frequency channels, the associated frequency information must also be notified along with the reported C and I values. The 79 frequencies in the ISM need 7 bits of data to notify. An example of data packet format prior to data whitening and coding is illustrated in the LMP _measurement_report PDU, as shown in Figure 6.

A DH1 packet can contain up to 12 measured units including C, I and frequency values because no coding is utilized. A DM1 packet contains only 9 measured units because 2/3 coding is used. A summary of the reporting format is shown in Table 3. This reporting format can be defined by the master device with the LMP_accepted_start PDU.

The measurement results can be further processed by channel windowing so that it is possible to take into account the BT 2.0 channel width which might differ from the channel measurement resolution. The window for channel windowing can be, for example, a slide average window which is originally slid through the measurement data of 1MHz resolution. The width of the sliding window can be, for example, the same as the channel bandwidth of the BT 2.0 channels. An example of channel windowing which is used in channel measurements is shown in Figure 14. It is also possible to utilize different weighting for adjacent channels or the whole set of channels, if so desired. Because of channel selection filtering, interference in adjacent channels is usually not as significant as interference in the channels that are in use. In Figure 14, the I value as processed by channel windowing is denoted by si=∑k=0N-1If⁢i+k where N is the number of frequency channels over which channel windowing is carried out. With N=4, s₂ is the channel-windowing average value of I over f₂, f₃, f₄ and f₅, for example. As shown in Figure 14, s₀ has the lowest level of interference. Thus, any one of the channels f₀, f₁, f₂, and f₃ can be used for BT 2.0 transmission because s₀ is the sum of interference in those channels. For that reason, the sum of interference after channel 76 is not available.

Figure 15 is a block diagram illustrating a system 20 for the allocation of adaption transmission channels. As shown in Figure 15, the system 20 includes a plurality of mechanisms included in the electronic devices in a piconet. In particular, a slave device 30 includes a requesting mechanism 32 for sending a request 200 (see Figure 1a) to a master device 40, requesting the establishment of a BT 2.0 connection link. The master device includes a deciding mechanism 42 for determining whether it is able to support a BT 2.0 connection link, at least at the time of request. The slave device further includes a mechanism 34 for channel measurements, a mechanism 36 for processing the measurement results and reporting the measurement results to the master device. Preferably, the slave device also includes a mechanism 38 to recognize that the master device fails to respond to the request. Both the master device and the slave device also include a mechanism 50 for establishing a BT 2.0 or BT 1.0 connection link therebetween. As shown in Figure 15, other messages 230, such as the response 202 in Figure 1b and the response 204 in Figure 1c, can also be sent from one device to another.

Although the invention has been described with respect to a preferred embodiment thereof, it will be understood by those skilled in the art that the foregoing and various other changes, omissions and deviations in the form and detail thereof may be made without departing from the scope of this invention, as defined in the appended claims.