Method for subcarrier allocation and for modulation scheme selection in a wireless multicarrier transmission system

The present invention relates to wireless communications systems, and more particularly, to a method for allocating subcarriers and selecting a modulation scheme per subcarrier in a high-speed wireless fixed or mobile multi-carrier system.

An approach to multi-carrier modulation is Orthogonal Frequency Division Multiplexing (OFDM). OFDM was for example standardized for the two high-speed digital radio transmission systems Digital Audio Broadcasting (DAB) and Digital Video Broadcasting Terrestrial transmission mode (DVB-T), which are used for transmitting digital radio and television signals. A further field of application, for which the OFDM transmission system plays an increasing role, is the mobile access to wire-bound networks with the aid of a local radio network, whereby high data rates shall also be transmitted. In this connection, the HIPERLAN/2 standard as well as an extension of the IEEE 802.11 a standard for the 5 Ghz area can be mentioned. In both systems OFDM transmission is applied. OFDM could also offer a sensible alternative for high-speed mobile applications, and thus represents an important step for next generation mobile radio systems or for a 4^th generation air interface.

In conventional multi-carrier schemes the transmitted data is split into a number of parallel data streams, each one used to modulate a separate subcarrier. If an OFDM transmission system is applied, the broadband radio channel is subdivided into a plurality of narrow-band subchannels or subcarriers being independently modulated with e.g. QPSK, 16 QAM, 64 QAM or higher modulation order allowing higher data rate per subcarrier. The higher modulation orders however can only be used if the signal to noise ratio (SNR) at the receiver is high enough to allow the demodulation. The subcarrier frequencies can be allocated to a user channel on a short term basis and the modulation order per subcarrier has to be selected to define a transmission channel for each user.

In a terrestrial mobile environment with multipath radio channels there is the possibility of a very strong channel attenuation of single subcarriers. This means that some subcarriers allocated to the user may be useless because of fading. In order to compensate the fading of some subcarriers, a known method, which is disclosed in document "<nplcit id="ncit0001" npl-type="s"><text>Improving performance of multi-user OFDM systems using bit-wise interleaver" (Electronic Letters, 13th September 2001, vol.37, No.19) by Z.Wang and R.A. Stirling Gallacher</text></nplcit>, proposes the allocation of subcarriers to a user channel according to a frequency interleaving scheme, that is the frequencies for the subcarriers allocated to a user channel are picked far enough apart so that they can experience different attenuation. By following this method, the bits are distributed over these subcarriers and an averaging effect is achieved, so that on average the raw bit error rate is acceptably low and a Forward Error Correction (FEC) coding mechanism is able to correct the bit errors.

While this allocation method significantly reduces the problem of fading, it still allocates subcarriers which can be useless for downlink transmission with a mobile station. Moreover, subcarriers may be allocated to a specific user channel which experience severe attenuation for this user and are thus not useful for downlink transmission to him, but which could be perfectly adequate, because not faded, for transmission to another user. Still another disadvantage of frequency interleaving allocation is that channel coding for the resulting bit error rate needs a lot of redundancy which reduces the usable net bit rate that can be transmitted over the channel.

A method and system of adaptive channel allocation in a frequency division multiplexed cellular system is disclosed in <patcit id="pcit0001" dnum="US5726978A"><text>US Patent 5,726,978</text></patcit> Certain subcarriers from an initial subset of M subcarriers are reconfigured based on signal quality level and interference level information reported by a plurality of user terminals to the network.

It is the object of the invention to provide an improved subcarrier allocation method for wireless mobile multi-carrier systems.

The object is achieved according to the invention by a method for allocating subcarrier frequencies to a user terminal in a wireless system using multi-carrier modulation according to claim 1, a network element comprising means for subcarrier frequency allocation to a plurality of user terminals in a wireless system using multi-carrier modulation according to claim 9, and a computer program product for controlling a network element of a wireless communications system according to claim 11.

Advantageous configurations of the invention emerge from the dependent claims, the following description and the drawings. For example, it is seen advantageous that, by using the proposed invention, the modulation scheme for every subcarrier frequency can be also selected, and the highest order modulation (e.g. 64 QAM) is used for as many terminals as possible, so that the maximum radio cell traffic throughput for the whole system can be improved in order to serve as many subscriber as possible and that higher data transmission rates are achieved. Also advantageous is that less channel coding/decoding complexity is needed in the user terminal.

An embodiment example of the invention is now explained with the aid of Figures 1 to 5. <ul id="ul0001" list-style="none"><li>Fig. 1 shows an example of a mapping of subcarriers to a user channel into a time-frequency grid of OFDM.</li><li>Fig. 2 shows an example of a diagram of a user channel transfer function with frequency-selective fading.</li><li>Fig. 3 shows a block diagram of a mobile communications system comprising means for subcarrier frequency allocation to a user terminal according to the invention.</li><li>Fig. 4 shows an example of a suitable partitioning of the subcarrier frequency set in two logical blocks of frequency subsets according to the invention.</li><li>Fig. 5 shows an example of a way for sending specific information to the mobile radio network about the user channel transfer function.</li></ul>

Figure 1 shows an exemplary allocation of subcarriers S1 to SN to four user channels A, B, C, D into an OFDM time-frequency T-F grid.

OFDM offers the possibility to flexibly allocate one or more subcarriers S1 to SN to one user or one logical channel A, B, C, D to control the data rate for this user channel. Since this can change also over time in a TDMA system (e.g. with a change period of K symbol periods Ts e.g. a period of 2ms), we have a 2-dimensional resource allocation grid as indicated in Figure 1.

Some of the time-frequency grid locations may not be available for data transmission, because they are used for carrying pilot or signaling information. User assignment of remaining locations can be done based on frequency or time or a combination of both.

Figure 2 shows an example of a diagram of a user channel transfer function Hf with frequency-selective fading over a 5Mhz bandwidth.

In wideband wireless communications, severe frequency-selective fading due to multipath spreading causes a decline in user channel quality. The OFDM signal can have, for example 300 or 700 subcarriers over the bandwidth of 5Mhz and as thus the SNR per subcarrier at the receiver is also highly different over the frequency, that is, some carriers are strongly attenuated while others are not.

A user channel transfer function Hf, such as the one shown in figure 2, is unique and specific for each user terminal and its current reception condition. The user channel transfer function Hf is subject to variations, for example caused by the mobility of the users or by adjacent transmission interference. Such channel estimation must be then performed in the OFDM receiver by methods known in the art in order to improve the quality of data reception.

Figure 3 shows a block diagram of a mobile communications system in which a mobile radio network N, including a plurality of network elements NE, and a plurality of user terminals UE exchange data information via an air interface downlink channel DC and a feedback channel FC using multi-carrier modulation schemes, at least in the downlink, such as OFDM; the network N and the user terminal UE comprising means for subcarrier frequency allocation according to the invention.

The invention starts from the first observation that since, as indicated in the Figure 2 example, there are many user channel transfer functions Hf on a cell at the same time which are typically different, the probability that one subcarrier frequency is attenuated for all the user terminals UE in a radio cell is nearly zero. Furthermore, the invention starts from the second observation that, while some user terminals UE are moving fast and experience a quickly changing and unpredictable channel transfer function Hf with unpredictable frequency selectivity, there is a number of user terminals UE that experience a roughly stationary or slowly changing channel situation where the channel transfer function Hf is predictable for some transmission time intervals in advance.

According to the invention then, and the aforementioned observations, it is possible to allocate specific subcarriers to the user terminals UE based on the knowledge of their unique channel transfer function Hf, that is, such knowledge about the useful or useless subcarriers for a specific terminal is used then by the network N for the allocation process.

According to the invention, once the terminals UE have estimated their channel transfer function Hf over a certain period of time, they report to the network N, over their feedback channel FC, if their channel transfer function Hf is predictable or not and if they want to participate in the frequency-selective subcarrier allocation method or not. Depending on the information, for example, about the number of terminals UE participating in the frequency-selective allocation procedure, their respective traffic demands and their channel transfer function Hf characteristics, the network N partitions the available OFDM time-frequency T-F grid shown in Figure 1 in at least two logical blocks or frequency subsets, at least a first subset containing the frequencies that are used for frequency-selective allocation for the terminals with predictable channel transfer function Hf and which want to participate on this allocation process and at least a second subset with the frequencies that are used for conventional frequency interleaving allocation in case of terminals with no predictable transfer function Hf or the ones which do not want to participate in the frequency-selective allocation method. The participating terminals UE report then back to the network N information about their measured channel transfer function Hf and, if required, interference noise estimate through the feedback channel FC. The network N, according to this information, allocates the subcarrier frequencies from the first subset to the participating terminals in the frequency-selective allocation method by assigning to said terminals subcarriers that are useful for transmission according to their channel transfer function Hf and allocates the other subcarrier frequencies from the second subset to the rest of the terminals according to a conventional frequency interleaving scheme as described previously.

Also according to the invention, the network N, when allocating the subcarrier frequencies to the participating terminals, can further determine the_modulation scheme (e.g. QPSK, 16 QAM, 64 QAM) for said subcarriers.

This allocation procedure is done repetitively, and the frequency-selective allocation lasts an appropriate time interval, in the order of the radio channel coherence time during which stationarity can be assumed, until new participating user terminal UE channel transfer function Hf reports are available again.

The subcarrier allocation principle is also open to incorporate in the allocation process other factors like user/service demand or user/service priority achieving always the best cell throughput possible under these constraints.

The way the participating terminals UE report back to the network N information about their measured channel transfer function Hf and interference noise estimate according to the invention through the feedback channel FC should, in an preferred embodiment to reduce feedback traffic, be in a coded compressed form. For example, this can be done by sending the quantized measurement values on specific pilot subcarriers and, if required, possibly an additional quantized noise estimate; or by sending the estimated delay and amplitude values of the channel impulse response in quantized form and an additional noise estimate; or by just describing on which subcarriers a certain modulation scheme is useful and where not. This last method is done for example by sending an integer function over the axis of subcarrier frequencies indicating which respective modulation scheme (e.g. QPSK, 16 QAM, 64 QAM) on which frequencies are adequate for transmission and on which frequencies not. This last scheme uses the observation that the sending network usually only has the choice between a very limited number of modulation schemes.

Fig. 4 shows an example of a suitable partitioning of the subcarrier frequency set available in two logical blocks of frequency subsets SU1 and SU2, in which the first subset SU1 comprises rasters R0, R1 and R2 and the second subset SU2 comprises rasters R3 to R7.

As shown in the example of the figure, the mobile radio network N partitions the axis of discrete subcarrier frequencies in e.g. 8 interleaved subsets of equally spaced frequency points which we call frequency rasters. So in our example we have frequency rasters R0 to R7. This way, frequency diversity is provided in both subsets SU 1 and SU2, the first subset SU 1 being used for frequency-selective allocation procedure and the second subset SU2 being used for frequency interleaving allocation.

It shall be understood that Figure 4 is just an example of a possible partition method and that more ways of doing said partition are possible, for example, it is possible to do a partition in more subsets or it is possible that any of the subsets is assigned zero frequencies, which would mean that a single set of frequencies would be available for allocation. As also mentioned above when explaining Figure 1, the partitioning of the time-frequency grid can also be made over time and frequency and said partitioning being used for the allocation method.

Fig. 5 shows an example of a way for sending specific information to the network N about the user channel transfer function. This is done by sending the considered modulation scheme (e.g. QPSK, 16 QAM, 64 QAM) as an index and an integer function A(f) indicating which frequencies are adequate for transmission with this modulation scheme and which not.

Said integer function can be also sent in a further compressed form by a source coding algorithm such as run-length coding to minimize traffic on the feedback channel FC. The principle can also be generalized to the case of a multiple input multiple output (MIMO) system where the feedback information is given separately for the different sending antennas of the network N.

For the sake of generalization, it shall be understood, that although for the explanation of the present invention an OFDM modulation scheme has been used, the above proposals can in principle as well be adapted to any multi-carrier modulation scheme apart from OFDM.

It shall also understood that the allocation method herein described can be implemented anywhere in the mobile radio network N, that is, in a network element NE such as a base station or a radio network controller or by means of a radio resource manager entity, inside or outside the network elements NE, which carries out the allocation algorithm as well. This radio resource manager entity can be implemented by hardware or software means.

It is also possible to apply a subcarrier allocation and modulation scheme selection in the uplink channel for data transmission from the terminal to the network using the same principles.