WO2007051849A1

WO2007051849A1 - Method for calibrating transmissions in a radio communications system

Info

Publication number: WO2007051849A1
Application number: PCT/EP2006/068096
Authority: WO
Inventors: Martin DÖTTLING; Bernhard Raaf; Ingo Viering
Original assignee: Nokia Siemens Networks Gmbh & Co. Kg
Priority date: 2005-11-07
Filing date: 2006-11-06
Publication date: 2007-05-10

Abstract

According to the invention, a method for calibrating transmissions between radio stations in a radio communications system is disclosed, wherein when establishing a packet data flow, medium access control (MAC) protocol functions are used for transmitting signals relating to the calibration of at least one of the radio stations.

Description

_

Description

Method for calibrating transmissions in a radio communica^¬ tions system

Future OFDM-based (OFDM - Orthogonal Frequency Division Mul^¬ tiplex) cellular radio communications systems, like they are currently discussed in the framework of the WINNER (Wireless World Initiative New Radio) project (internet www.ist-win- ner.org) sponsored by the European Union, will exploit reci^¬ procity for spatial processing based on short-term channel state information (CSI) in the TDD (Time Division Duplex) physical layer mode. The idea is to avoid overhead caused by periodic measurement feedback signalling by estimating the channel of the forward link based on return link measurements at the transmitter.

Although the pure propagation channel of a TDD system is reciprocal, the imperfections of the RF (Radio Frequency) front-ends, which also contribute to the effective experi^¬ enced channel, normally prevent from the exploitation of the reciprocity principle. A calibration step is thus required to compensate for these impairments and restore the channel re^¬ ciprocity. Also for beamforming (both TDD and FDD (Frequency Division Duplex) ) calibration is required in order to be able to design beams with a given direction.

It is thus an object of the present invention to provide a calibration functionality for such systems. This object is addressed by the features of independent claim 1.

According to a first aspect of the invention, a method for calibrating transmissions between radio stations in a radio communications system is proposed, wherein when establishing a packet data flow, medium access control (MAC) protocol functions are used for transmitting signals relating to the calibration of at least one of the radio stations.

According to a second aspect of the invention, signals are used to calibrate radio frequency equipment in the at least one radio station.

Furthermore, a radio communications system is proposed com- prising means for realising the inventive method.

The present invention will become more apparent from the de^¬ scription given herein below and from the accompanying drawings which are given by way of information only, and thus are not limitative of the present invention, and wherein:

FIG 1 depicts a flow setup (context establishment) for a new downlink flow, and

FIG 2 depicts a modified flow setup (context establish^¬ ment) for a new downlink flow including calibration function .

A number of alternative approaches to perform calibration were already discussed within the WINNER project. Calibration may be based on RF (Radio Frequency) approaches that model separately each contribution of the RF imperfections, on self-calibration approaches or on global approaches involving over-the-air signalling of calibration information.

Nevertheless, in the current system design, no functional de^¬ scription or service specification of such a calibration function are available and its integration in the MAC design is not specified. In the following, a short overview of associated MAC func^¬ tions is provided.

New uplink (UL) and downlink (DL) flows (flow: packet stream from one source to one destination or several destinations) are established by the so called RLC (Radio Link Control layer) Flow establishment function. It requires a detailed setup of a flow context over each involved hop. Such flow context establishment is executed by the MAC (Medium Access Control layer) flow setup function, whereas the flow context release is initiated by the RLC and executed by the MAC flow termination function.

When a new downlink flow is established, it is given a local flow address (flow_address) that is unique within the cell. Its destination UT (User Terminal) is notified and a resource scheduling buffer queue is initialised.

In the FDD mode, flows to/from half-duplex terminals are as^¬ signed to one of four groups :

Group 1 transmits in the downlink the first half of the frame, and in the uplink during the latter half. Group 2 transmits/receives in the opposite way. Group 3 contains half-duplex terminals that have adaptable and flexible uplink and downlink transmission periods. Group 4 relates to full-duplex terminals.

Point-to-point flows are initially assigned either for adap- tive or for non-frequency adaptive transmission by the flow setup controller. The choice is based on the capability of the terminal, the average SINR (Signal to Interference Noise Ratio) and the velocity of the terminal. The choice is also affected by the potentially available spatial schemes, and their CSI and CQI (Channel Quality Indicator) requirements. The initial assignment may be changed later if the circum^¬ stances change.

Currently, the setting up of a new flow, e.g. a new downlink flow, is realised as described in the following steps with reference to FIG 1.

1. Start: a. For uplink flows: Start when receiving Flow_context_ind (UL, I) from RLC. b. For downlink flows: Start when receiving Flow_context_ind (DL, I) from RLC.

2. If the flow is to/from a UT/RN (Relay Node) without other flows, the UT/RN is polled (Setup_request ) and responds by transmitting Terminal_capabilities and Ba- sic_LA_measurements .

3. In FDD mode, assign the flow to one of the following half-duplex groups : a. Group 1 has downlink transmission in the first slot. b. Group 2 has downlink transmission in the second slot of the frame . c. Group 3 has flexible assignments of uplinks and downlinks, d. Group 4 refers to full-duplex terminals.

If a terminal already has active flows that belong to one of the groups, the new flow is assigned to that group. To dis- tinguish between assignment to Groups 1 and 2, the long-term part of the spatial scheme pre-selection function is used to help making this decision and to obtain relevant background information on other terminals that are spatially close. 4. For DL flows: Define new downlink queues for the data.

5. Assign a flow_address that is unique within the cell to each queue assigned to the flow. Update the flow_table of the cell with relevant parameters.

6. Determine parameters for retransmission and segmentation (ARQ_parameters table) , based on RLC layer input (RLC_HARQparameters) .

7. Create Adaptation_parameters based on Ba- sic_LA_measurements, i.e. determine if adaptive or non-fre^¬ quency adaptive transmission is to be used. The potentially available spatial schemes, and their requirements, are taken into account in the decision. In case of adaptive transmis^¬ sion, assign the new flow to one of the competition bands. In case of non-frequency adaptive transmission, determine ini^¬ tial slow link adaptation and power control settings.

8. Determine spatial transmission scheme by calling the spatial scheme controller, with Adaptation_parameters and Ba- sic_LA_measurements as input argument, which returns an up^¬ date of the Spatial_scheme table for the entry flow_address .

9. Send Flow_setup_message to terminals over PHY Transmit control signalling service.

The inventive features are discussed in the following in more details .

• Introduction of a calibration function in the MAC design and associated input and output variables: o new MAC variable that contains the calibration status .

• Description of the interaction of this functionality with other functions (triggers for involving it, signals to and from other functions) : o Integration in MAC flow state control, o Triggered additionally each time the input variables change . o The calibration function is a function that triggers different calibration procedures, including manage^¬ ment of partial unavailability of radio resources for data transmission, triggering, execution and signalling of calibration procedures. o The calibration function receives and transmits sig^¬ nals from/to the PHY (Physical layer) services meas^¬ urements, transmit control signalling, and transfer over physical layer, as well as from/to the MAC func^¬ tions constraint processor, resource partitioning and resource scheduling.

• Overall description of calibration procedure, in particular the update of calibration matrices when multiple us^¬ ers perform calibration.

The following detailed descriptions of the inventive func^¬ tional features provide an example of a calibration procedure suitable for the system concept and MAC design of the WINNER project. Required input and output variables, as well as sig- nailing and measurements are furthermore provided. The cali^¬ bration function is embedded in the MAC flow state control and spatial scheme control. As an illustrative example the current names of functions and services as defined in WINNER are used. However, the scheme is applicable in any similar context of the MAC design

A) Integration of Calibration Function in MAC design and flow state control

The MAC flow setup procedure described above is amended as follows :

• For each terminal an additional variable Calibra- tion_Status is maintained. It contains an actual calibration status of all flows to that terminal. The initial status is uncalibrated, furthermore different levels of calibration ex^¬ ist depending on the actual calibration procedure are performed. The uncalibrated status can be reached again after specific events, like expiry of a timer, change of the MAC connection state (e.g. transition into passive state), etc.

• A new Calibration function is introduced, which uses the input variables Calibration_Status and Spatial_Scheme and triggers the required calibration procedures for each flow according to the given requirements.

• The Calibration function should be invoked each time at least one of the input variables has changed, initially for example between step 8 and 9 in the flow setup procedure described above. This initial call of the calibration func- tion is also depicted in FIG 2.

• The Calibration function receives and transmits sig^¬ nals from/to the PHY services measurements, transmit control signalling, and transfer over physical layer, as well as from/to the MAC functions constraint processor and resource scheduling.

B) Details on the Calibration Function Based on the input variables Calibration_Status and Spa- tial_Scheme, the required calibration procedures for each flow are detected. This function therefore maintains a list comprising the calibration requirements for each possible Spatial_Scheme and compares it with the current Calibra- tion_Status. This list might either be static, updated peri^¬ odically or event triggered by a dedicated higher layer con^¬ trol message. It is important to distinguish different impact of calibration on other system functions and services: • temporal unavailability of transmitter and receiver,

• triggering measurements on the PHY layer,

• signalling to different functions and services,

• over-the-air control signalling to peer entities.

Various calibration procedures are conceivable. It would be advantageous to distinguish three procedural classes:

• No calibration activity: either no calibration required for this flow or existing calibration sufficient.

• Self-calibration: all calibration procedures that do not involve over-the-air signalling of calibration information (e.g. calibration based on models of RF impairments or self-calibration approaches.

• Global calibration: calibration procedures that include over-the-air signalling of calibration information.

Internal calibration procedures (e.g. based on cascading transmitter and receiver) may result in a temporal unavail^¬ ability of (parts of) the transmitter and receiver in case they need to be applied during operation. This needs to be signalled to the entities involved, such as the MAC resource scheduler. Therefore, a protocol based on request and confir^¬ mation is required. As an example it may be based on the fol^¬ lowing signals: • Suspend_tx_request, suspend_rx_request : the request of the calibration function to suspend transmission or reception in order to perform calibration.

• Confirm_suspend_tx, Confirm_suspend_rx : the corresponding confirmation generated by the function that has the overall control to all involved functions and entities. Such message should also contain the exact time duration of the unavailability.

It should be noted that although no explicit calibration in^¬ formation is transmitted over the air, over-the-air signalling of information may still be required, e.g., to prevent user terminals from transmitting signals to a base station that is currently unavailable.

Over-the-air calibration requires explicit signalling of calibration information and may include the following protocol parts :

• The Calibration function notifies the involved MAC functions (constraint processing, resource partitioning, and/or resource scheduling) that a certain part of the re^¬ sources (chunk layers) (Chunk: basic resource unit on radio channel. A time-frequency resource consisting of n_sub adjacent subcarriers and n_symb consecutive OFDM symbols with chunk du- ration T_chunk. Chunk layer: chunk within one spatial channel (layer) . There are Q_c layers in the cell) will be used for calibration . o This may either cause a reservation of additional re^¬ sources for the calibration reference signal, or o cause a reuse of existing control signals (like syn^¬ chronisation or pilots) for calibration purposes. In this case certain additional restrictions may apply to these control signals, therefore the involved MAC functions must be notified as well.

• The Calibration function notifies the PHY service measurements which calibration measurements must be started and how the measurement shall be performed (type and time du^¬ ration of calibration, etc.). o This involves signalling to the local PHY measurement service in the BS (Base Station) , but also to the UT in order to start the measurements there and/or to trigger the message Calibration_meas_feedback, containing the calibration measurement .

• The Calibration function triggers the PHY service transmission over PHY layer to insert the calibration test signal (reference signal) to be transmitted in the reserved resources in case dedicated calibration reference signal is used (or no existing control signal that can be reused is configured at this time) .

• The PHY service measurements send a Calibra- tion_meas_feedback message over the air conveying the calibration measurement (e.g. a channel measurement) Calibra- tion_measurement (e.g. from UT to BS, or from UT to RN) .

• After decoding this feedback it is forwarded to the Calibration function where it is further processed to calcu^¬ late (or update) the required calibration information.

• The Calibration function sends Calibra- tion_matrix_feedback messages conveying the calibration information to the involved entities. o Typically the involved entities reside in both ends of the links. The calibration matrices can for exam^¬ ple be taken into account while calculating the link adaptation parameters (in particular the linear pre- coding weights) .

Calibration is not only required for spatial processing techniques relying on channel reciprocity. In order to allow in^¬ terference management, constraints may apply on the power ra- diated in different directions for different chunks. These constraints are handled by the constraint processor function. Calibration is also required here in order to enable the gen^¬ eration of beam pattern with known directions in space.

The calibration reference signal may reuse existing control signals, such as synchronisation or pilot signals. The timing (start, duration, kind of averaging, etc.) of the measurement and signalling may either be signalled when the procedure is triggered or reported along with the measurement feedback. Antennas per pilot can be used right away, for beam-specific pilots additional calculations at the central unit are re^¬ quired which impose restrictions on the averaging possible in the measurements and involve exact knowledge of the measure^¬ ment time. This is required to avoid averaging over measure- ments with different beamforming weights, and to be able to know which beamforming weights were applied at the time of the measurement. Therefore in case of beam-specific pilots the central unit must be able to control the averaging window and at least know (or better control) the measurement time. Explicit signalling of these parameters can be avoided by fixed timing relationships in the calibration procedure, i.e. after receiving the request to perform the measurements it is a priori known in all involved entities at which time and how long the measurement will be performed and at which time it will be reported back.

The calibration procedure can also be performed sequentially per antenna, leading to partial unavailability of resources (like chunk layers, antennas, etc.) or other means for achieving orthogonality, such as code multiplexing, frequency multiplexing may be adopted.

If multiple users are calibrated, the central unit (base sta^¬ tion or relay node) obtains multiple estimates of its own calibration matrices, which can be used to iteratively in^¬ crease the accuracy. Therefore it is required to be able to send an update of the Calibration_matrix_feedback even with- out additional transmission of the calibration reference sig^¬ nal and measurement .

It should be noted that in order to benefit from the calibra^¬ tion and reciprocity, pilot transmission in the return link is required even in cases where no return link data is trans^¬ mitted. Therefore, depending on the Spatial_Scheme, Calibra- tion_Status and the MAC state of the flow (e.g. for all flows not in the passive state) , and potentially additional input variables, this permanent or periodic pilot transmission (even without associated data) should be triggered in the PHY service transfer over the physical layer.

According to the invention, a calibration function is introduced into the MAC design. Without this function spatial processing based on short-term CSI would not be feasible or would suffer from significant performance losses due to ex^¬ plicit return link CSI feedback. The calibration function could be easily integrated in the current MAC design and therefore correct interworking with other functions of the WINNER framework is ensured

Claims

Patent claims

1. Method for calibrating transmissions between radio sta^¬ tions in a radio communications system, wherein when establishing a packet data flow, medium access control (MAC) protocol functions are used for transmitting signals relating to the calibration of at least one of the radio sta^¬ tions .

2. Method according to claim 1, wherein the signals are used to calibrate radio frequency equipment in the at least one radio station.

3. Radio communications system, comprising means for realis- ing the method according to claim 1.