WO2008089595A1

WO2008089595A1 - Time domain interpolation method and apparatus for channel estimation

Info

Publication number: WO2008089595A1
Application number: PCT/CN2007/000207
Authority: WO
Inventors: Yiling Wu; Li Zou; Dun Liu
Original assignee: Thomson Licensing
Priority date: 2007-01-19
Filing date: 2007-01-19
Publication date: 2008-07-31
Also published as: WO2008089595A8

Abstract

A time domain interpolation method for channel estimation under Doppler channels in a multi-carrier system comprises the steps of: dividing the received data into scattered pilots and non-scattered pilots to get the CSI of the scattered pilots; and implementing an interpolation to get the CSI of the non-scattered pilots. Implementing the interpolation comprises the steps of: sorting the non-scattered pilots into a first type pilots capable of being estimated by predicting; and a second type pilots capable of being estimated by filtering; estimating the CSI of the first type pilots by predicting; and estimating the CSI of the second type pilots by filtering. A time domain interpolation apparatus comprises a separation unit, a division unit, and an interpolation unit. The interpolation unit comprises a sorting unit, a predicting unit, and a filtering unit. This apparatus can achieve the balance between performance and hardware cost.

Description

Time Domain Interpolation Method and Apparatus for channel Estimation

FIELD OF THE INVENTION

The present invention generally relates to communication technologies, and particularly to time domain channel estimation in wireless communication systems, and more particularly to a time domain interpolation method and apparatus with high efficiency for channel estimation under Doppler channels in multi- carrier systems .

BACKGROUND OF THE INVENTION

Frequency division multiplexing (FDM) is a technology for transmitting multiple signals simultaneously over a single transmission path, such as a cable or a wireless system. Each signal travels within its own unique frequency range (carrier) , which is modulated by the data (text, voice, video, etc.) .

Orthogonal FDM' s (OFDM) spread spectrum technique distributes the data over a large number of carriers that are spaced apart at precise frequencies. This spacing provides the "orthogonality" in this technique which prevents the demodulators from seeing frequencies other than their own. OFDM (Orthogonal Frequency Division Multiplexing) works by splitting the radio signal into multiple smaller sub-signals that are then transmitted simultaneously at different frequencies to the receiver. OFDM reduces the amount of crosstalk in signal transmissions. The benefits of OFDM are high spectral efficiency, resiliency to RF interference, and lower multi-path distortion. This is useful because in a typical terrestrial broadcasting scenario there are multipath-channels (i.e. the transmitted signal arrives at the receiver using various paths of different length) . Since multiple versions of the signal interfere with each other (inter symbol interference (ISI)), it becomes very hard to extract the original information.

OFDM is sometimes called multi-carrier or discrete multi-tone modulation. It is the modulation technique used for digital TV in Europe, Japan, and Australia. 802.11a WLAN, 802.16, and WiMAX technologies use OFDM.

In general, the OFDM system inserts some pilots in the frame at the transmitter site. Then at the receiver, a channel estimation can get the channel coefficients at the pilots and then estimate the non-point samples' channel coefficients by interpolating in both time domain and frequency domain.

Conventionally, filtering or predicting is used for the time domain interpolating. The predicting method can save hardware cost but can not achieve acceptable performance in the practical application. Compared with the predicting method, the filtering method can achieve better performance but costs much more hardware resource such as memory size.

If only filtering is used, it will cost more hardware resource such as memory, while only predicting will suffer worse performance. However, in traditional methods (for example, referring to "Optimum Receiver Design for OFDM-Based Broadband Transmission—Part II: A Case Study", IEEE TRANSACTIONS ON COMMUNICATIONS, VOL.49, NO.4, APRIL 2001), time-domain interpolating only focus on filtering method or predicting method. In other words, filtering or predicting is used for time domain interpolating independently.

So it needs to figure out a best tradeoff between the filtering and predicting methods. Therefore, one object of this disclosure is to provide a solution which can reach a balance between cost of hardware resource and performance .

In order to achieve the above object, the present invention proposes a new method to finish time domain interpolating for channel estimation in pilot-aided OFDM systems, which brings the prediction concept into the traditional filtering method. In order to achieve acceptable performance for mobile reception, the traditional method has to use the filtering method to resist Doppler effects although the high-order filter results in larger memory size. So this disclosure proposes a new design which can combine prediction and filtering methods efficiently and the performance is maintained good performance. The combination is not a simple ratio multiplying but a special signal processing aiming at an OFDM system such as DVB-T demodulator design.

SUMMARY OF THE INVENTION

In one aspect of the present invention, a time domain interpolation method for channel estimation under Doppler channels in a multi-carrier system is provided to meet this need. The time domain interpolation method comprises the steps of: separating the received data into scattered pilots and non-scattered pilots;

dividing the known transmitted scattered pilots to get CSI of the received scattered pilots; and

implementing an interpolation to get the CSI of the non-scattered pilots.

According to the present invention, implementing the interpolation comprises the steps of:

sorting the non-scattered pilots into a first type pilots capable of being estimated by predicting; and a second type pilots capable of being estimated by filtering;

estimating the CSI of the first type pilots by predicting; and

estimating the CSI of the second type pilots by filtering .

According to the present invention, the predicting process is performed by using the following Equation (5) :

csi_t_intp(m-I,n) = 2-SP_csi(m~2,ή)-csi_t_Jntp(m-2>,n) (5)

wherein, csi_t_intp(m,ri) is the CSI of mth (in DVB-T, m =12) symbol and n th (for example, n =3) sub- carrier; SP_csi(m,n) is the scattered pilot CSI of m th symbol and nth sub-carrier.

According to the present invention, the filtering process is performed by using the following Equation (3) and (4) : csi _t _intp(m,ή) = —SP_csi(m-2,ή) + —SP_csi(m + 2,ri) ( 3 )

1 3 csi __ t _ intp(m + l,ή) = -SP_ csi(m -2,n)+—SP_ csi{m + 2,ή) ( 4 )

wherein, csi_t_intp(m,ή) is the CSI of røth symbol and nth sub-carrier; SP_csi(m,ή) is the scattered pilot CSI of m th symbol and nth sub-carrier.

According to the present invention, the multi- carrier system is an OFDM (Orthogonal Frequency Division Multiplexing) system.

In another aspect of the present invention, a time domain interpolation apparatus for channel estimation under Doppler channels in a multi-carrier system is provided to meet this need. The time domain interpolation apparatus comprises a separation unit, a division unit, and an interpolation unit. The separation unit is used for separating the received data into scattered pilots and non-scattered pilots. The division unit is used for dividing the known transmitted scattered pilots. By the division operation, the CSI of scattered pilots are got.

The interpolation unit is used for implementing an interpolation to get the CSI of the non-scattered pilots.

The CSI of some non-scattered pilot can be estimated by predicting_/ and the CSI of some other non- scattered pilot can be estimated by filtering. According to the present invention, the interpolation unit comprises a sorting unit, a predicting unit, and a filtering unit. The sorting unit is used for sorting the non-scattered pilots into a first type pilots capable of being estimated by predicting; and a second type pilots capable of being estimated by filtering. The predicting unit is used for estimating the CSI of the first type pilots by predicting. The filtering unit is used for estimating the CSI of the second type pilots by filtering.

According to the present invention, the predicting unit performs the predicting by using the following Equation (5) :

csi_t_\x&p(m-1₅ή) = 2-SP__csi(m-2,ή)-csi_t_wXp{m-3,ή) (5)

wherein, csi_t_intp(m,ri) is the CSI of m th symbol and n th sub-carrier; SP_csi(m,n) is the scattered pilot CSI of mth symbol and nth sub-carrier.

According to the present invention, the filtering unit performs the filtering by using the following Equations (3) and (4) :

csi_t__intp(m,n) =-SP_csi(m-2,n)+-SP_csi(m+2,n) (3)

1 3 csi t intp{m-Y\,n)-~SP csi(m-2,ή)+—SP_csi(m+2,ή) (4)

4 4

wherein, csi_t_intp(m,ή) is the CSI of m th symbol and H th sub-carrier; SP_csi(m,ή) is the scattered pilot CSI of mth symbol and «th sub-carrier.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment ( s ) of the invention together with the description which serves to explain the principle of the invention. In the drawings:

Fig. 1 shows the pattern of scattered pilots in DVB-T systems;

Fig. 2 shows the simplified pattern graph of scattered pilots;

Fig. 3 shows time domain interpolation according to the related art;

Fig. 4 shows the sorted pattern graph of scattered pilots of Fig. 2 according to the present invention;

Fig. 5 shows the modified time domain interpolation method for Fig. 4 according to the present invention;

Fig. 6 shows the architecture of time domain interpolator which combines filtering and predicting together;

Fig. 7 shows a flow diagram of a time domain interpolation method according to the present invention; and

Fig. 8 shows a block diagram of a time domain interpolation apparatus according to the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

The technical features of the present invention will be described further with reference to the embodiments. The embodiments are only preferable examples without being limited to the present invention. It will be well understood by the skilled person in the art upon reading the following detailed description in conjunction with the accompanying drawings.

The present invention proposes a new solution to finish time domain interpolating for channel estimation in pilot-aided OFDM systems. In pilot-aided OFDM systems, the process of channel estimation is usually preformed by using scattered pilot information contained in the OFDM signal. Scattered pilots provide a reference signal of known amplitude and phase on every n OFDM sub-carriers per OFDM symbol. Channel estimation can be achieved by interpolating in both time domain and frequency domain.

Usually, the received data is separated to two parts: pilots, useful data and TPS. For the TPS, it is need not to estimate the CSI on TPS represented sub- carriers. TO pilots, the CSI can be easily achieved because its transmitted power is obvious larger than useful data. To the useful data, the CSI is estimated based on the CSI of pilots. The estimation method is usually one FIR filter. In this invention disclosure, the estimating method for the CSI of useful data, not only predicting but also filtering is used.

The time domain interpolation in DVB-T (Digital Video Broadcasting) systems is taken as an example and linear interpolation is used for demonstrating. The pattern of scattered pilots in DVB-T systems can be seen in Fig. 1. Here, black points represent scattered pilots while grey points represent received data, TPS (Tracking Pilot Signal) and continue pilots. Fig. 2 shows the simplified pattern graph of scattered pilots. The function of time domain interpolation is to estimate the Channel State Information (CSI) of grey points in Fig. 2 by linear filtering using the information provided by scattered pilots .

Fig. 3 shows time domain interpolation according to the related art. As shown in Fig. 3, taking the first column as an example, linear filtering can be completed using the following equations (2) ~ (4) :

3 1 csi _t_ intpim -l,ή) = -SP_ csi(m, ή)+—SP_ csi(m + 2,ή) ( 2 )

csi _ t __ intpim, ή) = -SP __ csi(m - 2, ή) + - SP _ csi(m + 2,ή) ( 3 )

csi _ t _ intpim + l,ή) = —SP_ csi(m -2,ή)+—SP_ csiim + 2,ή) ( 4 )

4 4

wherein, csi_t_intp(m,ri) is the CSI of /wth (in DVB-T, m =12) symbol and n th (for example, n =3) sub- carrier; SP_csi(m,ή) is the scattered pilot CSI of m th symbol and nth sub-carrier.

It is clear that at least three OFDM symbols need to be stored if only linear filtering is used in time domain interpolation.

Besides, prediction can also be used to do time domain interpolating and it stores only one OFDM symbol when linear predicting is used. However, the performance of channel estimation will suffer degradation if only predicting is used. Especially over large Doppler channels, the performance of this predicting method is not acceptable.

To achieve a balance between cost of hardware resource and performance of system, this disclosure proposed a new method which combines filtering and predicting methods together. Fig. 4 shows the sorted pattern graph of scattered pilots of Fig. 2 according to the present invention. As shown in Fig. 4, the grey- points in Fig. 2 can be sorted into three classes. The grid points represent the received data whose CSI can be estimated by predicting. And the CSI of dotted ones and lined ones can be estimated by filtering. Both prediction and filtering are performed in a linear way.

Fig. 5 shows the modified time domain interpolation method for Fig. 4 according to the present invention. As shown in Fig. 5, calculation of the CSI of dotted points and lined points remain the same way as shown in (3) and (4), and the CSI of grid ones can be obtained by using the following Equation (5):

csi_t_mtp(m-l,n) = 2-SP__csi(m-2,ri)-csi_t_mtp(m-3,n) (5)

wherein, csi_t_intp(m,ή) is the CSI of mth symbol and nth sub-carrier;

SP_csi(m,ή) is the scattered pilot CSI of mth symbol and nth sub-carrier.

Then, only two OFDM symbols need to be stored. The architecture of time domain interpolator which combines filtering and predicting together is shown in Fig. 6.

Fig. 7 shows a flow diagram of a time domain interpolation method according to the present invention. As shown in Fig. 7, the time domain interpolation method comprises the steps of:

step S702, separating the received data into scattered pilots and non-scattered pilots;

step s703, dividing the known transmitted scattered pilots to get the CSI of the received scattered pilots; and

step S704, implementing an interpolation to get the CSI of the non-scattered pilots.

Step S704 comprises the steps of:

step S706, sorting the non-scattered pilots into a first type pilots capable of being estimated by predicting; and a second type pilots capable of being estimated by filtering;

step S708, estimating the CSI of the first type pilots by predicting and estimating the CSI of the second type pilots by filtering.

Fig. 8 shows a. block diagram of a time domain interpolation apparatus 800 according to the present invention.

As shown in Fig. 8, the time domain interpolation apparatus 800 comprises: a separation unit 802 for separating the received data into scattered pilots and non-scattered pilots, a division unit 803 for dividing the known transmitted scattered pilots to get the CSI of scattered pilots; and a interpolation unit 804 for implementing an interpolation to get the CSI of the non- scattered pilots. The interpolation unit 804 comprises: a sorting unit 806 for sorting the non-scattered pilots into a first type pilots capable of being estimated by predicting and a second type pilots capable of being estimated by filtering, a predicting unit 808 for estimating the CSI of the first type pilots by predicting and a filtering unit 810 for estimating the CSI of the second type pilots by filtering.

Since the number of sub-carriers in many OFDM systems is large (For example, there are 1705 or 6817 useful sub-carriers in DVB-T systems) , storing less symbols means saving a lot of memory resources. Comparison among three types of time domain interpolation can be seen in table 1.

Table 1 - Comparison between three methods

The BER performance simulation is done in DVB-T system. The simulation parameters are listed in table 2. To know the performance of time-domain interpolation accurately, only the module of time-domain interpolation use fixed-point simulation and other modules use float- point simulation. Table 2 - Parameter settings in BER performance simulation

Based on the comparisons, it is clear that the proposed method not only achieves the acceptable performance but also saves around 30% memory size, which is a key parameter for chip size.

Obviously, those skilled in the art should understand that, the general idea of combining filtering and predicting can be used for interpolation even outside the field of channel estimation.

As described above, this disclosure provided a solution which can reach a balance between cost of hardware resource and performance. The simulation result and hardware test proved its efficiency. The above Table 1 gives both the simulation result and hardware test result in detail.

Obviously, those skilled in the art should understand that the above mentioned modules or steps can be achieved using a general computing device, and they can be resided on an individual computing device, or distributed in the network consisting of multiple computing device. Alternately, they can be achieved by a computer program which can be executed by a computing device, and then they can be executed after storing the same in a storage device, or they can be achieved by being made into integrate circuit chips, or by making multiple modules or steps of them into an individual integrate circuit chip. In this case, the present invention is not limited to the combination of any specific hardware and software. Understandable that changes in these specific embodiments are obvious for those skilled in the art, and do not deviate from the range of the present invention.

Alternately, the present apparatus can be wrapped as IP core, and based on memory size, it can detecte whether this invention is used or not by the present apparatus .

Alternately, the present method can be implemented in one chip, and it can be detected by testing the maximum tolerable Doppler frequency offset for this receiver chip.

The specific applications could be channel estimation module in the receiver of pilot-based multi- carrier system, such as DVB-T demodulator IP core, DVB-T demodulator chip, DVB-H demodulator IP core, DVB-H demodulator chip, 802.16a demodulator IP core, 802.16 demodulator chip, etc. The complexity of receiver to handle Doppler frequency offset will be reduced obviously.

Whilst there has been described in the forgoing description preferred embodiments and aspects of the present invention, it will be understood by those skilled in the art that many variations in details of design or construction may be made without departing from the present invention. The present invention extends to all features disclosed both individually, and in all possible permutations and combinations .

Claims

1. A time domain interpolation method for channel estimation under Doppler channels in a multi-carrier system, comprising the steps of:

separating the received data into scattered pilots and non-scattered pilots;

dividing the known transmitted scattered pilots to get the CSI of the received scattered pilots; and

implementing an interpolation to get the CSI of the non-scattered pilots.

2. The time domain interpolation method according to claim 1, wherein implementing the interpolation comprising the steps of:

estimating the GSI of the first type pilots by predicting; and

estimating the CSI of the second type pilots by filtering.

3. The time domain interpolation method according to claim 2, wherein the predicting process is performed by using the following equation (5) : csi_t_wtp(m-\,ri)=2-SP_csi(m-2,ή)-csi_t_intp(m-3,ή)

(5)

wherein, csi_J_intp(m,ή) is the CSI of m th symbol and wth sub-carrier; and

SP_csi(m,ή) is the scattered pilot CSI of m th symbol and nth sub-carrier.

4. The time domain interpolation method according to claim 2, wherein the filtering process is performed by using the following equations (3) and ( 4 ) :

csi _ t _ intp(m, ri) = -SP_ csi(m -2,ri) +—SP_ csi(m + 2,ή) ( 3 )

1 3 csi t mtp(m+l,ή) = —SP csi(m -2,ή)+—SP csi(m + 2,ή) ( 4 )

4 4

wherein, csi_t__intp(m,ri) is the CSI of m th symbol and »th sub-carrier; and

SP_csi(m,n) is the scattered pilot CSI of m th symbol and nth sub-carrier.

5. The time domain interpolation method according to anyone of claims 1 to 4, wherein the multi-carrier system is an OFDM system.

6. A time domain interpolation apparatus for channel estimation under Doppler channels in a multi-carrier system, comprising:

A separation unit for separating the received data into scattered pilots and non-scattered pilots; a division unit for dividing the known transmitted scattered pilots to get the CSI of the scattered pilots; and

an interpolation unit for implementing an interpolation to get the CSI of the non-scattered pilots .

7. The time domain interpolation apparatus according to claim β, wherein the interpolation unit comprising :

a sorting unit for sorting the non-scattered pilots into a first type pilots capable of being estimated by predicting; and a second type pilots capable of being estimated by filtering;

a predicting unit for estimating the CSI of the first type pilots by predicting; and

a filtering unit for estimating the CSI of the second type pilots by filtering.

8. The time domain interpolation apparatus according to claim 7, wherein the predicting unit performs the predicting by using the following equation (5) :

csi_t_intp(m-1,ή) = 2•SP_csi(m-2,ri)-csi_t_intp(m-3,ή)

(5)

wherein, csi_t_intp(m,ή) is the CSI of m th symbol and nth sub-carrier; and

SP_csi(m,ή) is the scattered pilot CSI of m th symbol and nth sub-carrier.

9. The time domain interpolation apparatus according to claim 7, wherein the filtering unit performs the filtering by using the following equations (3) and (4) :

csi_t_intp(m,n) =—SP_csi{m-2_iή)+—SP_csi(m+2,n) (3)

csi _ t _ intp{m +l,n) = -SP_ csi(m - 2, n) + -SP _ csi(m + 2,ή) ( 4 )

wherein, csi_t_intp(m,n) is the CSI of m th symbol and nth sub-carrier; and

SP_csi(m,ή) is the scattered pilot CSI of m th symbol and nth sub-carrier.

10. The time domain interpolation apparatus according to anyone of claims 6 to 9, wherein the multi-carrier system is used in an DVB-T system.