Packet interleaving method

Abstract

The packet interleaving method includes selecting successive input sets of consecutive input packets (X1 . . . XNin) received from a forward correction module (14), each input packet (Xj) being a vector of constellation points of a predetermined constellation diagram. For each input set, it further includes generating an output set of output packets (O1 . . . ONout), each output packet (Om) being a vector of constellation points, by distributing the constellation points of each input packet (Xj) of the input set, and sending the output packets (O1 . . . ONout) of the output set to a modulator (18). The input set including Nin input packets (X1 . . . XNin) and each of the Nin input packets (X1 . . . XNin) including a same number Lin of constellation points, the number Nout of output packets in the output set is related to Lin by the relation Lin=A×Nout, where A is a fixed whole number.

Description

The present invention relates to a packet interleaving method, intended to be used in a wireless emitting device.

It is known to interleave packets by:

• • selecting successive input sets of consecutive input packets received from a forward correction module, each input packet being a vector of constellation points of a predetermined constellation diagram, and
  • for each input set, generating an output set of output packets, each output packet being a vector of constellation points, from the input packets.

The output packets are sent to a modulator that generates symbols emitted by an antenna. Each constellation point of the output packet is thus emitted by the antenna in a symbol.

Several functions are known in the art, for performing the interleaving operation. However, they all have the same drawback of being dedicated to a predetermined set of modulation parameters (e.g. number of pilot carriers, guard interval, etc.).

When the modulation parameters of the modulator are changed, these functions are not able to correctly interleave the constellation points of the input packets between the output packets.

It is an object of the invention to provide a packet interleaving method that is able properly interleave the input packets, even when the modulation parameters of the modulator are changed.

Accordingly, the invention relates to a packet interleaving method comprising:

• • selecting successive input sets of consecutive input packets received from a forward correction module, each input packet being a vector of constellation points of a predetermined constellation diagram, the packet interleaving method being characterized by, for each input set:
  • generating an output set of output packets, each output packet being a vector of constellation points, by distributing the constellation points of each input packet of the input set between the output packets of the output set,
  • sending the output packets of the output set to a modulator configured for generating, from the output packets, a waveform, destined to be provided to an antenna,
  • the input set comprising N_in input packets and each of the N_in input packets comprising a same number L_in of constellation points, the number N_out of output packets in the output set is related to L_in by the relation L_in=A×N_out, where A is a fixed whole number.

Other features of the invention are introduced in the dependent claims.

The invention further relates to a computer program product according to claim 10.

The invention further relates to a packet interleaver according to claim 11.

The invention will be explained in more detail hereafter, with reference to the drawings, wherein:

FIG. 1 is a bloc diagram of a wireless data emitting device according to an embodiment of the invention,

FIGS. 2 and 3 are representations of data packets, illustrating the operation of a frame mapper of the wireless data emitting device of FIG. 1, and

FIG. 4 is a bloc diagram representing the steps of a method for setting up an emitting device.

Functional elements of a wireless data emitting device 10 are illustrated on FIG. 1. Of course, several physical implementations may be designed, which achieve the functions of the functional elements illustrated on FIG. 1, without departing from the scope of the invention.

In the following description, data is organized in data packets. A data packet comprises a plurality of ordered elements, and will be represented as a vector. The length of a data packet designates the number of elements it contains.

The emitting device 10 is configured for receiving a flow of data constituted as a sequence of information packets p_k=[p₁ . . . p_K]. The elements of the information packets are either bits or bytes. In the described example, each input data packet p_k is a MPEG data packet.

The emitting device 10 is configured for generating, from the input data packets p_k, a waveform W that is provided to an antenna 12 for emitting the data.

The emitting device 10 comprises three stages 14, 16, 18, implemented in the physical layer of the OSI model, through which the flow of data passes.

1. Forward Error Correction

The first stage is a forward error correction (FEC) module 14, which outputs FEC data packets X_j.

More precisely, the FEC module 14 first comprises a concatenation module 19 configured for taking a number Nb of one or several information packets p_k, and regrouping them into a large data packet P_j=[P_1,j . . . P_N,j].

Considering one large data packet P_j, the FEC module 14 comprises a FEC coder 20 configured for generating a coded data packet C_j=[C_1,j . . . C_M,j], which contains the initial information plus redundant information. The rate of redundant information relative to the initial information data is referred to as the code rate. The FEC coder 20 uses for example the Turbo code.

The FEC module 14 further comprises a bit interleaver 22 configured to mix up the bits, of the extended data packet Cj. The bit interleaver 22 thus outputs an interleaved data packet C′_j=[C′_1,j . . . C′_M,j].

The FEC module 14 further comprises a constellation mapper 24. The constellation mapper 24 is configured for encoding the interleaved data packet C′_j into a FEC data packet X_j=[X_1,j . . . X_Lin,j]. A FEC data packet is a vector of constellation points of a chosen constellation diagram. The constellation points are usually represented by a complex number. The constellation diagram is for instance one of the QAM diagrams: QAM-4, QAM-16, QAM-64, QAM-256, etc. The constellation mapper 24 is able to encode each interleaved data packet with a different QAM constellation. The higher the constellation is, the more bits can be encoded in a single constellation point.

As it can be appreciated, the forward error correction module 14 is parameterized by three parameters, which each affects the quantity of data encoded in the L_in constellation points. Those parameters are: the number Nb of input data packet p_k, the code rate, and the chosen constellation diagram.

In order to control the forward error correction module, the wireless data emitting device 10 comprises a controlling means 25 configured for setting parameters such that the forward error correction module 14 only outputs data packets comprising a fixed number L_in of constellation points. In particular, the wireless data emitting device is able to set different FEC parameters for different data packets in a single frame, as it will be explained later.

Of course this same number L_in is kept during normal operation of the emitting device. The fixed number L_in may be changed by reconfiguring the controlling means 25. This is done for instance by the manufacturer.

2. Packet Interleaver

The second stage is a packet interleaver 16. The packet interleaver 16 is configured for selecting successive input sets of consecutive FEC data packets (X₁ . . . X_Nin) received from the forward correction module 14. The functions of the packet interleaver 16 are for instance performed by a calculator configured accordingly.

For each set, supposing that this set comprises N_in consecutive FEC data packets X₁ . . . X_Nin, the packet interleaver 16 is configured for distributing their constellation points so as to form N_out data packets O₁ . . . O_Nout, which will be called symbols. As it will be explained later, a symbol O_m comprises the constellation points that are intended to be emitted by the antenna 12 in a modulation symbol, for instance an OFDM symbol when using OFDM modulation.

The packet interleaver 16 is configured so that the number N_out of formed symbols is related to the length L_in of each FEC data packet by the relation: L_in=A×N_out, where A is a fixed whole number (1, 2, 3, etc.).

With reference to FIGS. 2 and 3, the packet interleaver 16 is configured for segmenting each of the N_in FEC data packets X_j into N_out fragments G^j₁ . . . G^j_Nout of a number A of consecutive constellation points. The packet interleaver 16 is further configured for forming each symbol (O_m) by, on the one end, choosing one fragment of constellation points in each of the N_in data packets and, on the other end, aggregating the N_in chosen fragments (G¹_m . . . G^Nin_m) to form the symbol (O_m). The chosen fragments are different for each formed symbol (O_m).

In this way, the constellation points of each data packet is evenly distributed amongst a plurality of N_out symbol O_m. If a symbol is lost when it is transmitted by the antenna 12, only a part of each input data packet is lost, and the lost data can be retrieved by using the redundant information of the input data packet contained in the other symbols.

Aggregating the N_in chosen groups (G¹_m . . . G^Nin_m) comprises mixing up the constellation points of each aggregated fragment with the constellation points of the other aggregated fragments, by alternating constellation points of each aggregated fragment, so as to interleave the aggregated fragments.

For instance, the constellation points in a symbol O_m are organized into a number A (same A as before) of sequences of N_in constellation points, where the j-th element of the sequence comes from the aggregated fragment of the j-th FEC data packet X_j.

Preferably, the packet interleaver 16 is configured for aggregating the chosen fragments in such a way that the constellation points a same data packet X_j are positioned in the symbols O₁ . . . O_Nout at all the different possible positions. This is obtained in the described embodiment by shifting the position, in the symbols O₁ . . . O_Nout1, of the constellation points coming from a same data packet X_j, by a constant shift step B between a symbol O_m and the next symbol O_m+1.

The shifting is made circularly, that is, when the end of the symbol O_m is attained, the remaining constellation point are placed at the start of the symbol O_m. Preferably, the shifting uses a constant shift step B.

In a preferred embodiment, the frame mapper (16) is configured for realizing the following operation:O_k,m=X_i,j, i=1 . . . L_in and j=1 . . . N_in m being the ceil operation of i divided by A,k={((i−1)%A)×N_in+(j−1)+B×(m−1)}%L_out+1,where:X_i,j is the i-th constellation point of the j-th FEC data packet (X_j),O_k,m is the k-th constellation point of the m-th symbol (O_m), and. . . % . . . denotes the modulo division, i.e. c=a % b, means that c is the remaining of a/b—therefore, c is a value from the set {0, 1, . . . , a−1}.

For instance, m is calculated with the relation: m=|i/A|, where, | . . . | denotes the ceil operation, i.e. rounding to the nearest superior integer.

Alternatively, m is calculated with the relation: m=|(i−1)/A|+1, where, | . . . | denotes the floor operation, i.e. rounding to the nearest inferior integer.

So as to insure that the groups of elements are shifted on exactly all the length L_out of the symbols O₁ . . . O_Nout, B is chosen so that B and L_out are co prime (also called relative prime). Two numbers are called co prime if their Greatest Common Divider is one.

An example of using this relation is shown on FIGS. 2 and 3, with N_in=2, L_in=8, N_out=4, L_out=4 and A=L_in/N_out=2. In FIG. 2, B is set to zero (no shifting), while in FIG. 3, B is set to one.

3. Modulation

Referring back to FIG. 1, the modulator 18 further comprises a modulation unit 26 configured for receiving each symbol O_m and for generating a modulated symbol O′_m, taking into account modulation parameters Δ. The modulation parameters Δ are for example a number of pilots, a guard interval, etc.

The modulator 18 further comprises a waveform generator 28, which generates a waveform from the modulated symbols O′_m. More precisely, each constellation point of the symbol O′_m modulates a respective carrier during a time T_symbol.

4. Setting Up

Referring to FIG. 4, a method for setting up an emitting device, in order to obtain the wireless data emitting device of FIG. 1, will now be explained. It would be appreciated that the order of the following steps can be different, without departing from the scope of the present invention, in particular for the steps that do not depend on the completion of a previous step.

In a step 100, a time frame T_frame, i.e. the duration of a frame, is chosen. Preferably, the time frame is chosen between 100 and 200 milliseconds.

In a step 110, a symbol duration T_symbol, i.e. the time during which the antenna 12 sends a symbol O′_m before sending the next one O′_m+1, is chosen. T_symbol is chosen so that there is a whole number of symbol durations in a frame duration T_frame. Preferably, T_symbol is chosen between 100 microseconds and 4 milliseconds.

The number of symbols O′_m that are sent in a frame is also deduced. It will be called in the following N_{symbols/frame}. From what precedes, one can see that there is a number N_{symbols/frame} of modulated symbol O′_m in a frame. Similarly, there is a number N_{symbols/frame} of symbol O_m in a frame.

In a step 120, the length L_max of the modulated symbols O′_m is determined as a function of the symbol duration T_symbol and the available bandwidth. More precisely, L_max is the product of T_symbol with the available bandwidth. Basically, it may be considered that the length L_max is related to the maximum number of available frequency carriers.

In a step 130, the modulation parameters Δ are chosen.

In a step 140, the length L_out of the symbols O_m is determined as a function of the length L_max of the modulated symbols O′_m and the modulation parameters Δ. In fact, the modulation parameters Δ leads to some frequency carriers being not available for transmitting data, because they are used for obtaining a robust transmission (for example, they are used for pilot carriers). This leads to L_out being smaller than L_max.

In a step 150, the number L_in of constellation points in the FEC packets is set, by configuring accordingly the controlling means.

For each frame, the forward error correction module 14 generates N_{FEC packets/frame} FEC data packets, whose data is to be emitted in the waveform W during the duration of the frame. Since there is no added information in the packet interleaver 16, N_{FEC packets/frame} is derived, in a step 160, from the relation: N_{FEC packets/frame}=N_{symbols/frame}×L_out/L_in.

From the previous relation, it can be seen that the FEC parameters (influencing L_in) and the modulation parameters Δ (influencing N_out) must be chosen such that N_{symbols/frame}×L_out/L_in is a whole number.

In a step 170, it is chosen a whole number P, which represents the portion of the N_{FEC packets/frame} FEC packets of a frame that will be interleaved at a time by the packet interleaver 16.

Accordingly, N_in is chosen so that N_in=N_{FEC packets/frame}/P, where P is a divisor of N_{FEC packets/frame}, and also of N_{symbols/frame}, so that N_out=N_{symbols/frame}/P.

When P=1, the whole frame is interleaved. On the contrary, when P is greater than 1, only a part of the frame is interleaved at a time. This is advantageous for decreasing the size of the memory needed in a receptor, for performing a “de-interleaving” operation, i.e. the inverse operation of the packet interleaver 16.

It can be appreciated that it is possible to send different services, requiring a different quality of service and thus different modulation parameters, in a single frame.

In fact, when the modulation parameters are changed, L_out changes (cf. step 140), as well as N_{FEC packets/frame} (cf. step 160). However, because of the relation L_in=A×N_out, the packet interleaver 16 will still give an optimal distribution of the N_in FEC packet between the N_out data symbols.

Moreover, because of the previous relation (L_in=A×N_out) and because L_in is fixed for every FEC data packet, the emitting device 10 always generates the same number N_{symbols/frame}=N_out×P of symbols in a frame. A receiver is then able to demodulate efficiently received symbols. In fact, the receiver stays synchronized, because the number of symbols in a frame N_{symbols/frame} remains constant. Once again, the receiver will not be affected by the fact that several different services are transmitted in a single frame.

From the previous description, it can be seen that the wireless data emitting device 10 comprises:

• • a forward error correction module 14 configured for generating input packets X_j from information packets, each input packet X_j comprising L_in constellation points of a predetermined constellation diagram, the L_in constellation points encoding a quantity of data, the forward error correction module 14 being parameterized by at least two parameters, which each affects the quantity of data encoded in the L_in constellation points,
  • a controlling means 25 configured for setting the parameters of the forward error correction module 14,
  • a packet interleaver 16 configured for generating N_out output packets O₁ . . . O_Nout from N_in input packets X₁ . . . X_Nin, each output packet comprising L_out constellation points, the packet interleaver 16 being configured for distributing the constellation points of each of the N_in packet X_j between the N_out symbols O₁ . . . O_Nout,
  • a modulator 18 configured for generating a waveform W, destined to be provided to an antenna 12, from the symbols O_m,the controlling means 25 being configured for setting first parameters for the forward error correction module 14 to generate a first output packet, and second parameters, which are different from the first parameters, for the forward error correction module 14 to generate a second output packet, the first and second parameters being such that the first and second output packets both comprise a same number L_in of constellation points.

In particular, the controlling means 25 is configured for setting parameters such that the forward error correction module 14 only outputs packets comprising the number L_in of constellation points.

Claims

1. A packet interleaving method comprising: a calculator device selecting successive input sets of consecutive input packets (X1 . . . XNin) received from a forward correction module, each input packet (Xj) being a vector of constellation points of a predetermined constellation diagram, the consecutive input packet being FEC data packets from an FEC decoder, whereinthe packet interleaving method comprising the calculator device, for each input set: generating an output set of output packets (O1 . . . ONout), each output packet (Om) being a vector of constellation points, by distributing the constellation points of each input packet (Xj) of the input set between the output packets (O1 . . . ONout) of the output set,sending the output packets (O1 . . . ONout) of the output set to a modulator configured for generating, from the output packets (O1 . . . ONout) a waveform (W), destined to be provided to an antenna, andthe input set comprising Nin input packets (X1 . . . XNin) and each of the Nin input packets (X1 . . . XNin) comprising a same number Lin of constellation points, the number Nout of output packets in the output set is related to Lin by the relation Lin=A×Nout, where A is a fixed whole number.

2. The packet interleaving method of claim 1, comprising for each input set: segmenting each of the Nin input packets (X1 . . . XNin) of the input set into Nout fragments (Gj1 . . . GjNout) of A consecutive constellation points,generating each output packets (Om) of the output set, by:choosing one fragment of constellation points in each of the Nin input packets (X1 . . . XNin),aggregating the chosen fragments (G1m . . . GNinm) to form the output packet (Om),wherein the chosen fragments are different for each formed output packet (Om).

3. The packet interleaving method of claim 2, comprising, in each output packet (Om), aggregating the chosen fragments (G1m . . . GNinm) comprises mixing up the constellation points of each aggregated fragment with the constellation points of the other aggregated fragments.

4. The packet interleaving method of claim 3, wherein the mixing up of the constellation points in an output packet (Om) comprises alternating constellation points of each aggregated fragment, so as to interleave the aggregated fragments.

5. The packet interleaving method of claim 2, comprising aggregating the chosen fragments in such a way that the constellation points a same packet (Xj), are positioned in the output packets (O1 . . . ONout) at all the different possible positions.

6. The packet interleaving method of claim 5, wherein positioning at all the different possible positions comprises shifting the position, in the output packets (O1 . . . ONout), of the constellation points coming from a same packet (Xj), by a constant shift step B between a symbol (Om) and the next symbol (Om+1).

7. The packet interleaving method of claim 1, comprising realizing the following operation: Ok,m=Xi,j, i=1 . . . Lin and j=1 . . . Nin m being the ceil operation of i divided by A, k={((i−1)%A)×Nin+(j−1)+B×(m−1)}%Lout+1,where:Xi,j is the i-th constellation point of the j-th packet (Xj),Ok,m is the k-th constellation point of the m-th symbol (Om), and. . . % . . . denotes the modulo division.

8. The packet interleaving method of claim 1, wherein the time being divided into frames of constant duration, for each frame the forward error correction module generating NFEC packets/frame input packets, whose is to be emitted in the waveform (W) during the duration of the frame: selecting successive input sets comprises the NFEC packets/frame input packets.

9. The packet interleaving method of claim 1, wherein the time being divided into frames of constant duration, for each frame the forward error correction module generating NFEC packets/frame input packets, whose is to be emitted in the waveform (W) during the duration of the frame: selecting successive input sets comprises selecting a whole number P, greater than one, of consecutive sets of Nin input packets (X1 . . . XNin) of the NFEC packets/frame input packets, where Nin is a whole number verifying Nin=NFEC packets/frame/P.

10. A non-transitory machine readable storage medium containing thereon a computer program product directly loadable into the internal memory of a digital computer, comprising software code portions for performing the steps of claim 1 when said product is run on a computer.

11. A wireless emitting device, comprising: a calculator device configured to execute selecting successive input sets of consecutive input packets (X1 . . . XNin) received from a forward correction module, each input packet (Xj) being a vector of constellation points of a predetermined constellation diagram, the consecutive input packet being FEC data packets from an FEC decoder, whereinfor each input set the calculator device performs the steps of generating an output set of output packets (O1 . . . ONout), each output packet (Om) being a vector of constellation points, by distributing the constellation points of each input packet (Xj) of the input set between the output packets (O1 . . . ONout) of the output set,sending the output packets (O1 . . . ONout) of the output set to a modulator configured for generating, from the output packets (O1 . . . ONout), a waveform (W), destined to be provided to an antenna, andthe input set comprising Nin input packets (X1 . . . XNin) and each of the Nin input packets (X1 . . . XNin) comprising a same number Lin of constellation points, the number Nout of output packets in the output set is related to Lin by the relation Lin=A×Nout, where A is a fixed whole number.