Transmission system, receiver, and interconnection network

1. Field of the Invention

The invention relates to a transmission system comprising at least a transmitter and a receiver, said receiver comprising calculation modules designed for communicating with each other via an interconnection network comprising a forward path and a return path, said network serving to realize:

unidirectional local communications between mutually neighboring calculation modules along the forward path and the return path, and unidirectional global communications between non-neighboring calculation modules.

It also relates to a receiver and an interconnection network for interconnecting calculation modules in a data communication system which comprises a forward path and return path, said network serving to realize:

unidirectional local communications between mutually neighboring calculation modules along the forward path and the return path, and unidirectional global communications between non-neighboring calculation modules.

It has numerous applications, in particular in broadband digital communication systems for increasing the communication speed and shortening the processing times between the various elements of the receiver, and for providing a programmable network architecture in dependence on the envisaged type of application.

2. Background

The article by T.-Y. Feng in the journal IEEE Computer, pp. 12-27, December 1981, entitled “A Survey of Interconnection Networks” in particular discloses an interconnection network architecture which has an annular topology, or having two dimensions, formed by a forward path and a return path. The article does not describe how the various components of the architecture are mutually interconnected, neither does it state how the communications are realized.

The present invention describes a transmission system, a receiver, and a programmable network architecture for interconnecting several components of the receiver arranged in a forward path and a return path so as to realize communications at a very high data rate capable of comprising feedback loops.

To achieve this, a system, a receiver, and a network as mentioned in the opening paragraphs are characterized in that said local and global communications are realized by means of a programmable sequence of interconnection cells which comprise memory means for storing the data traversing between two non-neighboring modules situated at least in the forward path or on the return path.

According to a major characteristic of the invention, each interconnection cell is connected to at least one calculation module, referred to as current module, and comprises multiplexing means which are connected to neighboring cells for multiplexing the data originating from different modules in the forward path and the return path, said memory means being connected to the respective multiplexing means upstream and downstream of said current module so that said data pass through at most two neighboring cells in one calculation cycle.

The following description in combination with the annexed drawings, all by way of non-limitative example, will clarify how the invention may be put into practice.

FIG. 1 is a block diagram showing an example of a receiver according to the invention.

FIG. 2 shows an example of an interconnection network according to the invention.

FIG. 3 shows examples of interconnection cells in two embodiments of the invention.

FIG. 4 is a general diagram of a transmission system according to the invention.

FIG. 1 shows a simplified block diagram of an example of a receiver according to the invention. It refers in particular to a digital TV receiver, especially for demodulating a signal which was received in the form of data samples E and for providing symbols S at the output. The receiver comprises:

a receiver block 10 for receiving an analog input signal ε and converting it into digital input samples E,

calculation modules FU 1 , FU 2 , and FU 3 which form a forward communication path,

calculation modules BU 1 , BU 2 , and BU 3 forming a return path,

an interconnection network 11 comprising:

programmable interconnection cells referenced C 1 , C 2 , and C 3 for ensuring fast communications, shown in thick drawn lines and by way of example, between various calculation modules in the forward path and the return path, and

memory means 11 , 12 , and 13 for storing the data passing between two non-neighboring modules which lie in the forward path and the return path, respectively.

The calculation modules may be various filtering units connected in a cascade arrangement for realizing, for example: an anti-aliasing filter, a Nyquist filter, and an equalizer. Several successive digital filtering operations are in fact necessary in a digital TV for achieving the demodulation function. Moreover, these demodulations render necessary feedback loops for synchronizing the receiver with the transmitter. The invention provides a modular interconnection network architecture which particularly suits these applications, which involve considerable restraints as regards the calculation power and time. It comprises programmable interconnection cells C 1 , C 2 , and C 3 provided with respective memory means 11 , 12 , and 13 for storing the data transmitted between two non-neighboring calculation modules so that they will pass through at most two neighboring cells in one calculation cycle.

FIG. 2 shows an embodiment for the receiver of FIG. 1 and the interconnection network 11 according to the invention in more detail.

It is an object of the invention to define an interconnection network which has a simple topology and is capable of operating at a high frequency. The connections in general utilize multiplexers or data buses for multiplexing the data originating from the various calculation modules over the various communication paths. Now a bus renders possible the communication of only a single data in one clock cycle. The topology of the interconnection network according to the invention is defined such that all the calculation blocks can provide and receive a data in each cycle. The communication time between two neighboring modules being one clock cycle, the transfer time for the data depends on the distance traveled by these data and thus on the number of multiplexers traversed in one clock cycle. To increase the clock speed of the network, the invention provides means for safeguarding that the data do not pass through more than two multiplexers in one clock cycle.

The interconnection network is formed by several successive interconnection cells capable of realizing all the communications required by the system. They are in particular programmable for realizing:

unidirectional local communications between neighboring calculation modules along the forward path or the return path,

unidirectional global communications between the non-neighboring calculation modules.

Since the return path can be used in the feedback loops, local communications may accordingly be realized between two neighboring modules along the feedback loop, i.e. between two mutually facing modules, one situated in the forward path and the other in the return path.

FIG. 2 illustrates an example of an application which carries out data communications, shown in thick hatched lines, between the modules FU 2 and FU 3 in the forward direction and which realizes a feedback loop involving the modules BU 2 and BU 3 in the return path. This may relate to, for example, data communications between different elements of a digital TV receiver for the purpose of channel decoding.

In the embodiment shown, each cell Ci (i=1, 2, or 3) is directly connected via multiplexers MUX to a calculation module FU i in the forward path, referred to as current forward module, and to a module BU i in the return path, referred to as the current return module. The cell comprises a total of four multiplexers MUX controlled by a control member (not shown) provided with a clock for multiplexing the data originating from the different modules FU i and BU i of the various communication lines. A first memory MD, referred to as forward memory, is connected to the two multiplexers MUX 1 and MUX 2 which are situated on the forward path, upstream and downstream of the current forward module FU i , respectively, and a second memory MR, referred to as return memory, is connected to the two multiplexers MUX 3 and MUX 4 on the return path, which are situated upstream and downstream of the current return module PU 1 , respectively.

The first multiplexer MUX 1 has a first input for receiving the data coming from the forward path, a second input for receiving the data coming from the feedback loop, and an output connected to the forward memory MD and to the current module FU i of the forward path. The second multiplexer MUX 2 has a first input for receiving the data stored in the forward memory MD, a second input for receiving the data provided by the current module of the forward path FU i , and an output for transmitting the multiplexed input data over the forward path and over the return path through the feedback loop. The third multiplexer MUX 3 has a first input for receiving the data coming from the output of the second multiplexer MUX 2 , a second input for receiving data coming from the return path of the neighboring cell, and an output connected to the return memory MR and to the current calculation module of the return path BU i . The fourth multiplexer MUX 4 has a first input for receiving the data stored in the return memory MR, a second input for receiving the data provided by the current module of the return path BU i , and an output for transmitting the multiplexed input data over the return path and over the forward path via the feedback loop.

The role of the memories MD and MR, shown as registers in the Figure, is to guarantee that the data will not traverse through more than two successive multiplexers in one clock cycle. This renders it possible to choose a high clock speed and thus to speed up the local communications. Since most communications are local in this type of application, i.e. they take place at close range between neighboring calculation modules, it is more advantageous to advance this type of communication by adapting the clock speed of the network to fast communications rather than to long-distance or global communications, which are much less frequent.

FIG. 3 shows two modified embodiments of interconnection cells according to the invention, comprising three programmable multiplexers MUX 1 , MUX 2 and MUX 3 . FIG. 3A illustrates an example of a simplified cell which can be used for applications where a calculation module is absent in the return path. The simplified cell is directly connected to only a single module in the forward path. The output of the third multiplexer MUX 3 is connected only to the register MR. The output of the register MR is connected to an input of the first multiplexer MUX 1 and to the return path of the next cell.

FIG. 3B illustrates an example of a simplified cell which can be used for applications which comprise a direct reaction between two modules FU and BU which face each other in the forward path and the return path. The input of the first multiplexer MUX 1 is directly connected to an output of a calculation module of the return path BU, and the output of the register MR is only connected to the next cell via the return path.

FIG. 4 is a general diagram of a transmission system according to the invention comprising a transmitter 41 , a receiver 42 of the type as shown in FIG. 1 , and a transmission channel 43 for supporting, for example, satellite, cable, or radio wave transmissions, etc.