Background of the Invention In the past, trainline communications systems were based on a bus of several parallel wires extending from car to car across the trainline. Each wire was dedicated to a specific task, such as throttle, braking, door open control, intercom, etc. As rail cars and remote locomotives were improved by adding more control and diagnostic features, more wires were added to the bus. As the need for more wires increased, the electrical inter-car couplers also grew more complex, expensive and less reliable. To reduce the number of wires passing through the train cars, a technique new to trainlines was applied: Time Division Multiplexing.

This technique was widely implemented in telephone networks, allowing a large number of relatively low bandwidth signals to be combined and to share a single high bandwidth channel. In US 5,351,919 to Martin, there is disclosed what is known as a "pinless coupler", namely a low leakage RF link between rail cars providing a relatively high bandwidth channel able to be used for TDM communications. By eliminating the electro-mechanical pin coupler, this technique improved reliability.

These new multiplexed trainline communications links used standard PCM signals which are divided into 64kbit/s channels. Each system used a specific channel, e. g. channel 1 was for throttle, channel 2 was for door operation, channel 3 for monitoring, channel 4 for heating control, etc. In each car, a central box called a"trainline multiplexer" (TMX) was connected to each system by a separate cable. In terms of communications protocols, each system manufacturer had its own protocol for communicating between its own system devices in other cars. This meant that for different system or manufacturer, there was potentially a different communications protocol being used. Intercommunication between various systems was difficult if not impossible.

Distributed control systems based on high-performance, low-cost microprocessor/microcontrollers have found a growing number of applications in freight train, passenger train and subway systems, as well as applications in building, home and factory automation, automobiles and security systems.. Such a distributed control and monitoring system comprises a number of local control nodes, each using a special-purpose microprocessor or microcontroller connected to the various sensing and operating devices on some particular controlled equipment (propulsion, brake, door opening, lighting, air conditioning, visual annunciator, intercom, etc.).

The local control node is located close to the controlled equipment and automates

the operation of that equipment by inputting data from the sensing devices and outputting the appropriate control signals to the operating devices. The exchange of input and output data between the control node and the devices on the controlled machinery is normally accomplished by direct connections between ports on the control node and each device.

Interconnection between control nodes supports the exchange of information between them over a train monitoring and or control network in order to provide a global control system.

Such control networking can be provided by known networks such as the communication network disclosed in US 5,537,549 issued to Gee et al.

Data networking concepts have been adopted to establish train monitoring and control networking standards. However, unlike data networks, train monitoring and or control networks are often subjected to huge bursts of relatively short and time-critical control messages. Various standards for suitable train monitoring and or control networking architectures tailored to the characteristics of the control type-small but frequent and time-critical messages and to the consideration of protocol, interoperability, command status, and topology. The LonWorks network and corresponding LonTalkX protocol are an example of such train monitoring and or control networking standard.

Other examples of control networking standards include WorldFIP, CEBus, BACnet, CAN, ISP to name a few.

In general, a networking system is usually based on a serial link in order to simplify the connectivity between nodes attached to the network. The train monitoring and or control networking standards such as the LonWorks network and corresponding LonTalk protocol also make use of serial link for the same reason. With the increasing proliferation of monitoring sensors to predict future failures on the train, and security visual communication services, demands of both control and communications networking are equally growing.

Time Division Multiplexing (TDM) has been used in digital trainline communications to combine several electrical signals in a single multiplexed signal able to feed a single digital radio channel for permitting communications between adjacent cars of a multi-car vehicle disclosed in US 5,351,919 issued to Martin. The TDM link formed the basic structure in the development of standards established for digital networks, transmission systems and multiplexing equipment. Internetworking generally means the connection of two or more networks to allow an exchange of information between the various networks and within the individual networks arranged in a variety of formats.

The OSI Reference Model calls a device to interconnect two systems not connected directly to each other a relay. If the relay shares a common layer n protocol with other systems, but does not participate in a layer n+1 protocol, it is a layer n relay. Repeater is the physical layer relay. Bridge is the datalink layer relay. Router is the network layer relay.

Gateway is any relay at layer higher than network layer. Both bridges and routers can make

forwarding or routing decisions based on information in the packet headers. A bridge differs fundamentally from a router. A bridge typically relays Media Access Control (MAC) layer (or data link layer which is layer two in the OSI model) frames and decisions are made based on information in the frame header. A router relays network layer (layer three in the OSI model) datagrams and decisions are based on information in the network layer header.

This fundamental difference affects the way each type of device operates, and consequently, the applications to which it is best suited. In practice, the terminology is far from uniform.

The term gateway is often used for any relay operating at the datalink layer, network layer, or any higher layer.

In general, gateways are mostly used for interconnecting heterogeneous networks.

They may be separated into two half-gateways. When half gateways are used, each half normally operates between the protocols used in its network and an intermediate protocol.

There are significant advantages to a standard intermediate protocol. Without a common intermediate representation, designing relays between n different networks could require a distinct mapping between each network and each of the (n-1) others, or n (n-l) mappings.

With a common intermediate protocol, though, 2n mappings suffce-one from each network to the common protocol and one from the common protocol to each network.

Relays can be manufactured as half relays that fit together in various combinations, due to the common protocol. A prior art internetworking system using a common protocol for exchanging data between a plurality of data communication networks is disclosed in US Patent 5,490,252 issued to Macera et al.

However, there is still a need for interconnection between various train monitoring and control networks via a single trainline TDM link and for an integrated controUcommunications network, to benefit the availability of a wide range of wireline and wireless transmission equipment.

Summary of the Invention It is therefore an object of the present invention to provide interconnection between control networks to provide integrated control/communications networking.

It is another object of the present invention to provide transparency in communications between control networks according to their respective original control message formats.

According to the above objects, from a broad aspect of the present invention, there is provided an internetworking relay apparatus for exchanging control messages between control networks over a TDM link transporting both control data from said control networks and communications data from communications devices, said relay apparatus connecting one of the control networks to the TDM link. The apparatus comprises means for receiving control messages in an original control message format from one of the control networks and means for converting the control messages received into a

corresponding control data stream when addressed to a control node on another one of said control networks. There are also provided TDM channellization means for transferring to the TDM link the control data stream for transmission to a corresponding internetworking apparatus connected to the other control network, and means for receiving control data streams from the TDM link sent from other ones of said control networks. The apparatus further comprises means for converting ones of the control data streams which are addressed to a control node of said one control network into corresponding control messages using the original control message format, and means for transmitting the corresponding control messages to said one control network.

More conveniently, the TDM channellization means comprise first control means connected to the control data streams receiving means, said first control means detecting when the TDM link is available for control message transmission to produce a corresponding first control signal according to a predetermined access allocation, said first control means further detecting a beginning TDM frame slot to produce a corresponding second control signal according to a reference synchronization. The TDM channellization means further comprise first data buffer means coupled to the control messages converting means and connected to the first control means for temporarily storing the control data stream whenever receiving the first control signal, and first serial interface means connected to the TDM link and the first control means, the first serial interface means being coupled to the first data buffer means to transmit the control data stream to a sufficient number of allocated TDM frame time-slots whenever receiving the second control signal.

Still more conveniently, the control data streams receiving means comprise second serial interface means connected to the TDM link for receiving the control data streams sent from other ones of the control networks in the allocated TDM frame time-slots, second data buffer means coupled to the second serial interface and to the control data streams converting means, said second data buffer means temporarily storing the control data streams, and second control means connected to the second serial interface means and the second data buffer means for controlling transfer of the control data stream to the data streams converting means.

According to another broad aspect of the present invention, there is provided an internetworking system for exchanging control messages between a plurality of control networks over a TDM link transporting both control data from said control networks and communications data from communications devices, said system comprising a plurality of relays each connecting each said control networks to the TDM link. Each relay comprises means for receiving control messages in an original control message format from one of the control networks and means for converting the control messages received into a corresponding control data stream when addressed to a control node on another one of said control networks. Each relay further comprises TDM channellization means for transferring

to the TDM link the control data stream for transmission to a corresponding relay connected to the other control network, and means for receiving control data streams from the TDM link sent from other ones of the control networks. There are also provided means for converting ones of the control data streams which are addressed to a control node of said one control network into corresponding control messages using the original control message format, and means for transmitting the corresponding control messages to said one control network.

According to still another broad aspect of the present invention, there is provided a method for exchanging control messages between control networks over a TDM link transporting both control data from the control networks and communications data from communications devices. The method comprises steps of a) receiving control messages in an original control message format from one of the control networks; b) converting the control messages received into a corresponding control data stream when addressed to a control node on another one of the control networks; c) transferring to the TDM link the control data stream for transmission to the other control network; d) receiving control data streams from the TDM link sent from other ones of said control networks; e) converting ones of said control data streams which are addressed to a control node of said one control network into corresponding control messages using the original control message format; and f) transmitting the corresponding control messages to said one control network.

Brief Description of the Drawings Fig. I illustrates the overall structure of an integrated communications/control networks system based on a TDM link using a relay in accordance with the present invention.

Fig 2 depicts the logical configuration of intemetworking between a control network and the TDM network, showing the layered structure of a relay in accordance with the present invention Fig. 3 is a block diagram of a first preferred embodiment of a relay according to the present invention, showing how the relay can be implemented using one single Neurone Chip connected with an interface module.

Fig. 4 is a block diagram of a second preferred embodiment of a relay according to the invention, showing an implementation structure using two Neurone Chips with adapted design of the interface module.

Fig. 5 is a block diagram of a third preferred embodiment of a relay according to the invention, showing an implementation of the relay with multiplexing capability.

Fig. 6a illustrates a prior art system configuration to interconnect communications and control networked between buildings.

Figs. 6b-6d illustrates various application examples of the present invention to interconnect communications and control networks between buildings.

Fig. 7a illustrates a prior art system configuration to interconnect communications and train monitoring and or control networks between train cars.

Fig. 7b illustrates a prior art application example to interconnect communications, train monitoring and or control networks between adjacent cars of a multi-car vehicle using two wireless links.

Figs. 7c, 7d, 7e and 7f illustrate an application example of the present invention to interconnect communications, train monitoring and or control networks between adjacent cars of a multi-car vehicle using a single wireline link Detailed description of preferred embodiments The invention is directed to the scheme to transport a control network over a Time Division Multiplexing (TDM) link in order to offer an integrated control/communications network. In particular, it relates to techniques to support control networks based on a Carrier-Sense Multiple-Access (CSMA) scheme known as LonWorlssS networks, on a group of assigned channel time-slots of a TDM frame structures, methods to preserve the LonTallct) Protocol, and the schemes to encapsulate the LonTaLlcX control messages, to allow dynamic access allocation and to share the TDM channel time-slots. Other channel time-slots of the TDM link can be assigned to carry voice, video communications.

The relay according to the present invention provides intemetworking using a TDM link to support an integrated network including control networks, digital voice and

video communications so that the interface to international standard digital networks is possible. In addition, by having only one single full-duplex link, the transport of control and communications services over wireless systems becomes cost-effective. In order to transport control messages such as LonWorks network control messages over a TDM link, a TDM relay or gateway is introduced. To transfer control messages from a control network or control node to the assigned TDM channel time-slots, the functions of the TDM gateway are: a) to encapsulate the control messages; b) to organize the encapsulated control messages into suitable data streams, and to transfer them over TDM channel time- slots; and c) to provide dynamic access allocation and de-allocation mechanism by sending special codes indicating the assignment of the channel time-slots to another control node.

To transfer information from the TDM channel time-slots back to the control network or control node, the TDM gateway performs the following functions: d) recognizing the assignment of the TDM slots for transmission; e) receiving the encapsulated control messages; and f) reconverting them back to their original LonTalk control messages.

Interfacing the TDM gateway to the LonWorks network or LonWorks control node follows the procedures and functions defined by the LonTalk protocol. The TDM gateway provides all the functions required by the LonTalk protocol, including three layers: physical, datalink, and network. Interface between the TDM gateway and the TDM link requires a new set of design techniques, approaches and implementation procedures.

Therefore, the present invention uses concepts, approaches, structures and techniques to design the interfacing functions between the TDM gateway and the TDM link, including associated physical, datalink, and network layers. Furthermore, the present invention also provides the internetworking functions required to interconnect the above two mentioned sets of interfacing functions within the TDM gateway.

The present invention provides a TDM gateway which is a high performance, high availability internetworking relay. The TDM gateway can be used to interconnect a plurality of individual control networks and/or control nodes of a distributed control network such as many or all of the control networks operated by a large corporation whose operations could be located in different geographic areas, via the local, national or international TDM network (s). The TDM gateway can utilize an architecture based on the standard TDM hierarchy for Pulse-Code Modulation (PCM) multiplex equipment defined by the International Telecommunication Union (ITU) for digital networks transmission systems and multiplexing equipment in order to provide an integrated control and communications network.

1. General Structure: Referring now to Fig. 1, there is illustrated a general structure of an integrated communications/control networks system based on a TDM link using a relay in accordance

with the present invention, in which system each control node 15 performs its data acquisition, monitoring and local control as part of the distributed control network. In the example shown, control networks 11 use the LonTalk@ protocol to support the exchanges of control messages between control nodes 15. For transmission of control messages, each control node 15 continuously monitors its corresponding control network 11. If the control link of the corresponding control network 11 is idle, the control messages can be sent using a CSMA scheme. The idle period between control messages comprises a fixed ßl time, and ß2 slots. If more than one nodes transmit their packets at the same time, collision occurs.

The involved nodes can detect collision and re-transmit their packets after a randomly chosen time interval. Priority feature is also available for transmission without collision.

The TDM Gateway 10 may interconnect directly control nodes 12, or various control nodes 15 via their control networks 11, to the TDM link 13. The TDM link 13 also provides channels for communications device 14. Note that the TDM link 13 comprises several time-slot channels. For example, the frame structure of a primary PCM multiplex equipment operating at 1.544Mb/s (or Tl TDM link) has 24 channel time-slots, each can accommodate one 64kb/s circuit. Another example is the frame structure of a primary PCM multiplex equipment operating at 2.048Mb/s (or E1 TDM link) which has 32 channel time- slots: two reserved for framing and signaling and the remaining 30 for traffic, each also can accommodate one 64kb/s circuit. The TDM link capacity is time-shared between various circuits to support both communications devices or control networks. For example, the 30 traffic time-slots of a 2.048Mb/s El TDM link can be shared as follows. Two time-slots with a capacity of 128kb/s are dedicated for interconnecting control networks, and the remaining 28 time-slots are used to carry data, voice and/or video circuits. Fig. 1 shows the interconnection of various local control networks 11 and communications devices 14 in one location to the single TDM link 13. The TDM link 13 is connected to another location or to a public network. By using a single link and standard TDM structure, such connection becomes simple for both wireline or wireless internetworking. Some benefits of the invention are high-performance, transparent interworking, a unified architecture which provides seamless internetworking based on a well established, widely used TDM standard, and a comprehensive manageability and testability. The above examples use primary PCM multiplex equipment to illustrate the system operation. However, the present invention can directly applies other second-order and higher-order multiplex equipment TDM standards based on 64kb/s channellization. Furthermore, the above examples use the basic time- division multiplexing scheme. Therefore, it is also applicable to other TDM links, such as basic 2B+D ISDN (Integrated Services Digital Networks), primary ISDN and broadband ISDN.

2. Logical Configuration : Referring to Fig. 1, the control network 11 is assumed to be based on the

LonTalk@ protocol defined in the LonTalk Protocol Specification, Version 0.1, Revision Dec. 12,1996, Echelon Corp., which is incorporated herein by reference, and is known as a LonWorks network, each control node 15 being a LonWorks node. On one hand, the TDM gateway 10 accesses the control network 11 in a same manner as a control node 15, using the LonTalk@ protocol defined in the LonTalk Protocol Specification. On the other hand, the TDM Gateway 10 is interfaced to the TDM link 13 in a similar manner as a communications device 14 using the standard TDM specifications defined in the Digital <BR> <BR> Networks Transmission Systems and Multiplexing Equipment, Recommendations G. 701- G. 941, International Telecommunication Union, Yellow Book, Vol. in. Fascicle E. 3, Geneva, 1981, which is incorporated herein by reference. The TDM Gateway 10 serves as an application node so that the LonWorksg) network 11 and the TDM link 13 are totally separate. The installation of the TDM link 13 can ignore the LonWorks networks 11 and vise versa. Therefore, the LonWorks networks 11 and the TDM link 13 can be installed at different times by different tools.

Referring to Fig. 2, the TDM gateway 20 provides the complete function stacks 22 of the physical layer, datalink layer including the multiple-access control (MAC) sub-layer, and network layer required to support the control protocol for communications with the control network 21. Similarly, the TDM gateway 20'provides complete function stacks 24 of the physical layer, datalink layer including the multiple-access control (MAC) sub-layer, and network layer required to support the control protocol for communications with the control network 25. In addition, the TDM gateway 20 and 20'each provides new function stacks 23, equivalent to three layers: physical, datalink, and network for communications via the TDM link 13. The TDM gateways 20 and 20'each acts as a half-relay or half- gateway. The function stacks 23 is designed for the common intermediate protocol or format operating over TDM time-slots.

The introduction of the TDM Gateways 20 and 20'allows the internetworking between various control networks 21 and 25 via a common TDM network. For example, if the control networks 21 and 25 are based on the LonTalk protocol, the control nodes 15 shown in Fig. 1 can communicate to each other via the TDM link 13 using the LonTalk protocol exactly in the same manner as their communications via the LonWorks networks 21 and 25 shown in Fig. 2 In other words, the TDM gateways 20 and 20'provide a transparent connection with respect to the LonTalk@ protocol.

3. New function stacks for three layers: Referring to Fig. 2, the new function stacks 23 covers three layers: physical, datalink (including the MAC sub-layer), and network. The physical layer comprises the bit and frame synchronization functions, the serial transmission and reception required to accessing the allocated time-slots of the TDM bit stream according to the specifications defined in the Digital Networks Transmission Systems and Multiplexing Equipment,

Recommendations G. 701-G. 941. For example, consider the 2.048Mb/s standard is used to implement the TDM link 13 in FIG. 1. The 2.048Mb/s TDM structure is based on a frame unit or time-slot of 125 micro-seconds. Each frame contains 32 time-slots : TSO, 7S'7,'4, TS'37 as shown below : FRAME FRAE#N FRAME #-1 # +I TS31TSO TS1 TS2... TS15 TS16 TS17... TS30 TS31TSO Each time-slot can accommodate 8 bits. Normally, TSO is used to carry the framing codeword that denotes the beginning ofthe frame. Time-slot TS16 is normally reserved for signaling and can carry the multiframe codeword. The other 30 time-slots are used to carry traffic. Each time-slot has 8 bits per frame unit of 125 microseconds. Therefore, each traffic time-slot can accommodate a channel of 64kb/s. Accordingly, to support a service that requires a capacity of Nx64kb/s over a 2.048Mb/s link we need to allocate N time-slots.

For example, referring to Fig. 1, to support the control networks 11 operating at 78kb/s, we can allocate 2 time-slots, say 7X7 and TS2. The remaining time-slots, TS3-TS15, tel7- TS31, can be assigned to inter-connect communications devices 14 in Fig. 1. In this example, the TDM gateways 10 in Fig. 1 all transmit and receive information in the time- slots 7X7 and TS2. The physical layer of the new function stacks 23 in Fig. 2 establishes both the bit and frame synchronization, identifies the time-slots 7X7 and TS2, and supports the serial transmission and reception of data in these two time-slots. Since the total capacity provided by 7X7 and TS2 is 128kb/s (=2x64kb/s) and the capacity required to transport the control networks is only 78kb/s, the remaining capacity of 50kb/s (=128kb/s-78kb/s) can be used for other purposes to be discussed later.

As an illustrative example, consider the transport of the control networks 11 that use the LonTalk@ protocol. The activity over the control network 11 includes control messages separated by idle periods as follows: | ... CONTROL IDLE CONTROL IDLE CONTROL MESSAGE PERIOD MESSAGE PERIOD MESSAGE A control message has the following sequence: PREAMBLE, SYNC, DATA, CRC, CODE VIOLATION.

To transport the control messages over the allocated time-slots 7S7 and TS2, each control message is segmented into 7-bit words. Each time-slot, TS1 or TS2, has 8 bits: B0, Bu,...., B7 per frame unit of 125 microseconds. We can reserve one bit, say B0, as an activity indicator to denote whether the other 7 bits, BI B7, contain words of a control message or

idle period. For example, if ? 0=l then BI B7 carries a 7-bit word of a control message; otherwise (i. e., B0=0), BI B7 contain bits of an idle period. Since 1 out of 8 bits is used as an activity indicator, the capacity provided for traffic over TS1 and TS2 is reduced to 112kb/s (=128kb/s x 7/8) which is still higher than the required capacity of 78kb/s for control network. The Datalink functions include an optional forward error control scheme to increase the transmission quality by using part the extra capacity of 34kb/s (=112kb/s- 78kb/s), and the multiple-access control procedure defined in the following.

The access to the TDM time-slots is collision-free. There are three methods for multiple-access: a) Fixed allocation: In this technique, one of the TDM gateways 10 is appointed as a master. The master gateway will monitor the time-slots allocated to transport the control messages of the networks 11, e. g., 7S7 and TS2, and inject an allocation code message whenever the TDM slots are free (i. e., during the idle period) to one TDM gateway 10.

The allocation code message contains the ID (identification) of the assigned TDM gateway 10. Upon the reception of the allocation code message, the assigned TDM gateway 10 can send its control messages. If it does not have any message to send then the master will allocate the TDM slots to another. The master gateway can allocate TDM slots to TDM gateways 10 in a round-robin manner or in a certain priority basis. b) Chain allocation: This technique does not require an appointment of a master gateway. Each TDM gateway 10 has its own ID. The TDM Gateway with the lowest valued ID will have the first right to access the TDM slots. After transmission of its control messages (or if it does not have any message to send), it will send an allocation code message in the form of the next higher valued ID indicating that the TDM gateway with this designated ID has the access right. Such process is repeated until the TDM gateway with the highest valued ID gets its access right. After transmission, it then sends the allocation code message indicating the lowest valued ID. The chain is then repeated. c) Random capture: In this scheme, during the idle periods, the TDM gateways 10 that have messages to send will first send their allocation code messages containing their ID. The first successfully received ID at the TDM link 13 will obtain the access right. Note the allocation code messages may collide and hence are not successfully received.

Retransmission of the allocation code messages after a randomly selected delay time is used. We also note that possible collision only occurs for allocation code messages, and not for control message transmission.

The selection of one of the three above discussed access methods depends on the particular application and is the choice of the users.

In the three above discussed access methods, the allocation code message is transmitted only during an idle period and can have the same format. As discussed above, during an idle period, bit BO of the 8 bits in the allocated time-slots (e. g., TS1 and TS2) is

0. The idle period can be simply indicated by all-zero bits, BO-B7=0 in each allocated time- slot. The allocation code message can have the following format: ALLOCATION CODE, ID, PARITY where ALLOCATION CODE is a 7-bit code, ID is the identification of the assigned TDM gateway, and PARITY is the parity bits of a error-correction code selected to improve the detection performance of the allocation code message.

The allocation code message is also segmented into a number of 7-bit words to be transmitted over the allocated time-slots TSI-TS2 during the idle period, i. e., B0=0.

As a result, the activity over the allocated time-slots TS1 and TS2 is as follows: IDLE TRAFFIC IDLE TRAFFIC BO : 0 1 1 1 I 0 0 0 1 B1-B7 : 0... 0 CONTROL CONTROL... CONTROL 0... 0 ALLOCATION 0... 0 CONTROL MESSAGE MESSAGE MESSAGE | CODE MESSAGE The above provides an example on the construction of the allocation code message and bits in the allocated time-slots. Other choices of the values of bits in allocated time-slots in order to denote the activity are possible. Similarly, different formats of the allocation code message can be devised.

As a special case, when there are only two TDM gateways 10 connected to the TDM link 13, the communications between these two TDM gateways over the TDM time- slots are of the point-to-point type. In this point-to-point configuration, one TDM gateway transmits to the other and vice versa. Since there is only one transmitting or receiving TDM Gateway on each side, multiple-access is not required. Consequently, the monitoring and generation of the allocation code message are not needed.

The network layer supports the transfer of control messages from the control network side to the TDM link side, or vise versa. Network layer functions include the scanning, recognizing and filtering of the destination address contained in the preamble or header of the control messages.

4. Implementation Structures: There are different ways to implement the TDM Gateway to fulfill the functions discussed above.

4.1. Single-Neuron@ Chip Structure: Referring to Fig. 3, the TDM gateway 10 is implemented by using one Neurone Chip 32 as a data processor. A transceiver 31 provides the input and output interfaces to the control network link 38. The choice of the transceiver depends on the medium used for the control network link 38. Examples of transceiver designs are given in the LonWorksg Technology Device Data Manual, DL159/D, Motorola, October, 1995, which is

incorporated herein by reference. The interface between the transceiver 31 and the Neurone Chip 32 is done using serial ports. The Neuron Chip 32 is connected to the buffer 33 via its parallel port. Software functions implemented in the Neurone Chip 32 include: a) the datalink and network layers for the LonTalk protocol as identified in the function stacks 22 in Fig. 2; and b) the network layer to access the TDM link as identified in the new function stacks 23 of Fig. 2. When the Neuron@ Chip 32 has control messages to be transfer to the TDM link 13, it first checks an handshaking signal input to its parallel port. Note that this handshaking signal is produced by a buffer 33 provided on an interface module 37 for the single-Neuron Chip 32 as part of the gateway 10. If the handshaking signal indicates that the TDM link 13 is currently available to this TDM gateway to access, the Neuron Chip 32 transfers the control messages to the buffer 33 byte by byte. Under the control of a synchronization and control unit 34, data are transferred from the buffer 33 to a serial interface 36. As an option, data can also be forward-error control (FEC) encoded by an FEC encoder/decoder (codec) 35 before being loaded to the serial interface 36. The serial interface 36, on the transmit side, contains a parallel-in/serial-out register. In each TDM frame time interval, it receives a block of data from the FEC Codec 35 or directly from Buffer 33 as discussed above. Subsequently, this data block will be serially transmitted as part of a data stream into the allocated time-slots of the TDM link 13 by the serial interface 36 under the timing control of the synchronization and control unit 34. If the optional FEC codec 35 is not used, the serial data transmitted over the allocated TDM time-slots preserves the format of the control messages. If the optional FEC codec 35 is used, each control message is simply fragmented into a number of segments, each being encoded into a codeword of a selected error correction code. The FEC codec 35 is made as an option as previously discussed. This optional FEC codec 35 is applied with an appropriate FEC code selected to improve the transmission performance if necessary.

Concurrently, the serial interface 36 also receives data stream in the allocated time-slots of the TDM link 13. Data is then transferred from the serial interface 36 to the buffer 33 and finally to the Neuron@ Chip 32. If data is FEC encoded, it is FEC decoded by the FEC codec 35 before being transferred to the buffer 33. The software of the Neurone Chip 32 performs the network layer functions to scan, recognize and filter the destination address in the header of the control messages received from the TDM link 13. Only the control messages destined to the control network link 38 are further processed and transferred, others being discarded. The optional FEC codec 35 takes care of part of the datalink layer function.

The synchronization and control Unit 34 provides TDM synchronization, multiple- access control, and timing and control. There are three levels of synchronization: bit, frame and time-slot. The bit synchronization must be done first. The receive and transmit clock signals of all the TDM gateways 10 connected to the same TDM link 13 in Fig. 1 are

synchronized to a designated reference transmit clock of one of the TDM gateways selected as part of the network configuration. After the clock signal is established, the synchronization and control Unit 34 searches for the frame codeword that denotes the beginning of the TDM frame. Since the time-slots in a TDM frame have the fixed lengths, the positions of the time-slots allocated to transport the control messages can be derived from the position of the detected frame codeword, and based on the synchronized clock signal. For providing multiple-access control function, as previously discussed, the TDM time-slots allocated to transport the control messages are shared by the TDM gateways 10 connected to the same TDM link 13 in Fig. 1. On the receive side, all TDM gateways 10 simply transfer data contained in the allocated time-slots to their respective buffer 33 in Fig.

3. On the transmit side, only one TDM Gateway can send its data over the allocated time- slots on the TDM link 13 at a given time. The multiple-access control follows one of the three possible schemes as discussed before in Section 3, regarding new function stacks for the three layers. For this, the synchronization and control unit 34 in Fig. 3 monitors the received data at the input to buffer 33 and searches for the allocation code message mentioned in Section 3. Since the allocation code message is not part of the protocol used by the control network, it will be removed by the synchronization and control unit 34 and not transferred to the Neuron Chip 32. If the allocation code message indicates its identification number then the TDM gateway has its access right. In this case, the synchronization and control unit 34 informs the Neuron@ Chip 32 so that the latter can send its data. There are two possibilities: if the Neuron Chip 32 has data to send, it does so; if not, or at the end of the data transmission, it informs the synchronization and control Unit 34 to release its access to the allocated time-slots (i. e. setting the bit BO of the time- slots to 0). Depending on the selected multiple-access scheme, as previously discussed in Section 3, the synchronization and control unit 34 may also need to generate the allocation code message. If the chain allocation scheme is used, the synchronization and control unit 34 will send an allocation code message indicating the ID of the next TDM gateway that will have the access right. If the fixed allocation scheme is used, the master TDM gateway will take care of the allocation. In the random capture scheme, the time-slots are simply set to idle period (e. g., BO=0) so that other TDM gateways can send their allocation code message if needed. To provide timing and control functions, the buffer 33 in Fig. 3 is organized as a two-port memory block. Data transfer between the buffer 33 and Neuron@ Chip 32 is clocked by the Neuron@ Chip 32. Data transfer between the buffer 33 and FEC codec 35 or serial interface 36 is clocked by the synchronization and control unit 34. In other words, the synchronization and control unit 34 generates the various timing and control signals required to clock the buffer 33, FEC codec 35 and serial interface 36.

4.2. Dual-Neurone Chip Structure: Turning now to Fig. 4, an implementation of the TDM gateway using two Neuron

Chips as data processor means will be explained. This implementation structure is used to convert a regular LonWorks Router into a TDM Gateway. In Fig. 4, the transceiver 41 is a LonWorks transceiver interfaced to the Neuron@ Chip 42 via its serial ports. The Neuron Chips 42 and 49 communicate to each other via their parallel ports. The transceiver 41, Neurone Chips 42 and 49 form the structure of a regular LonWorks Router. The transceiver 50, as part of the interface module 37 of the Dual-Neuron Chip implementation structure, is interfaced to the Neuron Chip 49 via its serial ports. Under the control of the synchronization and control unit 44, the transceiver 50 transfers serial data from the Neuron@ Chip 49 to the buffer 33 and vice versa. The serial data transfer protocol for Special-Purpose Mode Physical 1/O described in pages 22-23 of the above cited LonTalk Protocol Specification (Draft) and Neuron Chip Special-Purpose Mode Transceiver Interface Specification, LonWorks Engineering Bulletin, Echelon, October 1991, can be used, the latter specification being also incorporated herein by reference.

When the Neurone Chip 49 has control messages to be transfer to the TDM link 13, it will first check the status byte on the serial port received from the transceiver 50. This status byte is generated by the synchronization and control unit 34 to indicate whether the time- slots are currently available to this TDM gateway to access. If the status byte indicates a busy situation, the Neuron Chip 49 holds the control messages waiting for its turn.

Otherwise, it transfers the control messages to the buffer 33 byte by byte. Under the control of the synchronization and control unit 34, data are transferred from the buffer 33 to the serial interface 36. As an option, data can also be forward-error control (FEC) encoded by the FEC codec 45 before being loaded to the serial interface 36. Subsequently, data will be serially transmitted into the allocated time-slots of the TDM link 13 by the serial interface 36 under the timing control of the synchronization and control unit 44.

Concurrently, the serial interface 36 also receives data from the allocated time-slots of the TDM link 13. Data is then transferred from the serial interface 36 to the buffer 33 and finally to the Neurone Chip 49. If data is FEC encoded, it is FEC decoded by the FEC codec 35 before being transferred to the buffer 33. The software of the Neuron@ Chips 49 and 42 perform the network layer functions to scan, recognize and filter the destination address in the header of the control messages received from the TDM link 13. Only the control messages destined to the control network link 38 are further processed and transferred, others being discarded. Such software is similar to that of the regular LonWorks@ Router.

The optional FEC codec 35 takes care of part of the datalink layer function. The synchronization and control unit 34 provides: a) the timing, bit, frame, and time-slot synchronization required to support the physical layer functions of the TDM link; b) the MAC functions to access the allocated time-slots and to provide the status information to the Neurone Chips 49; and c) the timing and control signals required to operate the Buffer

33, FEC codec 35, and serial interface 36. It is pointed out that the time-slots allocated to transport the control networks over the TDM link 13 can be accessed by various TDM gateways using one of the three possible schemes discussed in Section 3 above regarding new function stacks for the three layers above. The allocation code message is monitored and generated by the synchronization and control unit 34 in a similar manner as previously discussed in Section 4.1 above.

5. Transmission of control messages over TDM time-slots For 1.544Mb/s T1 or 2.048Mb/s E1, each TDM time-slot has a capacity of 64kb/s.

The selection of the number of time-slots assigned to transport the control messages depends on the system capacity requirement. For example, to support a 78kb/s control network, two TDM time-slots are needed. The effective capacity provided by two time slots is 128kb/s. Excessive capacity of 128kb/s-78kb/s = 50kb/s can be used for forward error-correction (FEC) coding to improve the transmission performance and for additional signaling functions to provide the system robustness (e. g., as previously discussed in Section 3, one of eight bits in each allocated time-slot is used as an activity indicator). A control message includes the following sequence: Preamble (Bit and Byte Sync's), Data, Cyclic redundancy Code (CRC), Code violation. The Data portion contains source and destination addresses, control information and other attributes. The complete LonTalkX formats for control messages are described in the above cited LonTalk Protocol Specification (Draft). For simplicity and transparency, the control message format is preserved in the transmission over the TDM time-slots. An example of the transparent transmission of the control messages was discussed in Section 3. As previously discussed, the TDM gateways access the allocated TDM time-slots on a collision-free basis. A TDM gateway only sends its control messages in the shared time-slots when it obtains its assignment. The TDM gateway releases the time-slots after it has finished its transmission.

This is done by sending a release code after the last transmission of the control message.

The time-slots then are idle and available for allocation. Three allocation schemes have been discussed above in Section 3 regarding new function stacks for three layers above. It is to be understood any other equivalent scheme of access allocation known in the art can be used without departing from the present invention.

The following example uses the chain allocation scheme (i. e., scheme b) described in Section 3 above) to illustrate in detail the releasing and allocation steps. In this scheme, there is no master gateway. Each TDM Gateway has its own identification (ID). At the initial of the system operation, the TDM Gateway with the lowest valued ID will have the first right to access the TDM slots. After transmission of its control messages (or if it does not have any message to send), it will send a release code message and subsequently, an allocation code message representing the next higher valued ID indicating that the TDM Gateway with this designated ID has the access right. Such process is repeated until the

TDM gateway with the highest valued ID gets its access right. After transmission, it then sends an allocation code message indicating the lowest valued ID, the chain being then repeated. Consider the case in which TDM gateway #n is currently transmitting its control messages. At the end of its transmission, it sends a release code message and subsequently an allocation code message indicating that TDM Gateway # (n+l) has the access right. The TDM gateway #n then monitors the channel and waits for a time interval called waiting window. The duration of this waiting window is programmable by the user and is set by software as a system configuration parameter. The TDM gateway # (n+l) is expected to respond within this waiting window. There are three following possibilities: 1) If the TDM gateway # (n+I) has control messages to send, it first sends an acknowledge Code message containing its ID and then start the transmission of the control messages, the acknowledge code message being simply the echo of the allocation code message; 2) If the TDM gateway # (n+l) does not have any control message to send at this moment, it sends an Allocation Code containing the ID of the TDM Gateway # (n+2) ; and 3) If the TDM gateway # (n+I) is not operational then it cannot response. Upon the expiration of the Waiting Window (i. e., time-out situation), The TDM gateway #n sends another Allocation Code but containing the ID of the TDM gateway # (n+2). In other words, the channel is now assigned to the TDM gateway # (n+2). Such an approach provide the robustness of the scheme in the case of failure of any TDM gateway in the system.

6. Multiplexing equipment: The TDM gateway can also be designed as part of a TDM multiplex equipment to provide an additional support on communications devices. Referring to Fig. 5, the interface module 37 is identical to that described in Section 4.1 above with reference to Fig. 3. The synchronization and control unit 34 of the module 37 already includes functions necessary to perform TDM bit, time-slot, and frame synchronization as well as associated timing and control signals. These functions and associated signals can also be used to operates a TDM multiplexer 52. In other words, by adding the TDM multiplexer 52, the TDM gateway 10 can support connections for communications devices which operate on the basis of 64kb/s channellization. Examples of such communications devices include 64kb/s PCM voice circuits, Nx64kb/s digital video connections (e. g., 384kb/s) and other voice, video and data circuits operating at rate of multiples of 64kb/s. In Fig. 5, the interface module 37 based on that illustrated on Fig. 3 is given for illustration. It can also be based on that illustrated on Fig. 4.

7. Examples of applications: Inter-building connection The current invention provides interconnecting control nodes of one control

network or various control networks via a standard TDM link. It facilitates the interface to the existing standard digital networks and provides a cost-effective solution to both wireless and wireline connections. Figs. 6a to 6d illustrates some application examples to interconnect communications and control networks between buildings. Referring to Fig.

6a, two separate wireless links 64 and 70 are used to inter-connect communications devices 53,53'and control networks 65,65'between two buildings A and B. In both buildings, various communications devices, respectively 53 and 53', are connected to a respective standard TDM multiplexer (MUX) 61 and 61'. Radio transceivers 62 and 62'are used to inter-connect multiplexers 61 and 61'respectively. The interface between MUX 61 or 61' and radio transceiver 62 or 62'uses a TDM standard such as T1 or E1 or higher-capacity DS3, both MUX 61,61'and radio transceiver 62 or 62'are widely available communications products from various manufacturers. Since the control networks 65 and 65'do not use a standard TDM format, the inter-connection of the control networks 65 and 65'between two buildings has to use a different set of radio transceivers 63 and 63'.

Although such a radio transceiver 63 or 63'is commercially available, two separate wireless links 64 and 70 are employed in this case. Two sets of radio transceivers and two sets of wireless channels are required.

Referring now to Fig. 6b, the TDM gateways 66 and 66'in accordance with the present invention are used to transport control messages of the control networks 65 and 65'over a single TDM wireless link 64. The output of each TDM gateway 66 and 66'can be directly connected to respective standard TDM MUX 61 and 61'. The inter-connection between the two buildings A and B can be implemented with only one wireless link 64 with a single set of standard TDM radio transceivers 62 and 62'. Compared to Figure 6a, the present invention provides a more cost-effective solution. Both cost and spectrum requirement are greatly reduced. In Fig. 6c, it can be seen that the inter-connection between the two buildings A and B can also use one single wireline link 67 using a TDM standard such as T1, El or HDSL. Figs. 6b and 6c show application examples of the present invention in private networks where wireless or wireline links belong to a private owner such as a company.

Since TDM standards are widely used in public digital networks, the present invention is also applicable to inter-connections of distant control networks via the existing public network 69 as shown in Fig. 6d. Links 68 connect the TDM Multiplexers in different buildings A, B, C, and D to their nearest Central Offices (CO). These links can be wireline or wireless. The connections between the CO's are parts of the public network.

Trainline communications system The invention provides interconnecting control nodes of one train monitoring and or control network or various train monitoring and or control networks via a standard

TDM link. It facilitates the interface to existing standard digital networks and provides a cost-effective solution to both wireless and wireline connections. Figs. 7a to 7e illustrates some application examples to interconnect communications and train monitoring and or control networks between cars according to the invention.

In the prior art, referring to Figs 7a and 7b, two separate wireless links 64 and 70 are used to inter-connect communications devices 53,53'and train monitoring and or control networks 65,65'between two cars A and B. In both cars, various communications devices, respectively 53 and 53', are connected to a respective standard TDM multiplexer (MUX) 61 and 61'. Radio transceivers 62 and 62'are used to inter-connect multiplexers 61 and 61'respectively. The interface between MUX 61 or 61'and radio transceiver 62 or 62' uses a TDM standard such as T1 or E1 or higher-capacity DS3, both MUX 61, 61'and radio transceiver 62 or 62'are widely available communications products from various manufacturers. Since the train monitoring and or control networks 65 and 65'do not use a standard TDM format, the inter-connection of the train monitoring and or control networks 65 and 65'between two cars uses a difFerent set of radio transceivers 63 and 63'. Although such a radio transceiver 63 or 63'is commercially available, two separate wireless links 64 and 70 are employed in this case. Two sets of radio transceivers and two sets of wireless channels are required.

Referring now to Figs. 7c, 7d and 7e, the TDM gateways 66 and 66'in accordance with the present invention are used to transport control messages of the train monitoring and or control networks 65 and 65'over a single TDM wireless link 64 The output of each TDM gateway 66 and 66'can be directly connected to respective standard TDM MUX 61 and 61'. The inter-connection between the two cars A and B can be implemented with only one wireless link 64 with a single set of standard TDM radio transceivers 62 and 62'.

Compared to Figure 7a, the present invention provides a more cost-effective solution. Both cost and spectrum requirements are greatly reduced. In Fig. 7c, it can be seen that the inter- connection between the two cars A and B can also use one single wireline link 67 using a TDM standard such as T1, E1 or HDSL.

Fig. 7f illustrates some of the control and monitoring components, such as propulsion or throttle control, brake control, air pressure control, and temperature monitoring. Communications components include an intercom and a close circuit television device. The inter-car connection may be wireless 64 or wireline 67.

It is to be understood that it is within the ambit of the present invention to cover any obvious modification of the above described embodiments, provided it falls within the scope of the appended claims.