Wireless real-time transmission system

Abstract

Method for a broadcast-type wireless real-time transmission system comprising at least two nodes (N1, N2), each node (N1, N2) having an associated transmitting device (TX) and receiving device (RX), the real-time transmission system having a logical bus structure, a first node (N1) being logically positioned in front of a second node (N2), each of the at least two nodes (N1, N2) having a channel quality matrix (G1, G2), the channel quality matrix (G1, G2) having at least one link quality to neighboring nodes (N1, N2), comprising the steps: • Sending a token (T) from the first node (N1) to the second node (N2), wherein the token (T) contains information about at least parts of the channel quality matrix (G1) of the first node (N1), • Receiving the token (T) from the first node (N1) to the second node (N2), where the second node (N2) receives parts of the channel quality matrix (G1) of the first node (N1), • Determining the channel quality between the receiving node and the sending node, • Revising the channel quality matrix (G2) of the receiving node (N2) using the received parts of the channel quality matrix (G1) of the first node (N1), • Determine, using the revised channel quality matrix (G'2) of the second node (N2), whether a transmission of data or tokens (T) to a third node (N3), which is logically downstream of the second node (N2), is possible directly with a given probability, • If not possible with a given probability, determine one or more relay nodes.

Description

[0001] The invention relates to a wireless real-time transmission system. Background of the invention

[0002] There is currently a growing need for machine-to-machine (M2M) communication. This involves individual devices or assemblies communicating with each other, for example within a larger system, to implement process or safety control.

[0003] However, this results in specific requirements that are not easily defined. For example, in process control and especially in safety engineering, it is generally necessary to meet a so-called "hard real-time" requirement. Within such requirements, a relatively small amount of user data often needs to be transmitted to a receiver within a specific timeframe, typically in the range of a few milliseconds or less. Furthermore, a very high level of fault tolerance is also a requirement, so that the probability of a failure is often in the range of 10 -5 or must be below that.

[0004] Existing systems in this area rely on wired communication. However, the wiring effort and the associated reliability issues are considerable. Therefore, wireless systems would be desirable. However, current wireless transmission systems cannot meet the required reliability and real-time specifications, or can only do so inadequately. These wireless transmission systems are generally based on so-called CSMA protocols, which grant a station access to the transmission medium when it detects that the medium is de-energized. If power is detected, the station waits for a certain safety interval and checks again. With each unsuccessful access attempt, the safety interval is increased. This approach allows for very simple implementation and requires minimal configuration.On the other hand, this approach is highly unfair and cannot reliably meet the requirements for maximum delay and minimum transmission rate.

[0005] In wired communication, there are various ways to implement channel access in a shared transmission medium for a group of nodes. One class of these channel access mechanisms is called "token passing." With these mechanisms, channel access is governed by possession of a logical element, the token. Only the node that possesses the token may use the channel resources for a predetermined maximum time. After that, the node must pass the access to its successor. Such a token-passing-based communication network has a predetermined maximum number of nodes. By limiting the number of nodes and the maximum holding time of a token per node, a predetermined token rotation time can be ensured. This guarantees that each node has at least one communication window for its own transmissions within the token rotation time.This is often referred to in the literature as ensuring QoS requirements [see X. Sun, Y. Zhang, J. Li, “Wireless Dynamic Token Protocol for MANET”, International Conference on Parallel Processing Workshop ICPPW, Sep. 2007].

[0006] In the literature, various methods exist for physically connecting nodes [see J. Zhang, Kh. Liu, X. Shen, “A Novel Overlay Token Ring Protocol for Inter-Vehicle Communication”, IEEE International Conference on Communications ICC, pp. 4904-4909, May 2008]: Most commonly, a distinction is made between bus structures (Token Bus) and ring structures (Token Ring). In a bus structure, all nodes can communicate directly with each other; that is, all nodes are connected to all other nodes. In a physical ring structure, each node has two fixed communication partners (predecessor and successor) resulting from the physical connection. However, the logical structure of the network is decoupled from the previously described, actual physical connection. According to current technology, the logical structure is based exclusively on a ring principle [see M. Ergen, DLR. Sengupta, P.Varaiya, “WTRP - Wireless Token Ring Protocol”, IEEE Transactions on Vehicular Technology, Nov 2004], in which the token is exchanged between nodes in a ring sequence. If a node wants to join a logical ring, it must wait for an invitation from the existing, incomplete ring. This changes the predecessor and / or successor relationship at up to three nodes.

[0007] In addition to such a method, other channel access methods can also be used [see I. Liu, F. Takawira, Hj. Xu, “A Hybrid Token-CDMA MAC Protocol for Wireless ad Hoc Networks”, IEEE Transactions on Mobile Computing, pp. 557-569, May 2008].

[0008] The Japanese patent application JP2007-13825A discusses making station membership conditional, thereby improving the channel quality of the predecessor-successor relationship. The data used in this process are stored locally in the form of a channel quality matrix. Nodes can also be removed from the ring.

[0009] A critical element of this channel access mechanism is the existence of the token. Once lost, it must be regenerated. Most state-of-the-art mechanisms use timers and timeout events for regeneration. If the token hasn't returned to a node within the token rotation time, that node creates a new token during a timeout event, sends its own payload, and then sends the token to the successor or even the successor of the successor, thus excluding the direct successor from the ring. This is possible because transmission errors are relatively rare in wired networks. In such networks, it's more likely that, in the event of a repeated timeout, a node was shut down before it could disconnect from the network.

[0010] However, transmission errors are significantly more frequent in wireless networks, especially than node shutdowns. Nevertheless, token-passing mechanisms have fundamental problems [M. Ergen, DLR. Sengupta, P. Varaiya, “WTRP - Wireless Token Ring Protocol”, IEEE Transactions on Vehicular Technology, Nov 2004], which is why they have rarely been implemented in practice in wireless networks in the past, due to the high frequency of errors. This issue is significantly mitigated by the present invention, which will be described in more detail below.

[0011] It is therefore an object of the invention to provide a wireless real-time transmission system that avoids one or more disadvantages of the prior art.

[0012] The problem is solved by a method for a wireless real-time transmission system of the broadcast type according to claim 1. Further advantageous embodiments are the subject of the dependent claims and the further description.

[0013] The invention will be explained in more detail below with reference to the figures.

[0014] These show: Fig. 1 an exemplary system according to embodiments of the invention, Fig. 2 an exemplary schematic design of a node or a central instance according to embodiments of the invention, Fig. 3 an exemplary state diagram according to embodiments of the invention.

[0015] In Fig. Figure 1 shows an exemplary system according to different embodiments of the invention, and it will be discussed in more detail below whether certain components are optional or necessary.

[0016] In Fig. Figure 1 shows a general transmission system with several nodes N1, N2, N3, and N4. At least two nodes are necessary for the functioning of a transmission system according to the invention, although the advantages become more apparent with three or more nodes.

[0017] A node N can, for example, be physically integrated into a transmit / receive module or simply represent a logical arrangement of a functional arrangement. An example of a physical arrangement with respect to a node N1 is shown in Fig. Figure 2 shows a node N with an associated transmitter TX and a corresponding receiver RX. This receiver RX and the transmitter TX can also be part of an I / O module, which is shown as a dashed frame around RX and TX. For example, the transmitter TX and the receiver can be used for communication in a specific frequency band, such as an ISM band.

[0018] The node N1 in Fig. 2 further includes an assigned and suitable processing unit (CPU), such as a microcontroller, microprocessor, FPGA, ASIC, etc., to execute the steps of the procedure. Furthermore, the processing unit (CPU) can be assigned an internal or external memory (MEM) to store a channel quality matrix (G1) or at least parts of a received channel quality matrix (G2). Any suitable rewritable memory, such as RAM, can serve as the memory.

[0019] In a transmission system according to the invention, a logical bus structure is assumed. This is represented by the arrows with solid lines in Fig. Figure 1 illustrates this as an example. It should be noted that this bus structure does not necessarily have to be fully implemented in every case, but merely represents a suggested structure, as will be explained later. Without loss of generality, the logical bus can, for example, be implemented as a token bus or a token ring.

[0020] In the following, it is assumed that a first node N1 is logically positioned before a second node N2.

[0021] Each of the nodes N1, N2 carries a channel quality matrix G1, G2. A channel quality matrix G1, G2 has at least one connection quality to the respective neighboring node.

[0022] That is, in a system with four nodes as in Fig. As illustrated by example in Figure 1, the channel quality matrix of the first node N1 has at least one connection quality to the neighboring node N2, while the channel quality matrix of the second node N2 has at least one connection quality to the neighboring node N3, and so on. Preferably, further connection qualities are also included, e.g., to all nodes that are within the reception range of each node.

[0023] For example, nodes N1 and N3 are located in the receive range E2 of node E1, while only node N4 is located in the receive range E3 of node N3, and nodes N1 through N3 are located in the receive range E4 of node N4. Accordingly, the respective channel quality matrices can also include these connection qualities located in the receive range, but are not limited to them. Furthermore, it is possible to also consider connection qualities between other nodes. For example, node N1 may have connection qualities between nodes N2 and N4, and vice versa, in its own quality matrix. Without loss of generality, the connection quality can be determined based on a single value or averaged over a number of values.

[0024] The system presented in the invention is token-based. This means that in an undisturbed system, only the node possessing the token may transmit user data. To provide the fastest possible system, the token can be passed on immediately if no user data is present at a node. Alternatively, the respective nodes can be configured to pass the token on only after a certain period of time has elapsed.

[0025] In the following, it is assumed that the first node N1 sends the token T to the second node N2, where the token T contains information about at least parts of the channel quality matrix G1 of the first node N1.

[0026] For example, all or some of the connection qualities determined by node N1 can be transmitted here. This is particularly advantageous because the token T itself represents a very small unit of information, but in practice is contained within a larger data packet that would otherwise be unused. Therefore, the amount of additional information can be limited by the size of the data packet, so that not all available connection qualities are delivered to node N2, but only a selection of, for example, neighboring or (immediately) subsequent nodes that is favorable for the subsequent process. Furthermore, it is possible to transmit the entire channel quality matrix across multiple rotations of the token.

[0027] The token T is now received, for example, by the second node N2 – the logically subsequent node. This gives the second node N2 parts of the channel quality matrix G1 of the first node N1.

[0028] However, other nodes, such as node N4, can also receive the token T without being named as a destination. This is in Fig. 1. For example, the reception areas E are represented. It can be seen that in reception area E3 of node N3 - represented as a solid ellipse - node N1 is not included, while in reception area E2 of node N2 - represented as a dotted ellipse - and in reception area E4 of node N4 - represented as a dashed ellipse - node N1 is included.

[0029] Each receiving node can now determine the channel quality between itself and the sending node based on the transmitted token. That is, in the example above, nodes N2 and N4 can determine the channel quality between N1 and themselves. Without loss of generality, the connection quality can be determined based on a single channel quality value or averaged over a number of channel quality values.

[0030] Furthermore, each receiving node - here nodes N2 and N4 - can now revise the respective channel quality matrix - nodes G2 and G4 - using the received parts of the channel quality matrix G1 of the first node N1,

[0031] The node N2, intended as the actual recipient of the token T, can now use the revised channel quality matrix G'2 to determine whether a transmission of data or tokens to a third node N3, which is logically downstream of the second node N2, is possible directly with a given probability.

[0032] For example, it would not be possible to send a message directly from node N3 to node N1. Node N3 would have to use the relay node N4 to send a message to destination N3, with node N4 then relaying the message onward.

[0033] If sending data or tokens is not possible with a predefined probability, one or more relay nodes are determined. This means a repetition scheme is determined dynamically and ad-hoc based on a (local) channel quality matrix, whereby the relay node(s) are generally selected using adapted routing methods, e.g., with the lowest probability of failure as the target metric. The maximum relaying level can be taken into account here.

[0034] Both the determination of the revised quality matrix and the selection of one or more relay nodes can be outsourced to a central instance (ZI). This could be, for example, a bus master. The connection to such a central instance can also be wireless. These connections are in Fig. 1. The central instance can be represented by dashed arrows, just like a node, according to Fig. 2 and be designed as described above.

[0035] In an advantageous embodiment of the invention, it can further be provided that another node, e.g. node N3, receives the token T, which was sent from the first node N1 to the second node N2. The further node, here node N3, can then check whether it should act as a relay node for the second node.

[0036] Whether a node should act as a relay node can be communicated, for example, in the token T itself.

[0037] If node N3 is to act as a relay node, node N3 checks, for example, based on further received events, whether the token T has been forwarded. If the token T has not been forwarded, node N3 would forward the token T. That is, here N3 would forward the token T to node N2 if N3 determined that N2 was not forwarding the token and therefore had to assume that node N2 had not received the token T from node N1.

[0038] Alternatively or additionally, the first node N1 can check whether the token T has been forwarded from the second node N2 to the third node N3, which is logically downstream from the second node N2, within a predetermined time window. If the token T has not been forwarded, the first node N1 sends the token T to one or more relay nodes.

[0039] While previously only the token T itself was used for evaluation, the transmission of user data can now also be used to advantage to update the channel quality matrices.

[0040] In the following, it is assumed that node N1 possesses the token T and therefore the authorization to send payload data.

[0041] Node N1 now sends user data to a second node N2, optionally assigning a sequence number as an identifier to the user data, which is optionally transmitted along with the user data. A sequence number allows, for example, the filtering out of duplicates.

[0042] This user data is received, for example, at the second node N2.

[0043] However, other nodes, such as node N4, can also receive the payload without being named as the destination. This is in Fig. 1. For example, the reception areas E are represented. It can be seen that in reception area E3 of node N3 - represented as a solid ellipse - node N1 is not included, while in reception area E2 of node N2 - represented as a dotted ellipse - and in reception area E4 of node N4 - represented as a dashed ellipse - node N1 is included.

[0044] Each receiving node can now determine the channel quality between itself and the sending node based on the transmitted user data. This means that, in the example above, nodes N2 and N4 can determine the channel quality between N1 and themselves. Without loss of generality, the connection quality can be determined based on a single channel quality value or averaged over a number of channel quality values.

[0045] The user data could also be received at a third node N3 in the same way as before.

[0046] If node N3 is to act as a relay node, it checks, for example, based on further received events, whether the payload has been forwarded. If the payload has not been forwarded, node N3 would forward it. That is, N3 would forward the payload to node N2 if it determined that N2 had not forwarded the payload and therefore had to assume that node N2 had not received the payload from node N1.

[0047] Alternatively or additionally, the first node N1 can check whether the payload data has been forwarded from the second node N2 to the third node N3, which is logically downstream from the second node N2, within a predetermined time window. If the payload data has not been forwarded, the first node N1 sends the payload data to one or more relay nodes.

[0048] Preferably, the relay node(s) within the transmission of data or token T in a transmission to another node are named, as this can increase the response speed.

[0049] By using suitable methods known from the prior art, the bus can be dynamically extended by adding further nodes or reduced by removing existing nodes.

[0050] Depending on the desired level of security for the transmission, various other mechanisms may be provided.

[0051] For example, it may be provided that the receipt of a transmission addressed to a specific node is explicitly or implicitly confirmed by that node.

[0052] Another example is that the data security measures include a checksum or other safeguards, such as appropriate code selection, so that transmission errors can not only be detected but also corrected. These safeguards can be dynamically configured depending on the expected transmission probability.

[0053] Furthermore, it is possible, for example, to adjust the modulation rate and / or code rate to be used for the transmission depending on the previously determined number of relay nodes, in order to make optimal use of the time available for transmission.

[0054] For example, N1 can send tokens / data to N2 and has 10 ms to do so, with a potential additional node in the network acting as a relay. One approach could be to send data at a fixed modulation / code rate within a 5 ms window, so that if N2 doesn't respond in the second 5 ms window, the relay node can transmit within the 10 ms window. Alternatively, data at a lower modulation rate or higher code rate could be transmitted to N2 without the relay node, utilizing the full 10 ms. Dynamically determining the optimal relays and / or the number of relays can account for changes in reliability associated with rate variations.

[0055] The invention can also readily benefit from adaptive transmission schemes, such as those from DE 10 2011 001 911 A1.

[0056] Advantageously, the individual nodes notify the respective user whether a transmission is or could be faulty.

[0057] The novel approach presented by the invention thus addresses the issue at the medium access level. In contrast to typical (wireless) communication systems, which are based on carrier sense multiple access (CSMA) and collision detection (CD), the invention's approach is token-based. Unlike the CSMA approach, in a token-based approach, access to the transmission medium is granted when a station receives a special, exclusive authorization, a so-called token. This allows the token-based approach – also known as TPMA – to guarantee a minimum data rate and a maximum latency, with the resulting reduction in transmission efficiency due to the necessary token transmission being comparatively minor.

[0058] In contrast to query-based approaches, the token-based approach presented here allows for a decentralized transmission scheme and thus avoids the so-called single-point-of-failure problem, which occurs when the coordinating instance fails. Therefore, its fault tolerance is higher.

[0059] The presented system allows for the implementation of both pre-configured and dynamically changing topologies. For example, the system can be based on a Wireless Token Ring Protocol (WTRP), where the exact topology can be implemented as either a token ring or a token bus, e.g., by defining the ring structure at a logical level – in this example, the Fig. 1 N1→N2→N3→N4→N1 ... - remains unchanged.

[0060] A finite state machine, which describes an approach according to the invention, is in Fig.Figure 3 illustrates this. The states can be divided into different categories. The states "Sending", "Monitoring", and "Idie" constitute the core of the protocol, while the other states can optionally be used for the orderly addition ("Joining") or orderly departure ("Leaving") of a node into the respective bus topology, as well as for handling error states.

[0061] Where the system is referred to above as a "broadcast type", this means that a node can receive tokens T or payload data from at least some of the other nodes in the topology.

[0062] The invention improves the failure probability by exploiting the broadcast property of the medium. Current channel states between all (or only some of) nodes can be estimated based on the token or user data transmission. This can be achieved by utilizing either the preamble of a token packet or a user data packet (estimating the channel direction from the sender to all current receivers), or the token itself can contain channel state information in digitally encoded form.

[0063] Furthermore, the invention allows for the proactive transmission of user and control information via alternative paths, i.e., via other nodes located in the same network, thereby increasing the degree of diversity. It also enables the dynamic determination of the optimal repetition scheme (number and sequence of nodes forwarding information) based on channel information derived from the necessarily exchanged control information (tokens). In addition, the available time for transmitting user and control information can be divided into time slots depending on a defined maximum relaying level. Within each time slot, a different node (or group of nodes) has the opportunity to transmit the information repeatedly. Moreover, the invention allows for the transport of user and control information using an adaptive coding and modulation method, which is determined by the maximum relaying level.The relaying level depends on limiting the length of all time slots to the total available time (without applying this procedure).

Claims

[1] Method for a broadcast-type wireless real-time transmission system comprising at least two nodes (N1, N2), each node (N1, N2) having an associated transmitting device (TX) and receiving device (RX), the real-time transmission system having a logical bus structure, a first node (N1) being logically positioned in front of a second node (N2), each of the at least two nodes (N1, N2) having a channel quality matrix (G1, G2), the channel quality matrix (G1, G2) having at least one link quality to neighboring nodes (N1, N2), comprising the steps: • Sending a token (T) from the first node (N1) to the second node (N2), wherein the token (T) contains information about at least parts of the channel quality matrix (G1) of the first node (N1), • Receiving the token (T) from the first node (N1) to the second node (N2), where the second node (N2) receives parts of the channel quality matrix (G1) of the first node (N1), • Determining the channel quality between the receiving node and the sending node, • Revising the channel quality matrix (G2) of the receiving node (N2) using the received parts of the channel quality matrix (G1) of the first node (N1), • Determine, using the revised channel quality matrix (G'2) of the second node (N2), whether a transmission of data or tokens (T) to a third node (N3), which is logically downstream of the second node (N2), is possible directly with a given probability, • If not possible with a given probability, determine one or more relay nodes. [2] The method of claim 1, comprising the steps of: • Receiving the token (T) from the first node (N1) directed to the second node (N2) at the third node (N3), • Checking on the part of the third node (N3) whether the third node (N3) should act as a relay node for the second node (N2), • If yes, check whether the token (T) has been forwarded based on further received events, and if the token (T) has not been forwarded, send the token (T). [3] Method according to claim 1 or 2, comprising the steps: • The first node (N1) checks whether the token (T) has been passed from the second node (N2) to the third node (N3), which is logically downstream from the second node (N2), within a predetermined time window. • If the token (T) was not forwarded, the token (T) is sent again from the first node (N1) to the relay node. [4] Method according to claim 1, 2 or 3, comprising the steps: • Sending payload data from a first node (N1) to a second node (N2), where the payload data is optionally assigned a sequence number as an identification feature, which is optionally transmitted with the payload data, • Receiving the user data from the first node (N1) directed to the second node (N2). [5] The method of claim 4, comprising the steps of: • Receiving the user data from the first node (N1) directed to the second node (N2) at the third node (N3), • Checking on the part of the third node (N3) whether the third node (N3) should act as a relay node for the second node (N2), • If yes, check whether the payload has been forwarded based on further reception events; if the payload has not been forwarded, send the received payload. [6] Method according to any of the preceding claims, wherein the relay node(s) within the transmission of data or token (T) to the second node (N2) is named. [7] Method according to any of the preceding claims, wherein nodes (N) can be dynamically inserted into or removed from the bus system. [8] Method according to any of the preceding claims, wherein the receipt of a transmission addressed to a specific node (N) is confirmed. [9] Method according to any one of the preceding claims, characterized by , that an erroneous transmission is provided as an error message to a user of the procedure. [10] Method according to any one of the preceding claims, characterized by that the channel quality is averaged over a number of reception events. [11] Method according to any one of the preceding claims, characterized by, that the selection of one or more relay nodes is made by a central control instance (CI). [12] Method according to any one of the preceding claims, characterized by , that depending on a previously determined relay node, the modulation rate and / or code rate to be used for the transmission is adjusted.

Images