Patent Application
20110149946
The present invention relates to a method for synchronizing at least two nodes in communication with at least one terminal in a wireless network and to a device for implementing such a method.
The invention relates more particularly to the synchronization of home base stations (HeNBs). Such home base stations are deployed, for example, for providing a seamless wireless network inside houses or in a building of a company, a mall, a station or more generally any large building. Thus, in such a network, the number of HeNBs is large.
According to the 3GPP-LTE (“3rd Generation Partnership Project Long Term Evolution”) standard, in a typical deployment of HeNBs, the nodes of the wireless network are activated in an uncoordinated way and non-synchronized one from each other.
However, it would be advantageous to synchronize such home base stations.
The first advantage of this synchronization is an improvement of the overall performance of the wireless network.
For example, a terminal in the network will perceive the synchronization signals of the different HeNBs as perfectly orthogonal which improves, for example, the detection algorithms used for mobility purposes.
A second advantage of this synchronization is the possibility of providing new telecommunication services such as multimedia broadcasting inside the building wherein the seamless wireless network is deployed.
The third advantage of this synchronization is the possibility of providing other services such as an accurate geolocation of the terminals.
There exists several means for providing nodes synchronization, over a fixed network, but limited in most cases to low accuracy (around a millisecond in practical scenarios with several nodes), and insufficient for taking benefit from it.
These algorithms depend a lot on the topology of the fixed network, and thus cannot be used with high reliability requirements.
For wireless networks, a target of around microsecond must be achieved. Several algorithms for distributed synchronization of nodes on a wireless network exist, most of them being developed for ad-hoc sensor networks. The known algorithms are based on correction of the synchronization time of each node from observation of neighboring nodes.
The existing algorithms for distributed synchronization of nodes most of the times apply an on-the-fly correction of the time of synchronization. Indeed, in most applications, the synchronization is an initial phase before communication between the nodes, and the synchronization is a necessary condition before any transmission.
However, in the context of the HeNBs, the node i.e. the HeNB itself is not in wireless communication i.e. there is no data exchange with other nodes (HeNBs).
However, the HeNB is in communication with several terminals. Thus, the synchronization between nodes is not a necessary condition but is a way for improving the overall system performance.
As several mobiles are in communication via the HeNB and synchronized to it, it is not possible to modify the synchronization time of the HeNB as it would cause the terminals to loose their synchronization with it, thus drastically degrading or cutting the communication.
The invention accordingly gives a solution to the above problem by proposing a synchronization method that does not disturb the communication between the nodes and the terminals.
The invention relates to a method for synchronizing at least two nodes in communication with at least one terminal in a wireless network, each node transmitting synchronization sub-frames, each synchronization sub-frame of a node being aligned to an actual synchronization time dedicated to said node, characterized in that it comprises
The invention also relates to a device for synchronizing at least two nodes in communication with at least one terminal in a wireless network, each node transmitting synchronization sub-frames, each synchronization sub-frame of a node being aligned to an actual synchronization time dedicated to said node, characterized in that it comprises means for defining a virtual synchronization time for each node and means, provided at each node, for:
A better understanding of the invention will be facilitated by reading the following description, which is given solely by way of examples and with reference to drawings, in which
The principle of the present invention consists on computing the classical (for example distributed) synchronization algorithm on a virtual synchronization time and updating the actual synchronization time of the node according to its virtual synchronization time.
The implementation of the method for synchronizing nodes 4, 6, 8 is explained with more details in the following description with reference to
The nodes 4, 6, 8 transmit symbols which are grouped into sub-frames.
According to a particular embodiment of the present invention, the transmission is OFDM (“Orthogonal Frequency Division Multiplexing”) based.
The nodes 4, 6, 8 transmit data sub-frames as well as sub-frames dedicated for synchronization. These dedicated sub-frames may be for example, a blank sub-frame or a MBSFN (“Multicast Broadcast over Single Frequency Network”) sub-frame of the 3GPP-LTE standard.
According to the invention, each synchronization sub-frame 20, 22, 24 contains a signal that can be spread over the multiple symbols of the sub-frame such that a pre-determined processing allows for locating a given point of the sub-frame, for example the middle of the sub-frame, as illustrated by arrows in the
This point is called the virtual synchronization time or VST of the node in the following description.
Thus, as illustrated, in
In the following, it is assumed that at a given time, some nodes are transmitting the synchronization sub-frames while other nodes switch into a receiving mode.
The synchronization method of the invention is described for node 4 but the same processing is carried for the other nodes 6 and 8.
In 40, the node 4 checks if its current transmitted sub-frame is a sub-frame containing signals for nodes synchronization, i.e. a synchronization sub-frame. If the result of the checking is negative, i.e. the sub-frame is not a synchronization sub-frame, the node 4 processes the data transmission to the connected terminal. Else, the node 4 checks in 42 if it must be in transmission or reception mode for synchronization between nodes.
This information can be given by a coordinator (for example, a gateway connected to several nodes) or by a scheduling of a known synchronization algorithm.
If the node 4 is in transmission mode for the synchronization between nodes, it transmits in 44 a signal aligned to its virtual synchronization time or VST.
If the node 4 is in the reception mode for the synchronization between nodes, it listens in 46 the signals sent in the time period corresponding to a sub-frame, aligned to its actual synchronization time (which is for example the middle time of the synchronization sub-frame) and determines the VST of other nodes 6, 8 from the received signals. This step 46 is illustrated in
As illustrated in
Node 4 detects the VST 32 and 34 of nodes 6 and 8 respectively.
These VSTs are not in the same position in the sub-frame as the nodes are not synchronized one with the others.
Measuring the virtual synchronization time of neighboring nodes is done either by the node itself using the synchronization sub-frames or by the terminals of the node, and provided to the node by measurements reports containing messages comprising the determined VSTs of the other nodes.
In 48, the node 4 applies a classical synchronization algorithm, for example a distributed synchronization algorithm known in ad-hoc networks, that computes in 50 an update of its virtual synchronization time as a function of the observed VSTs of the other nodes 6, 8. However, even if this VST is updated, the node 4 desynchronization with the neighboring nodes 6, 8 remains the same.
According to a particular embodiment of the invention, in 51, the node 4 modifies slowly its AST to be equal to its VST. The modification must be slow enough to allow a terminal in communication with the node 4 to be able to track this slow time synchronization drifting, and to keep a good communication quality. For example, the slow drifting might be processed by removing or adding one sample in the discrete version of the time domain OFDM signal. This allows for updating the AST of the node 4 while it is in communication with a terminal, and thus makes the system converge more quickly. This slow modification of the AST can be done by the node in parallel with other tasks such as transmitting or receiving data to/from the mobile terminals.
Then, in 52, the node 4 checks if it switches to the idle mode, i.e. it checks if there are active terminal connected to it or no. If node 4 is not in the idle node, the processing comes back to step 40.
As soon as the node 4 falls into an idle node, it updates in 54 its actual synchronizing time AST to its virtual synchronization time VST.
In the particular embodiment wherein a slow correction in 51 is performed, steps 52 and 54 are only optional.
This means that when the node 4 will start again a communication, the next sub-frames will be aligned to the VST determined in 50. The node 4 will be more synchronized to its neighboring nodes 6, 8 thus the nodes converge to a synchronized state and the performance of the network is improved.
Then, the processing comes back to step 46 so that the node synchronization goes on being improved while the node 4 is in idle mode.
In 60, the three nodes 4, 6 and 8 are not synchronized. Their virtual synchronization times VSTs are drawn by arrows and their actual synchronization time ASTs are draws by circles in
In 62, the node 8 listens and computes from the virtual synchronization times of nodes 4 and 6 an updated VST, shown as a dotted line arrow.
In 64, the node 8 transmits a synchronization sub-frame with a signal making its VST appear to the updated value (different from its actual synchronization time). The node 4 listens and computes from the VSTs of nodes 6 and 8 an updated VST shown as a dotted line arrow.
In 66, the node 4 transmits, a synchronization sub-frame with a signal making its virtual synchronization time appear to the updated value.
It is assumed that the node 8 felt in an idle mode between 64 and 66 and updated its actual synchronization time value to be equal to its virtual synchronization time.
In 68, it is assumed that the node 4 felt in an idle mode between 66 and 68 and updates its actual synchronization time value to be equal to its virtual synchronization time.
Thus, at the end of 68, the three nodes 4, 6 and 8 are synchronous.
Sub-frames 70, 72, 74, 76, 78 and 80 are normal sub-frames used for transmitting data.
The node 4 transmits a synchronization signal 82 (illustrated by a triangle) such that its virtual synchronization time VST4 is equal to its actual synchronization time AST4.
Thus, the synchronization signal is not truncated in synchronization sub-frame 20 of node 4.
For node 6, the virtual synchronization time VST6 is not equal to the actual synchronization AST6. In order to align the synchronization signal, referenced 84, of node 6 to the actual synchronization time AST6, this synchronization signal 84 is shifted by a value VST6-AST6 so the synchronization signal 84 is truncated.
The node 8 is assumed to be listening during the synchronization sub-frame 24. Node 8 thus makes a correlation with the synchronization signal 82 of node 4 and with the synchronization signal 84 of node 6.
For simplicity the propagation delay between nodes is regulated so that node 8 can determine VST4 and VST6.
It is worthy to observe that the peaks of the correlations vary as a function of the mis-synchronization of nodes and as a function of the virtual synchronization times as illustrated by amplitude of the arrows in
Concerning the choice of the synchronization signals transmitted within the synchronization sub-frames of the nodes, there are two possible embodiments.
According to the first embodiment, all the nodes use the same synchronization signal. In this case, a listening node can perform an autocorrelation, or a cross-correlation of the received signal and the synchronization signal, and detect the VST at the output of the auto/cross-correlation.
According to the second embodiment, all the nodes have a synchronization signal chosen from a set of signals having good cross-correlations such as Zadoff-chu sequences or Hadamard sequences. The listening node thus computes multiple cross correlations, one with each signal of the set, and can discriminate the VST of each neighbor. This is particularly useful if the synchronization algorithm needs to perfectly know the number of neighboring nodes and their respective VSTs.
Thus, the method of the invention aims at synchronizing the nodes 4, 6, 8 without disturbing the communication between the terminals 10, 12, 14.
An idea to reach this purpose is to take benefit from the idle periods during which no terminal is active to change the synchronization times of the nodes 4, 6, 8.
Unfortunately, by only considering idle time periods of the nodes for synchronization time update, the number of exchanges between the nodes which must be idle at the same time, and thus the convergence speed of the system is limited.
The invention gives a solution that allows for processing existing synchronization algorithms for updating the synchronization time of a given node, while active terminals are in communication with said node.
In some cases, a node can obtain a precise actual synchronisation time (AST) from another way of synchronization than using the above described algorithm of the invention. For example, said node is equipped with a GPS receiver or connected to an optical network providing a way of synchronization to the node.
Thus, as this node obtains its AST from the other way of synchronization, it aligns its VST to its AST and remains in a transmission mode of synchronization sub-frames as described in step 44 of the invention's algorithm. Thus, such a node gives an improved robustness to the overall system performance.
| Number | Date | Country | Kind |
|---|---|---|---|
| 09306314.7 | Dec 2009 | EP | regional |
