CA2091014A1 - Method for making handover decisions in a radio communications network - Google Patents
Method for making handover decisions in a radio communications networkInfo
- Publication number
- CA2091014A1 CA2091014A1 CA002091014A CA2091014A CA2091014A1 CA 2091014 A1 CA2091014 A1 CA 2091014A1 CA 002091014 A CA002091014 A CA 002091014A CA 2091014 A CA2091014 A CA 2091014A CA 2091014 A1 CA2091014 A1 CA 2091014A1
- Authority
- CA
- Canada
- Prior art keywords
- base station
- network
- neural network
- fixed base
- handover
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000004891 communication Methods 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 47
- 238000004088 simulation Methods 0.000 claims abstract description 3
- 238000012549 training Methods 0.000 claims description 17
- 238000005259 measurement Methods 0.000 claims description 13
- 210000002569 neuron Anatomy 0.000 claims description 9
- 238000012546 transfer Methods 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 4
- 230000002596 correlated effect Effects 0.000 claims description 2
- 230000006399 behavior Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 238000010606 normalization Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 150000002500 ions Chemical class 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W36/00—Hand-off or reselection arrangements
- H04W36/24—Reselection being triggered by specific parameters
- H04W36/32—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data
- H04W36/322—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data by location data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W36/00—Hand-off or reselection arrangements
- H04W36/24—Reselection being triggered by specific parameters
- H04W36/32—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data
Abstract
ABSTRACT A METHOD FOR MAKING HANDOVER DECISIONS IN A RADIO COMMUNICATIONS NETWORK The invention relates to a method for making handover decisions in a radio communication network comprising a number of fixed base stations and a number of mobile units. The method utilises an artificial neural network which is an image of the real network of a respective base station and which exhibit a behaviour pattern learnt through the acquisition of information from the network. Thereafter, simulation is carried out in the neural network through the generation of a list of eligible base stations to which handover can be effected, every one of the eligible stations being given points. Thereafter, a decision is made whether, or not, a handover will be effected by the network.
Description
2 0 ~
A M~QP ~(?R MAKIN(~7 l~ NUOVER DECI~IONS IN ~ RADTO
COMMUNIÇ~TIONS NETWORK
The invention relates to a method for making handover 5 decisions in a radio communications network comprising radio base stations and mobile units.
In accordance with the method of the present invention, decisions will be made concerning whether, and when, the handover of a mobile unit from one base station to another base 10 station unit will occur. It is important that such decisions are correct because a wrong decision can result in communication being interrupted. Furthermore, wrong decisions concerning handover could also have the result that other mobile units are deprived of the possibility to communicate.
Neural networks are generally described in US Patent Nos.
4 796 199, 4 918 617 and 4 983 962. These patents are concerned with the organisation of a proposed architectural structure and provide a good source of information in relation to neural networks.
In the following text, this information concerning neural networks 20 will be utilised in connection with mobile telephony.
In mobile radio systems, communication is established between a mobile unit and a fixed unit, i.e. a base station. In practice, a mobile unit makes contact with a base station which is the most suitable for the purpose of establishing effective 25 communication. The selection of a base station can be preprogrammed in relation to the geographic position of the mobile unit with respect to the base stations. Another possibility is for 2 2 ~ 9 ~
measurement to be made at the terminal, or at the base station.
The parameters which will be considered are, for example, signal strength, distance and so forth. When a mobile unit has to change base station, it will normally be done in accordance with one of the 5 above-mentioned principles. In the case where a wrong decision is made in respect of such a change, i.e. handover from one base station to another base station, an inferior transmission channel will be obtained, or alternatively, the communication will be interrupted. Any lack of handover can result in other connections 10 being broken due to interference. It is thus of great significance that decisions concerning handover are made with the correct assumptions.
It is an object of the present invention to provide a method for making handover decisions in a radio communication network 15 at the correct time. To this end, an artificial neural network has been created. The artificial neural network represents a model of the real radio communication network.
The present invention provides a method for making handover decisions in radio communication network which 20 comprises a number of fixed base stations and a number of mobile units, and in which communication between a mobile unit and a first radio base station is discontinued to enable the communication to be handed over to a second radio base station, characterised in that, the method includes the steps of causing an artificial neural 25 network which is an image of the real network of a respective base station, to exhibit a behaviour pattern learnt through the acquisition of information from the network, the information 2 ~
acquired by the artificial neural network being correlated to thc position of the mobile unit relative to the fixed base stations in the radio communication network with whicll the mobile unit can communicate; providing a list Or fixed base stations with whicll thc 5 mobile unit can communicate, the artificial neural network determining the degree of suitability for communication with a respective fixed base station; providing information for the radio communication network concerning the most suitable fixed base station for communicating with the mobile unit; and making a 10 decision concerning the connection of the mobile unit to a selected fixed base station.
The term "handover" means that a mobile unit in a radio communication network changes from communication with one radio base station of the network to another radio base station of 15 the network which is better placed from the point of view of communlcation.
According to one aspect of the present invention, the neural network is allocated a number of layers comprising an input layer, an output layer and a number of intermediate layers, the input 20 layer being allocated a number of nodes representing selected measurement data units, and the output layer being allocated a number of nodes representing a fixed base station to which the mobile unit is presently communicating and the fixed stations to which handover is possible, the nodes in the intermediate layers 25 being utilised for simulation.
According to another aspect of the present invention, each node of the neural network is a neuron, and weights input data 2 ~) e~
from all neurons in the preceding la~er, all weighted signals being added together and thereafter subjec~c d to transformation.
According to a further aspect of the present invention, the neural network is allocated training data, which is nortnalised, 5 which is selc cted to be as representative as possible of network behaviour, and which is allocated for the entire reception area to be served by the fixed base station and for the areas adjoining thc reception area, a desired structure for the neural network being determined for each set of training data.
Thus, the method according to the present invention solves handover problems wi~h the aid of neural networks, which are characterised by being able to learn certain behaviour through the acquisition of information.
The neural network is an image of the real radio communication network and every one of the radio base stations of the network obtains access to a neural network that reflects the network of the base station concerned. A mobile unit identifies its geographic position and communicates this information to the radio base station which thereafter utilises the neural network for deciding whether, or not, there should be a change of radio base station i.e. a handover.
The foregoing and other features according to the present invention will be better understood from the following description with reference to the accompanying drawings, in which:
Figure 1 illustrates, in a pictorial view, a radio communication network including mobile radio units and a number of base stations connected to a central exchange.
2 Q ~
Figure 2 shows, in the from of a block diagram, the manner in which measurement data in the mobile radio system is transferred to a locating block and thereafter to a decision block which forwards a decision regarding hando~er to the mobile radio system, Figure 3 shows, in !he form of a block diagram, the manner in which the measurement data is introduced into a normalisation block; the normalised data being transferred to an ANN (Artificial Neural Network) the output of which is the locating information for a mobile radio unit, Figure 4 illustrates the structure of a neural network, Figure 5 illustrates the manner in which input data from all nodes (neurons) in a layer of the neural network are weighted and added together and thereafter transformed.
Figure 6 illustrates the measurement area for a radio base station, and the adjacent base stations, of a radio communication network.
The radio communication network illustrated in Figure 1 of the drawings includes a number of base stations 1, each one of which is connected to a central exchange and adapted to communicated with a number of mobile units. The central excharige includes a number of LMNN functions, one for each base station. The LMNNs contain, inter alia, the artificial neural networks hereinbefore referred to.
As illustrated in Figure 1 of the drawings, the mobile radio units communicate with a respective one of the radio base stations 1 over the airwaves. The radio base station 1 continuously receives 2 ~
information from a mobile radio unit about its position, received signal strength and so forth.
As shown in Figure 2 of the drawings, this information, i.e. the measurement data from the mobile radio system, is fed in~o a 5 locating block, associated with the radio base station l, for locating the mobile radio unit from which the information is received. The information received by the locating block is then processed, consideration being given to the geographic position and the direction of movement of the mobile radio unit concerned. This 10 processing results in the establishment of a list of radio base stations l which are eligible for communicating with the mobile radio unit. This list is a priority list, in that, it specifies an order of priority relating to the suitability of the eligible radio base stations for communicating with the mobile radio unit.
As shown in Figure 2, the priority list is then transferred to a decision block which decides, and provides an output indicative of, whether, or not, the mobile radio unit should change radio base stations. The decision is then forwarded to tlle mobile radio system.
In order to obtain the best possible decision with such an 20 arrangement, it is necessary, as shown in Figure 3 of the drawings, to create an artificial neural network (ANN). With this arrangement, the measurement data is applied to a normalisation block and the normalised data at the output of this block is transferred to the artificial neural network. The output of neural network is the 25 locating information for a mobile radio unit, The artificial neural network is an image of the actual radio network of the respective base station. The artificial neural 2 Q ~
network is given information regarding the limits of its own coverage area and the overlap with adjoining base stations.
In order to obtain the best possible decisions, the artificial neural network will have to learn a desired behaviour depending 5 on the position and direction of movement of the mobile radio units.
As illustrated in Figure 4 of the drawings, the artificial neural network is structured in a number of layers, i.e. an input layer, a number of intermediate layers and an output layer.
The input layer has a node, shown as a circle, for each selected measurement unit. As illustrated in Figure 4, every one of the nodes in the input layer transfers its information to all nodes, shown as circles, in the first intermediate layer. Respective nodes in the first intermediate layer transfer their information to the 15 nodes of the next intermediate layer and so forth. Only the first and last intermediate layers are illustrated in Figure 4. The number of intermediate layers in the neural network is selected freely which is why one or a number of intermediate layers can be utilised. Finally, every one of the nodes in the last intermediate 20 layer transfers its information to all nodes in the output layer which represents possible candidates for handover. This includes the base station to which the mobile radio unit is currently connected. The structure of the artificial neural network varies with measurement data which is why different radio environments 25 result in different optimum structures.
Each node of the neural network is a neuron which weights input data from all neurons in the previous layer. Figure 5 of the 2Q ~
drawings illustrates the manner in which input data from each neuron is weighted and added together, and thereafter subjected to a transformation. The transformation is carried out in all layers of the neural network except the input layer which only contains one input per neuron. In order to obtain a desired behaviour, a non-linear transfer function is selected.
In order to obtain correct decisions regarding handover, it is necessary for the neural network to be taught certain behaviour through the acquisition of information. This is effected by supplying training data to the system. The training data needs to be provided for the entire reception area to be handled by the base station, and the adjoining areas.
Figure 6 of the drawings illustrates the measurement area for a base station. The solid lines show the desired reception ~lrea for a respective base station l. The dotted lines represent measurement points for training data and the dashed areas illustrate the coverage area of the base station l. The adjoining base stations 2 and 3 are also illustrated, as is the areas of overlap between the three base stations.
The construction of the artificial neural network is determined for each set of training data. Normalisation (see Figure 3) is carried out for a value between 0 and l with the aid of a maximum value for each data item in the complete set of training data.
During the learning process, a mobile radio unit travels within the coverage area of the base station and the adjoining areas. In the mobile radio unit, measurements are made with respect to the 2 ~
relevant parameters referred to above. The mobile radio unit sends the resul~s to the radio base station which further processes the information. During the training phase, diffcrent geographic positions are defined whicll can also be defined for other mobile 5 radio units.
Also, during the learning process, weights, referred to abovc in relation to Figure 5 of the drawings, are first randomly selected whereafter the network works with training data. For each set of training data, the result at the output nodes of the neural network 10 is checked against a desired result. The difference between the desired result and the result obtained with the training data gives rise to an error. Calculation of weights is carried out, for example, by the optional gradient method and the error is spread back through the network. By changing the different weights, the result 15 converges towards a desired result. When the error drops below a predetermined level, the neural network is considered to be trained and learning is terminated.
When the neural network is trained, the geographic position of a mobile radio unit can be unambiguously determined with the 20 guidance of data obtained from the network. As stated above, the inforrnation regarding the position of a mobile radio unit is transferred to the locating block which transfers the information to the decision block. The decision block produces a vector with numbers which relate to the suitability for handover to a respective 25 radio base station. The numbcr varie~ within limits which (Icl-cn(l on the selected transfer function. A decision is made with respect to history and hysteresis. "History" means that the decision block stores a number of the la~est locating vectors. A demand is made that a certain number of these locating vectors should show the same result for handover to occur. A short history provicles a fast decision, but a long history provides a more reliable decision at the 5 cost of speed. Thus, an appropriate balance between these extreme cases must be carried out in the individual cases where different aspects get balanced against one another. "Hysteresis", referred to above, means that a candidate for handover would result in improved communication, by a certain measure, in relation to the 10 suitability number of the existing base station. For handover to be possible, the hysteresis value must lie within the transfer function.
Hysteresis relates the handover decision to the suitability number for the present base station and provides the same result in the handover function as the history.
A M~QP ~(?R MAKIN(~7 l~ NUOVER DECI~IONS IN ~ RADTO
COMMUNIÇ~TIONS NETWORK
The invention relates to a method for making handover 5 decisions in a radio communications network comprising radio base stations and mobile units.
In accordance with the method of the present invention, decisions will be made concerning whether, and when, the handover of a mobile unit from one base station to another base 10 station unit will occur. It is important that such decisions are correct because a wrong decision can result in communication being interrupted. Furthermore, wrong decisions concerning handover could also have the result that other mobile units are deprived of the possibility to communicate.
Neural networks are generally described in US Patent Nos.
4 796 199, 4 918 617 and 4 983 962. These patents are concerned with the organisation of a proposed architectural structure and provide a good source of information in relation to neural networks.
In the following text, this information concerning neural networks 20 will be utilised in connection with mobile telephony.
In mobile radio systems, communication is established between a mobile unit and a fixed unit, i.e. a base station. In practice, a mobile unit makes contact with a base station which is the most suitable for the purpose of establishing effective 25 communication. The selection of a base station can be preprogrammed in relation to the geographic position of the mobile unit with respect to the base stations. Another possibility is for 2 2 ~ 9 ~
measurement to be made at the terminal, or at the base station.
The parameters which will be considered are, for example, signal strength, distance and so forth. When a mobile unit has to change base station, it will normally be done in accordance with one of the 5 above-mentioned principles. In the case where a wrong decision is made in respect of such a change, i.e. handover from one base station to another base station, an inferior transmission channel will be obtained, or alternatively, the communication will be interrupted. Any lack of handover can result in other connections 10 being broken due to interference. It is thus of great significance that decisions concerning handover are made with the correct assumptions.
It is an object of the present invention to provide a method for making handover decisions in a radio communication network 15 at the correct time. To this end, an artificial neural network has been created. The artificial neural network represents a model of the real radio communication network.
The present invention provides a method for making handover decisions in radio communication network which 20 comprises a number of fixed base stations and a number of mobile units, and in which communication between a mobile unit and a first radio base station is discontinued to enable the communication to be handed over to a second radio base station, characterised in that, the method includes the steps of causing an artificial neural 25 network which is an image of the real network of a respective base station, to exhibit a behaviour pattern learnt through the acquisition of information from the network, the information 2 ~
acquired by the artificial neural network being correlated to thc position of the mobile unit relative to the fixed base stations in the radio communication network with whicll the mobile unit can communicate; providing a list Or fixed base stations with whicll thc 5 mobile unit can communicate, the artificial neural network determining the degree of suitability for communication with a respective fixed base station; providing information for the radio communication network concerning the most suitable fixed base station for communicating with the mobile unit; and making a 10 decision concerning the connection of the mobile unit to a selected fixed base station.
The term "handover" means that a mobile unit in a radio communication network changes from communication with one radio base station of the network to another radio base station of 15 the network which is better placed from the point of view of communlcation.
According to one aspect of the present invention, the neural network is allocated a number of layers comprising an input layer, an output layer and a number of intermediate layers, the input 20 layer being allocated a number of nodes representing selected measurement data units, and the output layer being allocated a number of nodes representing a fixed base station to which the mobile unit is presently communicating and the fixed stations to which handover is possible, the nodes in the intermediate layers 25 being utilised for simulation.
According to another aspect of the present invention, each node of the neural network is a neuron, and weights input data 2 ~) e~
from all neurons in the preceding la~er, all weighted signals being added together and thereafter subjec~c d to transformation.
According to a further aspect of the present invention, the neural network is allocated training data, which is nortnalised, 5 which is selc cted to be as representative as possible of network behaviour, and which is allocated for the entire reception area to be served by the fixed base station and for the areas adjoining thc reception area, a desired structure for the neural network being determined for each set of training data.
Thus, the method according to the present invention solves handover problems wi~h the aid of neural networks, which are characterised by being able to learn certain behaviour through the acquisition of information.
The neural network is an image of the real radio communication network and every one of the radio base stations of the network obtains access to a neural network that reflects the network of the base station concerned. A mobile unit identifies its geographic position and communicates this information to the radio base station which thereafter utilises the neural network for deciding whether, or not, there should be a change of radio base station i.e. a handover.
The foregoing and other features according to the present invention will be better understood from the following description with reference to the accompanying drawings, in which:
Figure 1 illustrates, in a pictorial view, a radio communication network including mobile radio units and a number of base stations connected to a central exchange.
2 Q ~
Figure 2 shows, in the from of a block diagram, the manner in which measurement data in the mobile radio system is transferred to a locating block and thereafter to a decision block which forwards a decision regarding hando~er to the mobile radio system, Figure 3 shows, in !he form of a block diagram, the manner in which the measurement data is introduced into a normalisation block; the normalised data being transferred to an ANN (Artificial Neural Network) the output of which is the locating information for a mobile radio unit, Figure 4 illustrates the structure of a neural network, Figure 5 illustrates the manner in which input data from all nodes (neurons) in a layer of the neural network are weighted and added together and thereafter transformed.
Figure 6 illustrates the measurement area for a radio base station, and the adjacent base stations, of a radio communication network.
The radio communication network illustrated in Figure 1 of the drawings includes a number of base stations 1, each one of which is connected to a central exchange and adapted to communicated with a number of mobile units. The central excharige includes a number of LMNN functions, one for each base station. The LMNNs contain, inter alia, the artificial neural networks hereinbefore referred to.
As illustrated in Figure 1 of the drawings, the mobile radio units communicate with a respective one of the radio base stations 1 over the airwaves. The radio base station 1 continuously receives 2 ~
information from a mobile radio unit about its position, received signal strength and so forth.
As shown in Figure 2 of the drawings, this information, i.e. the measurement data from the mobile radio system, is fed in~o a 5 locating block, associated with the radio base station l, for locating the mobile radio unit from which the information is received. The information received by the locating block is then processed, consideration being given to the geographic position and the direction of movement of the mobile radio unit concerned. This 10 processing results in the establishment of a list of radio base stations l which are eligible for communicating with the mobile radio unit. This list is a priority list, in that, it specifies an order of priority relating to the suitability of the eligible radio base stations for communicating with the mobile radio unit.
As shown in Figure 2, the priority list is then transferred to a decision block which decides, and provides an output indicative of, whether, or not, the mobile radio unit should change radio base stations. The decision is then forwarded to tlle mobile radio system.
In order to obtain the best possible decision with such an 20 arrangement, it is necessary, as shown in Figure 3 of the drawings, to create an artificial neural network (ANN). With this arrangement, the measurement data is applied to a normalisation block and the normalised data at the output of this block is transferred to the artificial neural network. The output of neural network is the 25 locating information for a mobile radio unit, The artificial neural network is an image of the actual radio network of the respective base station. The artificial neural 2 Q ~
network is given information regarding the limits of its own coverage area and the overlap with adjoining base stations.
In order to obtain the best possible decisions, the artificial neural network will have to learn a desired behaviour depending 5 on the position and direction of movement of the mobile radio units.
As illustrated in Figure 4 of the drawings, the artificial neural network is structured in a number of layers, i.e. an input layer, a number of intermediate layers and an output layer.
The input layer has a node, shown as a circle, for each selected measurement unit. As illustrated in Figure 4, every one of the nodes in the input layer transfers its information to all nodes, shown as circles, in the first intermediate layer. Respective nodes in the first intermediate layer transfer their information to the 15 nodes of the next intermediate layer and so forth. Only the first and last intermediate layers are illustrated in Figure 4. The number of intermediate layers in the neural network is selected freely which is why one or a number of intermediate layers can be utilised. Finally, every one of the nodes in the last intermediate 20 layer transfers its information to all nodes in the output layer which represents possible candidates for handover. This includes the base station to which the mobile radio unit is currently connected. The structure of the artificial neural network varies with measurement data which is why different radio environments 25 result in different optimum structures.
Each node of the neural network is a neuron which weights input data from all neurons in the previous layer. Figure 5 of the 2Q ~
drawings illustrates the manner in which input data from each neuron is weighted and added together, and thereafter subjected to a transformation. The transformation is carried out in all layers of the neural network except the input layer which only contains one input per neuron. In order to obtain a desired behaviour, a non-linear transfer function is selected.
In order to obtain correct decisions regarding handover, it is necessary for the neural network to be taught certain behaviour through the acquisition of information. This is effected by supplying training data to the system. The training data needs to be provided for the entire reception area to be handled by the base station, and the adjoining areas.
Figure 6 of the drawings illustrates the measurement area for a base station. The solid lines show the desired reception ~lrea for a respective base station l. The dotted lines represent measurement points for training data and the dashed areas illustrate the coverage area of the base station l. The adjoining base stations 2 and 3 are also illustrated, as is the areas of overlap between the three base stations.
The construction of the artificial neural network is determined for each set of training data. Normalisation (see Figure 3) is carried out for a value between 0 and l with the aid of a maximum value for each data item in the complete set of training data.
During the learning process, a mobile radio unit travels within the coverage area of the base station and the adjoining areas. In the mobile radio unit, measurements are made with respect to the 2 ~
relevant parameters referred to above. The mobile radio unit sends the resul~s to the radio base station which further processes the information. During the training phase, diffcrent geographic positions are defined whicll can also be defined for other mobile 5 radio units.
Also, during the learning process, weights, referred to abovc in relation to Figure 5 of the drawings, are first randomly selected whereafter the network works with training data. For each set of training data, the result at the output nodes of the neural network 10 is checked against a desired result. The difference between the desired result and the result obtained with the training data gives rise to an error. Calculation of weights is carried out, for example, by the optional gradient method and the error is spread back through the network. By changing the different weights, the result 15 converges towards a desired result. When the error drops below a predetermined level, the neural network is considered to be trained and learning is terminated.
When the neural network is trained, the geographic position of a mobile radio unit can be unambiguously determined with the 20 guidance of data obtained from the network. As stated above, the inforrnation regarding the position of a mobile radio unit is transferred to the locating block which transfers the information to the decision block. The decision block produces a vector with numbers which relate to the suitability for handover to a respective 25 radio base station. The numbcr varie~ within limits which (Icl-cn(l on the selected transfer function. A decision is made with respect to history and hysteresis. "History" means that the decision block stores a number of the la~est locating vectors. A demand is made that a certain number of these locating vectors should show the same result for handover to occur. A short history provicles a fast decision, but a long history provides a more reliable decision at the 5 cost of speed. Thus, an appropriate balance between these extreme cases must be carried out in the individual cases where different aspects get balanced against one another. "Hysteresis", referred to above, means that a candidate for handover would result in improved communication, by a certain measure, in relation to the 10 suitability number of the existing base station. For handover to be possible, the hysteresis value must lie within the transfer function.
Hysteresis relates the handover decision to the suitability number for the present base station and provides the same result in the handover function as the history.
Claims (7)
1. A method for making handover decisions in a radio communication network which comprises a number of fixed base stations and a number of mobile units, and in which communication between a mobile unit and a first radio base station is discontinued to enable the communication to be handed over to a second radio base station, characterised in that, the method includes the steps of causing an artificial neural network which is an image of the real network of a respective base station, to exhibit a behaviour pattern learnt through the acquisition of information from the network, the information acquired by the artificial neural network being correlated to the position of the mobile unit relative to the fixed base stations in the radio communication network with which the mobile unit can communicate; providing a list of fixed base stations with which the mobile unit can communicate, the artificial neural network determining the degree of suitability for communication with a respective fixed base station; providing information for the radio communication network concerning the most suitable fixed base station for communicating with the mobile unit; and making a decision concerning the connection of the mobile unit to a selected fixed base station.
2. A method as claimed in claim 1, characterised in that the neural network is allocated a number of layers comprising an input layer, an output layer and a number of intermediate layers, in that the input layer is allocated a number of nodes representing selected measurement data units, in that the output layer is allocated a number of nodes representing a fixed base station to which the mobile is presently communicating and the fixed stations to which handover is possible, and in that nodes in the intermediate layers are utilized for simulation.
3. A method as claimed in claim 2, characterised in that each node of the network is a neuron, in that each node weights input data from all neurons in the preceding layer, and in that all weighted signals are added together and are thereafter subjected to transformation.
4. A method as claimed in any one of the preceding claims, characterised in that the neural network is allocated training data, which is normalised, in that the training data is selected to be as representative as possible of network behaviour, in that training data is allocated for the entire reception area to be served by the fixed base station and for the adjoining areas, and in that a desired structure for the neural network is determined for each set of training data.
5. A method as claimed in claim 4, characterised in that, during the training process, weights are randomly selected whereafter the neural network operates with received data, in that the result at the output nodes of the neural network is compared with a desired result, the difference between the desired result and the received result constituting an error relating to the selection of the weights, in that new weights are calculated and the process is repeated until the result at the output nodes of the neural network converges towards a desired value, and in that the training process is terminated when the difference between the desired and received results falls below a predetermined level.
6. A method as claimed in any one of the preceding claims, characterised in that input data in vector form is transferred to a locating block for the mobile unit, in that the locating block transfers the said data to a decision block, in that the decision block produces a vector with a number that relates to the suitability for handover to another fixed base station, and in that a handover decision is based on historical information and whether handover to the said another fixed station would result in improved communication, by a certain measure, in relation to the suitability number of the fixed base station with which the mobile unit is presently communicating.
7. A radio communication network in which handover decisions are made in accordance with the method as claimed in any one of the preceding claims.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
SE9201164A SE470151B (en) | 1992-04-13 | 1992-04-13 | Method for making handover decisions in communication radio networks |
SE9201164-2 | 1992-04-13 |
Publications (1)
Publication Number | Publication Date |
---|---|
CA2091014A1 true CA2091014A1 (en) | 1993-10-14 |
Family
ID=20385938
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CA002091014A Abandoned CA2091014A1 (en) | 1992-04-13 | 1993-03-04 | Method for making handover decisions in a radio communications network |
Country Status (9)
Country | Link |
---|---|
US (1) | US5434950A (en) |
EP (1) | EP0566548B1 (en) |
JP (1) | JPH0622365A (en) |
AU (1) | AU655719B2 (en) |
CA (1) | CA2091014A1 (en) |
DE (1) | DE69319017T2 (en) |
ES (1) | ES2117702T3 (en) |
GB (1) | GB2266212B (en) |
SE (1) | SE470151B (en) |
Families Citing this family (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SE9302140L (en) * | 1993-06-21 | 1994-08-29 | Televerket | Procedure for locating mobile stations in digital telecommunications networks |
JP2885067B2 (en) * | 1994-05-30 | 1999-04-19 | 日本電気株式会社 | Mobile subscriber calling method and mobile communication system |
SE505915C2 (en) * | 1994-08-18 | 1997-10-20 | Ericsson Telefon Ab L M | Cellular mobile communication system |
SE505006C2 (en) * | 1994-08-31 | 1997-06-09 | Telia Ab | Device for predicting handover in cellular telecommunication system |
GB2334185B (en) * | 1994-12-09 | 1999-09-29 | Motorola Inc | Method for determining handover in a communications system |
US5717406A (en) * | 1995-06-07 | 1998-02-10 | Sanconix Inc. | Enhanced position calculation |
GB2311697B (en) * | 1996-03-22 | 1999-07-28 | Matsushita Electric Ind Co Ltd | Wireless communication system and method and system for detection of position of radio mobile station |
US6236365B1 (en) | 1996-09-09 | 2001-05-22 | Tracbeam, Llc | Location of a mobile station using a plurality of commercial wireless infrastructures |
US7764231B1 (en) | 1996-09-09 | 2010-07-27 | Tracbeam Llc | Wireless location using multiple mobile station location techniques |
US9134398B2 (en) | 1996-09-09 | 2015-09-15 | Tracbeam Llc | Wireless location using network centric location estimators |
WO1999018748A1 (en) * | 1996-09-27 | 1999-04-15 | Telia Ab | Neural network application for frequency planning |
US6553355B1 (en) | 1998-05-29 | 2003-04-22 | Indranet Technologies Limited | Autopoietic network system endowed with distributed artificial intelligence for the supply of high volume high-speed multimedia telesthesia telemetry, telekinesis, telepresence, telemanagement, telecommunications, and data processing services |
AU2003271394B2 (en) * | 1998-05-29 | 2007-03-08 | Indranet Technologies Limited | Computer network structure and cybernetic device |
US6097953A (en) * | 1998-12-22 | 2000-08-01 | Motorola, Inc. | Method of performing handoff in a wireless communication system |
JP3950602B2 (en) | 1999-12-06 | 2007-08-01 | 株式会社日立コミュニケーションテクノロジー | Mobile communication system, handover control apparatus, and handover method thereof |
US9875492B2 (en) | 2001-05-22 | 2018-01-23 | Dennis J. Dupray | Real estate transaction system |
US10684350B2 (en) | 2000-06-02 | 2020-06-16 | Tracbeam Llc | Services and applications for a communications network |
US10641861B2 (en) | 2000-06-02 | 2020-05-05 | Dennis J. Dupray | Services and applications for a communications network |
US8467821B1 (en) | 2000-08-16 | 2013-06-18 | International Business Machines Corporation | System and method for anticipating transmit power requirements in wireless mobile units communicating with a base station |
GB2377130A (en) * | 2001-06-29 | 2002-12-31 | Motorola Inc | Handover criteria for mobile transceivers |
US7092722B1 (en) * | 2001-07-26 | 2006-08-15 | Sprint Spectrum L.P. | Method and system for establishing mobile station active set based on mobile station location |
US8195204B1 (en) | 2007-07-25 | 2012-06-05 | Sprint Spectrum L.P. | Method and apparatus for scanning sectors in order of distance from mobile station |
US7881263B1 (en) | 2007-07-31 | 2011-02-01 | Sprint Spectrum L.P. | Method for use of azimuth and bearing data to select a serving sector for a mobile station |
ATE551846T1 (en) * | 2007-08-06 | 2012-04-15 | Schweizerische Bundesbahnen Sbb | METHOD FOR THE MOBILITY MANAGEMENT OF TERMINAL DEVICES IN MOBILE NETWORKS, MOBILE NETWORK AND MOBILE TERMINAL DEVICE |
US20090312036A1 (en) | 2008-06-16 | 2009-12-17 | Skyhook Wireless, Inc. | Methods and systems for improving the accuracy of expected error estimation in location determinations using a hybrid cellular and wlan positioning system |
US8478275B1 (en) | 2010-08-05 | 2013-07-02 | Sprint Spectrum L.P. | Conditional assignment of connection identifiers to help avoid communication errors |
US9538493B2 (en) | 2010-08-23 | 2017-01-03 | Finetrak, Llc | Locating a mobile station and applications therefor |
US8670425B1 (en) | 2011-08-09 | 2014-03-11 | Sprint Spectrum L.P. | Use of past duration of stay as trigger to scan for wireless coverage |
DE102014212989A1 (en) * | 2014-07-04 | 2016-01-07 | Bayerische Motoren Werke Aktiengesellschaft | Controlling data connections and / or data transmissions in a mobile radio device |
CN113661727B (en) * | 2019-02-19 | 2022-04-01 | 诺基亚通信公司 | Configuration of a neural network for a Radio Access Network (RAN) node of a wireless network |
KR20210063911A (en) * | 2019-11-25 | 2021-06-02 | 삼성전자주식회사 | Apparatus and method for performing handover in wireless communication system |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4796199A (en) * | 1987-02-24 | 1989-01-03 | Oregon Graduate Center | Neural-model, information-handling architecture and method |
US4983962A (en) * | 1987-09-14 | 1991-01-08 | Hammerstrom Daniel W | Neural-model, computational architecture employing broadcast hierarchy and hypergrid, point-to-point communication |
US5014235A (en) * | 1987-12-15 | 1991-05-07 | Steven G. Morton | Convolution memory |
US4874963A (en) * | 1988-02-11 | 1989-10-17 | Bell Communications Research, Inc. | Neuromorphic learning networks |
US4918617A (en) * | 1988-11-10 | 1990-04-17 | Oregon Graduate Center | Neural-model computational system with multi-directionally overlapping broadcast regions |
JP2822432B2 (en) * | 1989-03-30 | 1998-11-11 | 日本電気株式会社 | Channel selection method for wireless system |
GB2242805B (en) * | 1990-04-06 | 1994-08-03 | Stc Plc | Handover techniques |
DE69220683T2 (en) * | 1991-01-23 | 1997-11-06 | Fujitsu Ltd | Adaptive equalizers |
-
1992
- 1992-04-13 SE SE9201164A patent/SE470151B/en not_active IP Right Cessation
-
1993
- 1993-02-23 GB GB9303566A patent/GB2266212B/en not_active Expired - Fee Related
- 1993-03-04 CA CA002091014A patent/CA2091014A1/en not_active Abandoned
- 1993-03-17 US US08/032,664 patent/US5434950A/en not_active Expired - Fee Related
- 1993-03-23 DE DE69319017T patent/DE69319017T2/en not_active Expired - Fee Related
- 1993-03-23 ES ES93850054T patent/ES2117702T3/en not_active Expired - Lifetime
- 1993-03-23 EP EP93850054A patent/EP0566548B1/en not_active Expired - Lifetime
- 1993-03-26 AU AU35515/93A patent/AU655719B2/en not_active Ceased
- 1993-04-12 JP JP5109872A patent/JPH0622365A/en active Pending
Also Published As
Publication number | Publication date |
---|---|
US5434950A (en) | 1995-07-18 |
JPH0622365A (en) | 1994-01-28 |
AU655719B2 (en) | 1995-01-05 |
DE69319017D1 (en) | 1998-07-16 |
DE69319017T2 (en) | 1998-10-08 |
GB2266212A (en) | 1993-10-20 |
ES2117702T3 (en) | 1998-08-16 |
GB2266212B (en) | 1995-06-28 |
SE9201164D0 (en) | 1992-04-13 |
EP0566548B1 (en) | 1998-06-10 |
SE470151B (en) | 1993-11-15 |
EP0566548A1 (en) | 1993-10-20 |
GB9303566D0 (en) | 1993-04-07 |
AU3551593A (en) | 1993-10-14 |
SE9201164L (en) | 1993-10-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP0566548B1 (en) | A method for making handover decisions in a radio communications network | |
US6140964A (en) | Wireless communication system and method and system for detection of position of radio mobile station | |
EP0963130B1 (en) | Control of handoff in CDMA cellular systems | |
US4398063A (en) | Mobile telephone interoffice handoff limiting method and apparatus | |
US5905957A (en) | Method and apparatus for management of automatically initiated communication | |
Pathirana et al. | Location estimation and trajectory prediction for cellular networks with mobile base stations | |
US6014564A (en) | Method and apparatus for determining virtual cell area | |
Xu et al. | QoE-driven intelligent handover for user-centric mobile satellite networks | |
GB2329801A (en) | System for the detection of the position of a radio mobile station | |
MXPA01001746A (en) | System network and method for the transference of cell handover information. | |
KR20010032466A (en) | Method and apparatus for determining hand-off candidates in a communication system | |
CN106054875A (en) | Distributed multi-robot dynamic network connectivity control method | |
Xu et al. | Deep reinforcement learning based mobility load balancing under multiple behavior policies | |
CN114885340A (en) | Ultra-dense wireless network power distribution method based on deep transfer learning | |
CN115811788B (en) | D2D network distributed resource allocation method combining deep reinforcement learning and unsupervised learning | |
CN113395724B (en) | Mode optimization method and device for base station | |
Munoz-Rodriguez et al. | Neural supported hand off methodology in micro cellular systems | |
US6311067B1 (en) | Gradual frequency plan revision methodology | |
Hortos | Application of neural networks to the dynamic spatial distribution of nodes within an urban wireless network | |
Salcic | Agpcs–an automatic gsm-based positioning and communication system | |
US20220103272A1 (en) | Device and method for assessing a state of a radio channel | |
Duque-Antón et al. | Extending Kohonen's self-organizing mapping for adaptive resource management in cellular radio networks | |
Feng et al. | A practical heterogeneous network optimisation algorithm based on reinforcement learning | |
CN111770466A (en) | Effective energy supplement online scheme in wireless sensor and actuator network | |
CN116634517A (en) | Base station switching method and device for double-connection mobile communication system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FZDE | Discontinued | ||
FZDE | Discontinued |
Effective date: 20010305 |