Title of Invention

METHOD AND SYSTEM FOR CONTROLLING RADIO COMMUNICATIONS NETWORK AND RADIO NETWORK CONTROLLER

Abstract

(57) Abstract The invention relates to a method and system for controlling a radio communications network. The invention is advantageously applied in broadband radio networks (GRAN) offering fixed network services to their users. An idea of the invention is that in preparation for a handover a list in a neighbouring radio network controller is compiled of those base stations that would constitute a candidate set should said neighbouring radio network controller be made the active radio network controller. A second idea of the invention is that in conjunction with a coimection set-up a radio network controller is selected via which user information flows for the whole duration of the connection. That radio network controller is here called an anchor controller (aRNC). If a connection is handed over to a base station that belongs to another radio network controller (bRNC), the user information is directed so as to travel to the active radio network controller via the anchor controller. The invention makes it possible, among other things, to reduce the load of the core network of a cellular network and to decrease transmission power in conjunction with a handover. Fig. 7 .

Full Text

The invention relates to a method and system for controlling a radio communic¬ations network and a radio network controller. Particularly the invention relates to the handover procedure in a cellular system. The invention can be advantageously applied in broadband radio networks that offer fixed network services to their users. Below it will be described the prior art by first illustrating the operation of a popular second-generation cellular system and in particular the handover, or change of active base stations serving a mobile station moving in the cellular network's coverage area. Then it will be disclosed the characteristics of new, third-generation cellular systems and problems related to prior-art handover solutions. A terminal of a cellular radio system attempts to choose a base station so as to operate on said base station's coverage area, or cell. Conventionally, the choice has been based on tlie measurement of the strength of the received radio signal in the terminal and base station. For example, in GSM (Global System for Mobile tele¬communications) each base station transmits a signal on a so-called broadcast control chaimel (BCCH) and the terminals measure the strengths of the received BCCH signals and based on that, determine which cell is the most advantageous one as regards the quality of the radio link. Base stations also transmit to the terminals information about the BCCH frequencies used in the neighbouring cells so tliat the terminals Icnow what frequencies they have to hsten to in order to find the BCCH transmissions of the neighbouring cells. Fig. 1 shows a second-generation cellular system that comprises a mobile switching centre (MSC) belonging to the core network (CN) of the cellular system as well as base station controllers (BSC) and base stations (BS) belonging to the a radio access network (R-A.X), to which mobile stations (MS) are linked via radio interface. Fig. 2 shows the coverage areas C21-C29 of base stations BS21-BS29 of a second-generation cellular system. In second-generation cellular systems, such as GSM, communication between base stations BS and the core network CN occurs via base station controllers BSC. Usually, one base station controller controls a large ntmiber of base stations so tliat when a tenninal moves from the area of a cell to the area of anotiier cell, the base stations of both the old and the new cell are connected to the same base station controller. Thus the handover can be executed in the base station controller. So, in the conventional GSM system, for example, there occur fairly few handovers between a base station of a first base station controller and a base station of a second base station controller. In such a case, the switching centre has to release the connection with the first base station controller and establish a new connection with the new base station controller. Such an event involves a lot of signalling between the base station controllers and th? switching centre and as the distances between the base station controllers and the switching centre may be long there may occur disturbances in the connection during the handover. The prior-art handover arrangement is suitable for the so-called second-generation digital cellular radio systems such as GSM and its extension DCS 1800 (Digital Communications System at 1800 MH2), IS-54 (Interim Standard 54), and PDC (Personal Digital Cellular). However, it has been suggested that in future third-generation digital cellular systems the service levels offered to the terminals by the cells may differ considerably from a cell to another. Proposals for third-generation systems include UMTS (Universal Mobile Telecommunications System) and FPLMTS/IMT-2000 (Future Public Land Mobile Telecommunications System / International Mobile Telecommunications at 2000 MHz). In these plans cells are categorised according to their size and characteristics into pico-, nano-, micro- and macrocells, and an example of the service level is the bit rate. The bit rate is the highest in picocells and the lowest in macrocells. The cells may overlap partially or completely and there may be different terminals so that not all terminals necessarily are able to utilise all the service levels offered by the cells. Fig. 3 shows a version of a future cellular radio system which is not entirely new compared with the known GSM system but which includes both known elements and completely new elements. In current cellular radio systems the bottleneck that prevents more advanced services from being offered to the terminals comprises the radio access network RAN which includes tlie base stations and base station controllers. The core network of a cellular radio system comprises mobile services switching cenn:es (MSC), other network elements (in GSM, e.g. SGSN and GGSN, i.e. Serving GPRS Support Node and Gateway GPRS Support node, where GPRS stands for General Packet Radio Service) and the related transmission systems. According e.g. to the GSM+ specifications developed firom GSM the core network can also provide new services. In Fig. 3, the core network of a cellular radio system 30 comprises a GSM+ core network 31 which has three parallel radio access networks linked to it. Of those, net-worlcs 32 and 33 are UMTS radio access networks and network 34 is a GSM-i- radio access network. The upper UMTS radio access network 32 is e.g. a commercial radio access network, owned by a telecommunications operator offering mobile services, which equally serves all subscribers of said telecommunications operator. The lower UMTS radio access network 33 is e.g. private and owned e.g. by a company in whose premises said radio access network operates. Typically the cells of the private radio access network 33 are nano- and/or picocells in which only terminals of the employees of said company can operate. All three radio access net¬works may have cells of different sizes offering different types of services, Additionally, cells of all three radio access networks 32, 33 and 34 may overlap either entirely or in part. The bit rate used at a given moment of time depends, among other things, on the radio path conditions, characteristics of the services used, regional overall capacity of the cellular system and the capacity needs of other users. The new types of radio access networks mentioned above are called generic radio access networks (GRAN). Such a network can co-operate with different types of fixed core networks CN and especially with the GPRS network of the GSM system. The generic radio access network (GRAN) can be defined as a set of base stations (BS) and radio network controllers (RNC) that are capable of com¬municating with each other using signalling messages. Below, the generic radio access network will be called in short a radio network GRAN. The terminal 35 shown in Fig, 3 is preferably a so-called dual-mode terminal that can serve either as a second-generation GSM terminal or as a third-generation UMTS terminal according to what kind of services are available at each particular location and what the user's communication needs are. It may also be a multimode terminal that can function as terminal of several different communications systems according to need and the services available. Radio access networks and services available to the user are specified in a subscriber identity module 36 (SIM) connected to the terminal. Fig. 4 shows in more detail a core network CN of a third-generation cellular system, comprising a switching centre MSC, and a radio network GRAN coimected to the core network. The radio network GRAN comprises radio network controllers RNC and base stations BS connected to them. A given radio network controller RNC and the base stations connected to it are able to offer broadband services while a second radio network controller and base stations connected to it may be able to offer only conventional narrowband services but possibly covering a larger area. Fig. 5 shows coverage areas 51a-56a of base stations 51-56 in a third-generation cellular system. As can be seen from Fig. 5, a mobile station travelling only a short distance can choose from many base stations for the radio link. New cellular systems can employ a so-called macrodiversity combining technique related to CDMA systems. This means that on the downlink path a terminal receives user data from at least two base stations and correspondingly, the us'er data transmitted by the terminal is received by at least two base stations. Then, instead of one, there are two or more active base stations, or a so-called active set. Using macrodiversity combining it is possible to achieve a better quality of data com¬munications as momentary fade-outs and disturbances occurring on a given trans¬mission path can be compensated for by means of data transmitted via a second transmission path. For selecting an active set an active radio network controller determines, on the basis of the geographic location, for example, a candidate set of base stations, which is a set of the base stations that are used for measuring general signal strength information using e.g, a pilot signal. Below, this candidate set of base stations will be called a candidate set (CS) in short, In some systems, such as IS-41, separate candidate base stations are used. Document WO 95/04423 discloses a method for executing a handover between base stations. For selecting a new active base station a base station candidate list is compiled. However, the document does not disclose any solutions for handover situations in a third generation cellular system where the active radio network controller should be changed and where several base stations are simultaneously active. Let us consider the application of a prior-art arrangement to a proposed third-generation digital cellular system. In third-generation systems, base station handovers and radio network controller handovers are more firequent than in second-generation systems. One of the reasons behind this is that the cell sizes may be remarkably small and that there may occur need to change the service type e.g. from narrowband to broadband during a call. In accordance with the prior an a handover between radio network controllers would be carried out in such a manner that the user data connection between the switching centre and the so-called old active radio network controller/base station is released and a new connection is established between the switching centre and the so-called new active radio network controller/base station. Then the switching centre would have to release/set up many connections, which involves a lot of signalling between the switching centre and the radio network controller. Further¬more, there are very many small-sized cells in the area of one switching centre, and in broadband applications the amount of user data transmitted is great. This puts very tight requirements for capacity and speed on the switching centre hardware, which in large systems cannot be met at reasonable costs using current technology. Secondly, known systems have a problem of how to transmit signalling and data of the core network CN and signalling of the radio network to a terminal moving in the radio network's area. CN signalling and data are specifically meant for the terminal and routed via radio network controllers. Radio network signalling may be intended either for the terminal or for the radio network itself so that it can arrange optimal use of radio resources in the network area. The problem is caused by the moving terminal and its effect on the flow of data in the radio network's area. When using macrodiversity combining the prior art further has the problem that after a handover between radio network controllers the new radio network controller does not have knowledge of the base stations suitable for macrodiversity combining so that macrodiversity combining cannot be used before the new radio network controller has established a candidate set of its own. Therefore, transmission power has to be increased and only one transmission path can be used temporarily between the system and the terminal. This degrades the quality of communications and causes stability problems which must be corrected by constant adjustments. Handovers between active base stations serving a terminal can be categorised as follows: 1. handover between base stations (base station sectors) (intra-RNC HO) 2. handover between radio network controllers inside a generic radio network (inter-RNC HO) and 3. handover between generic radio networks (inter-GRAN HO). The present invention primarily relates to handovers between radio network controllers inside a generic radio network (item 2 above). One idea of the invention is that in preparation for a handover a list is compiled in a neighbouring radio network controller of those base stations that would constitute the candidate set should said neighbouring radio network controller be made the active radio network controller. Then the active set AS becomes in conjunction with the handover the new active set AS'. Said list is here called an external base station candidate set. When compiling external candidate sets it is advantageous to use a boundary base station list (BBSL) that can help determine whether a handover is likely. In addition, so-called intense monitoring can be used for an external base station set. The use of an external base station candidate set brings e.g. the following advantages. First, the transmission power change related to the handover is not great at the interface but the use of power is "smooth". This results in small total power consumption in the interface area and low interference-induced noise level. In addition, the solution achieves a continuous state as regards the network so that handovers will not cause deviations from the normal operation and thus a stability problem. Another idea of the invention is that a connection is assigned a radio network controller through which the user data are directed also when some other radio network controller is the active radio network controller. This radio network controller assigned to a connection is here called an anchor controller. If during a connection a base station connected to another radio network controller is chosen the active base station, the user data are directed such that they travel to the active radio network controller via the anchor controller. The use of an anchor controller in accordance with the invention brings considerable advantages compared to the prior art. First, the radio network topology becomes simple and clear, and the network can be easily extended and re-configured. Second, internal traffic events in the radio network are handled within the radio network controlled by the anchor function so that - a handover between radio network controllers is fast so that it is easier to meet the requirements for a seamless and lossless handover and - the load of the mobile switching centre MSC remains moderate. A particularly significant advantage is that the operation of the radio network can be made optimal as regards the use of radio resources. Furthermore, when using an anchor controller, data encryption can be performed in the anchor controller so that encryption keys need not be transmitted during a connection from a radio network controller to another. Transmission routing from the anchor controller to the active radio network controller can be performed by means of chaining so that all active radio network controllers used during a call remain transmission links for the duration of the call. Another alternative is to use optimum routing where radio network controllers between the anchor controller and the active radio network controller are bypassed. Optimum radio network controller routing used in connection with the invention also brings further advantages. First, the internal signalling load of the radio network remains moderate and signalling can be easily made fast enough. In addition, the radio network controller's processing requiremeiits remain reasonable, which makes the solution practical. The method according to the invention is characterised by the features of claim 1. The communications system according to the invention is characterised by the features of claim 10. A communications system radio network controller according to the invention is characterised by the features of claim 12. . Preferred embodiments of the invention are disclosed in the sub-claims. "Active" base station here means a base station that has a user data connection with a terminal. "Active" radio network controller here means a radio network controller with which the active base station is in direct connection so that user data can be transmitted to the active base station. "Old" base station and radio network controller mean a base station or radio network controller that was active before the handover, and "new" base station or radio network controller means a base station or radio network controller which is active after the handover. It is also possible that several radio network controllers are active simultaneously. "Handover" here refers to a handover between base stations, radio network controllers or radio networks. After the handover it is possible that also the old base station / radio network controller remains active. "User data" here means information usually transmitted on a so-called traffic channel between two cellular system users/tenninals or between a cellular system user/terminal and other terminal via a core nerwork. It may be e.g. coded voice data, facsimile data, or picture or text files. "Signalling" refers to communications related to the management of the internal functions of the communications system. The invention is described in more detail with reference to the preferred embodi¬ments presented by way of example and to the accompanying drawing wherein Fig. 1 shows a second-generation cellular system according to the prior ait, Fig. 2 shows the coverage areas of base stations of a second-generation cellular system according to the prior art, Fig. 3 shows a third-generation cellular system, Fig. 4 shows the core network CN of a third-generation cellular system ac¬cording to the prior art and the radio network GRAN in connection with it, Fig. 5 shows the coverage areas of base stations of a cellular system according to the prior art, Fig. 6 shows a flow diagram of the main steps of a method according to the invention for performing a handover between base stations, radio network controllers and radio networks, Fig. 7 shows a cellular system according to the invention and some embodi¬ments for arranging communications between radio network controllers. Fig, 8 shows an embodiment of the invention for arranging communications between radio network controllers of different radio networks by means of the active protocol of the core network, Fig. 9 shows a technique according to the invention for performing the routing between radio network controllers by means of chaining, Fig. 10 shows a technique according to the invention for performing the routing between radio network controllers optimally, Fig. II shows a signalhng flow chart of a backward handover in a cellular system according to the invention. Fig. 12 shows a signalling flow chart of a forward handover in a cellular system according to tlie invention, Fig. 13 shows functions of radio network controllers before a handover in a cellular system according to the invention, Fig. 14 shows functions of radio network controllers after a handover in a cellular system according to the invention, Fig. 15 shows a signalling diagram of a procedure according to the invention for adding a new neighbour base station to the active set during the preparation for a handover, Fig. 16 shows a signalling diagram of a procedure according to the invention for removing a neighbour base station from the active set during the preparation for a handover, and Fig. 17 shows a signalling flow chart of the execution of a handover in a cellular system according to the invention. Figs. 1 to 5 were discussed above in connection with the description of the prior art. Below, a method according to the invention is described briefly with reference to Fig. 6. Then, referring to Fig. 7, a cellular system according to the invention and embodiments for transmitting signalling and user data between two radio network controllers will be described. After that, referring to Fig. 8, it will be disclosed a handover between a radio network controller in a first radio network and a radio network controller in a second radio network. Next, referring to Figs. 9 and 10, it will be disclosed a chained and an optimised embodiment for setting up routing between radio network controllers. Then, referring to Figs. 11 and 12, two embodiments will be described for realising optimised routing. After that, two embodiments will be disclosed for realising macrodiversity combining in a radio network according to the invention. Next, functions of radio network controllers will be described in conjunction with a handover according to the invention with reference to Figs. 13 and 14. Finally, with reference to Figs. 13 to 17, it will be described the steps related to a handover in a radio network employing macrodiversity combining and external candidate set. The description will be followed by a list of abbreviations used in the Figures and in the description. Fig. 6 shows a flow diagram of a method according to the invention for a handover involving the active base station, active radio network controller and active radio network. First, a static configuration 600 of the system is performed comprising the steps below. In step 601, the connections between a switching centre MSC and the radio network controllers are detected, and in step 602 a GRAN-wide routing table for the radio network controllers is created. Then, the fixed connections in tlie radio network GRAN are established in step 603. Then it is performed a dynamic configuration 610 of the radio network, comprising connection setup steps and connection steps as follows. First, an anchor controller is specified, step 611, after which a fixed radio network specific connection between a radio network controller RNC[i] and base stations BS[a(i)...k(i)] is established, step 612, Then, radio connections are set up between radio network controllers RNC[i] and mobile station MS [a], and radio links are set up between base stations BS[a(i)...c(i)] and mobile station MS[a], step 614. After that, possible handovers within The radio network controller are carried out in step 615. If the mobile station receives a strong signal from a base station of an external radio network controller, step 620, a new RNC-to-RNC connection is added, step 621, and the routing is updated and optimised, steps 622 and 623. After that, a radio net¬work controller specific fixed connection is Set up between the radio network controller RNC[j] and base stations BS[aO').-.f(j)]> step 624. Next, radio connections are set up between the radio network controller RNC[j] and mobile station MS[a], and radio links are estabUshed between base stations BS[a(j).,.d(j)] and mobile station MS[a], step 625. In step 626, handover is executed between radio network controllers RNC[i] and RNCQ]. Both radio network controllers can be active as long as it is advantageous to use base stations of boHi radio network controllers. If all signal connections between the mobile station and base stations of a radio network controller are terminated, the radio network controller can be removed from the chain. A radio network controller can also be forced to be removed from the chain when base stations of another radio network controller offers better signal connections. In Figure 6 the radio connection between the radio network controller RNC[i] and the mobile station is removed in step 627, and the radio network controller specific fixed connection between the radio network controller RKC[i] and base stations BS[a(i)...c(i)] is also removed. Fig. 6 also shows a handover (Inter-GRAN HO) between radio network controllers belonging to two different radio networks GRAN A and GRAN B. In the case of such a handover, the dynamic configuration is repeated in the new radio network and the same procedures as in the old radio network are carried out in the new radio network, steps 631 and 632. Fig. 7 shows in closer detail a cellular system's core network CN which comprises a switching centre MSC, and a radio network GRAN connected to the core network. The radio network GRAN comprises radio network controllers aRNC and bRNC and base stations BSl to BS4 connected to them. A terminal TE is connected by radio to the system, via the base stations. It should be noted that Fig. 7 shows only a fracuon of the usual number of radio network controllers and base stations in a radio network. Fig. 7 illustrates some embodiments of the handover according to the invention. When setting up a connection, one radio network controller is made an anchor controller, which in the case depicted by Fig. 7 also serves as active radio network connroller at the initial stage of the connection. The anchor controller is here marked aRNC. The Figure shows a situation wherein a radio network controller bRNC is made the active controller during tlae connection. In an embodiment of the invention the inter-RNC handover signalling messages, like other radio resource management messages within the radio access network as well as the user data, are nransmitted encapsulated via the core network CN. Then the core network CN serves only as a message router and link between two radio network controllers fiinctioning as tunnelling points. The radio network controllers know how to create and decode these messages as well as how to realise the functions requested in them. An advantage of this embodiment is that no separate physical transmission paths are needed between the radio network controllers. In a second embodiment of the invention there exists a physical link between two radio network controllers, such as a cable or radio network connection, for example. Then the handover signalling can be transmitted direct from a radio network controller to another without participation of the core network CN. From the prior art it is known signalling between radio network controllers on protocol layers Ll-L2, which, however, does not take part in the handover signalling proper. A third embodiment of the invention relates to a situation wherein there is no continuous connection between two radio network controllers. Then a solution is applicable where one base station is connected with two network controllers. Thus a base station can actively choose which of the two radio network controllers it sends control messages to. Then a base station can also serve as a mediator between radio network controllers so that messages from a radio network controller to another travel transparently via the base station in both directions. In this case, identification codes are used to distinguish between the messages and traffic proper between the base station and radio network controller. Fig. 8 shows a situation wherein it is needed a handover between radio network controllers of different radio networks. Then the anchor function will not remain in the old radio network but a radio network controller of the new radio network is made the anchor controller. In such a handover, the signalling between two radio networks GRAN can be carried out using an actively participating protocol, such as MAP of the GSM system, for instance. MAP will then communicate separately with the anchor radio network controllers of both GRANs and will process the signalling handover messages related to the handover, like other messages between the core network CN and radio network GRAN. Let us examine a situation in which a terminal is moving in the coverage area of a radio network GRAN. Tlie radio network anchor function then remains in the radio network controller specified for the connection, which means that all messages from the core network to the terminal are first taken to the anchor radio network controller which directs them further via other radio network controllers to the target radio network controller which delivers them to the terminal via a base station. Use of the anchor function requires that the anchor-RNC knows how messages are transmitted to other radio network controllers of the radio network GRAN. This may be realised using a GRAN-wide address mechanism such that the anchor-RNC knows the routing to other radio network controllers, in which case a so-called fixed routing table is used. Alternatively, the radio network controller is connected to only one other radio network controller so that messages are always sent forward until a radio network controller detects from tlie address attached to the message that the message is addressed to it. When using such an arrangement, it has to be taken into account that the anchor-RNC can be any one of the radio network's radio network controllers. In a small radio network it is possible to realise an embodiment of the method that employs only one anchor-RNC, common to all terminals, so that no connection-specific anchor-RNC is needed. Then the anchor-RNC functions as master and the other network controllers function as slaves. If the radio network controller can be selected, the anchor decision can be made either in the core network CN or in the radio network GRAN. Both the core network and the radio network must know which radio network controllers serve as anchors in each of the connections between the terminal TE and switching centre MSC. Figs, 9 and 10 show two embodiments for realising the routing between radio network controllers during different stages of connection. Fig. 9 shows an arrange¬ment for routing the connection by means of chaining, and Fig. 10 shows an arrangement for routing the connection in an optimised manner. In Figs. 9 and 10, circles represent radio network controllers and lines represent connections between radio network controllers, realised e.g. in one of the above-described methods according to the invention. A thick line represents active connection routing between a terminal moving in a radio network and the core network CN. Location of the terminal is only represented in the Figure by the radio network controller. Stages AO and BO in Figs. 9 and 10 represent an initial situation where the terminal communicates with the core network through radio network controllers 100 and 900. Stages Al and Bl represent a situation where the terminal is handed over to radio network controllers 111 and 911 while the anchor remains in the old radio network controller. The advantage of the optimised embodiment can be seen in the situation where the connection of a terminal is further handed over to either an anchor radio network controller or some other radio network controller. In stages A2 and B2 the next handover is to radio network controller 122 and 922. In the chaining method, a new communications link is simply established between the old radio network controller 921 and the new radio network controller 922. In the optimised solution, a new communications link is established between the anchor-RNC 120 and tlie new radio network controller (122), and the link between the anchor-RNC 120 and the old radio network, controller 121 is removed. Stages A3 and B3 illustrate a situation where the connection of the terminal has been handed over back to the anchor-RNC from the initial state of stages A2 and B2. In the optimised case, the communication link between the old radio network controller 132 and the anchor-RNC 130 is removed. Since the new radio network controller is the anchor-RNC, no new communication link needs to be established. In the traditional chaining method, a loop is made from the anchor-RNC 930 back to the anchor-RNC 930 through all the radio network controllers that the terminal has used during the connection. Optimised handover can be carried out in two ways depending on whether it is possible to use the signalling connection with the old radio network controller during the handover. In a so-called backward handover the old radio network controller is used for signalling during the handover, and in a so-called forward handover the old radio network controller is not used for signalling during the hand¬over. Figs. 11 and 12 show some ways of carrying out the above-mentioned back¬ward and forward handovers. The description to follow also refers to handover simations according to Figs. 9 and 10. Abbreviations used in the Figures are listed in the abbreviation list that follows the description. Fig. 11 shows by way of example the signalling flow diagram of an optimised backward handover between radio network controllers. In a backward handover the old connection with the terminal is retained for the whole duration of the handover so that the radio path parameters of the new location can be transmitted to the terminal via the old radio network controller 111. In our example the terminal transits from state Al shown in Fig. 10 to state A2, i.e. from the old radio network controller 111 to the new radio network controller 112. An optimised backward handover according to Fig. 11 between radio network controllers comprises the following steps; A terminal TE requiring a handover between base stations sends a message to the old radio network controller oRNC. When the old radio network controller finds that the new base station required by the terminal belongs to another radio network controller nRNC, it informs tlie anchor controller aRNC about the request for a backward handover. Having received the message from the old radio network controller oRNC the anchor controller aRNC requests the new radio network controller nRNC to reserve fixed and radio connections according to the bearer information (BI) for the terminal. Having received from the new radio network controller an acknowledge for the reservation of connections under the new radio network controller nRNC the anchor controller aRNC negotiates widi the new radio network controller nRNC and they set up the user data transmission link. Next, the anchor controller aRNC requests the old radio netu'ork controller oRNC to send the radio path information of the radio path reserved under the new radio network controller nRNC to the terminal using the old, still operational connection. Having received from the old radio network controller oRNC an acknowledge for the sending of information of the new radio path to the terminal the anchor-RNC requests the new radio network controller to start transmission to the terminal. Finally, the anchor controller aRNC requests the old radio network controller oRNC to release the resources allocated to the terminal. This can be a forced release after the new base station set offers better signal connections, or alternatively the release can be made if none of the base stations of the network controller serves the mobile station. Forward handover Fig. 12 shows by way of example the signalling flow diagram of an optimised forward handover between radio network contTollers. In a forward handover it is assumed that the old connection via the old radio network controller oRNC 111 is no longer in use. In the example accordmg to Fig. 12 a terminal transits from state Al shown in Fig. 10 to state A2, i.e. from the old radio network controller oRNC 111 to the new radio network controller nRNC 112. An optimised forward handover according to Fig. 12 between radio network controllers comprises the following steps: When the terminal and/or new base station nBS find that the terminal needs a handover and the radio network controller nRNC controlling the new base station has detected that the old base station belongs to another radio network controller oRNC, the new radio network controller nRNC sends a message indicating the need for a forward handover to the old base station oRNC either directly (as in Fig. 12) or via the anchor controller aRNC. The old radio network controller oRNC sends a request-acknowledge to the new radio network controller nRNC and informs the anchor controller about the need for a handover. Then the anchor controller aRNC and the new radio network controller nRNC negotiate and set tip a dedicated user data transmission link. Having received from the anchor controller aRNC an acknowledge to its handover request the old radio network controller releases the fixed and radio coimections allocated to the terminal. At latest when the new radio network controller has the user data connections from the anchor controller aRNC up and operational will the new radio network controller nRNC establish the necessary fixed and radio connections between the base station and terminal. Finally, the new radio network controller nRNC sends a message to the anchor controller aRNC indicating that the handover is completed. Use of macrodiversity combining in radio network according to invention Used with a CDMA-type radio network, which facilitates the combining of signals from multiple base stations, or macrodiversity combining, the arrangement according to the invention is characterised by some special features. Macrodiversity combining employs multiple simultaneous coimections, first, between the terminal and base station sectors and, second, between the terminal and individual base stations. On the uplink path the terminal uses one signal and one spread code which is received at several base stations. Alternatively, the terminal may use one signal with several spread codes received at several base stations. The final signal is the result of macrodiversity combuiation. In the downlink direction, several base stations transmit one and the same signal spread using different spread codes to a terminal that performs the macrodiversity combining. The signal connections that provide sufficient signal strength at agreed power levels belong to the so-called active set. If the active set includes base stations connected with different radio network controllers, the macrodiversity combining can be carried out separately for each radio network controller, Then the final signal combination is completed only in the anchor-RNC. In another embodiment the signals are separately routed to the anchor-RNC where the macrodiversity combining proper is carried out. A prerequisite for each diversity combining is rough timmg information, e.g. with the accuracy of 256 chips, indicating the framework within which bit-level signal combining can be performed. Alternatively, macrodiversity combining can be carried out such that the base stations handle the chip-level timing and make the soft bit decisions. These bits, represented by a more detailed representation defined by several bits, are sent to the radio network controller where the combining is carried out using the diversity technique. In a preferred embodiment, packet transmission can be realised in such a manner that same packets are not transmitted via two different base stations. The solution may be such that it is decided on the moment of transmission of each packet which one of the radio paths is the more advantageous one at that moment. The decision may be based e.g. on a prediction on the quality of radio connections, quality calculations or quality measurements. The advantage of macrodiversity combining is then that the better-quality radio transmission path branch is used at each time. Retransmissions caused by failed packet receptions can be further directed e.g. according to the following selection criteria for the radio transmission path branch: - retransmission uses the radio transmission path branch used in the previous transmission, - retransmission uses other than the branch used in the previous transmission or - retransmission uses the branch the quality of which is estimated the best. This is to improve the probability of success through retransinission. An advantage of this embodiment is e.g. a reduced radio path load as the same data normally are not transmitted via two branches. The active set can be limited such that it includes only the base station connections the base stations of which are connected to the same radio network controller. However, this embodiment has the disadvantage that as the terminal crosses the boundary between two radio network controllers, the macrodiversity has to be abandoned momentarily. In an embodiment in which radio network controllers are connected only through the core network CN, macrodiversity combining is advantageously realised in the nearest radio network controller lest it be necessary to transmit unconnected signals via the CN. If the radio network controllers are directly connected, macrodiversity combining according to the invention has two embodiments. The first embodiment covers the cases wherein macrodiversity combining is carried out in successive radio network controllers and finally in the anchor-RNC. The second embodiment covers the cases wherein all signals are separately gathered in the anchor-RNC and macrodiversity combining is carried out there. This embodiment is advantageous in a solution in which the anchor-RNC is the same for all connections in the radio network GRAN and the other radio network controllers are just routers. Mechanisms according to the present invention easily lead to different radio net¬work topologies. However, in the preferred embodiment the radio network is not made topologically complex but it is allowed to utilise as efficiently as possible the core network to transmit its own messages, either passively or actively. As regards the use of radio network resources, it is advantageous to retain a sufficient functional distribution because it is preferable that the radio link layers are located as close as possible to the base stations the signals of which are best detected by the terminal. According to the invention, a radio network controller advantageously has the following new characteristics. - means for realising anchor functions, - means for storing information on routing to other controllers in the radio network, - means for realising data routing to the core network CN, - means for realising data routing to another radio network controller, - means for conmiunicating with another controller, and - means for carrying out macrodiversity combining by choosing the momentarily strongest signal connection or by combining the signals of different connections. Fig. 13 shows radio network controller functions prior to a handover and Fig. 14 shows radio network controller functions immediately after a handover. In the situation represented by Figs, 13 and 14, the radio network controller RNCO is the anchor controller and the radio network controller RNCl is active before the hand¬over and RNC2 is active after the handover. In Figs. 13 and 14, a thick hne in the fixed network represents transmission of user data and a thin line a signalling connection, A thin line between base stations and a terminal indicates measurement operations and a serrated line, or flash symbol, indicates transmission of user data. In addition to the anchor RNC functions (ARNCF) the anchor controller RNCO realises the user data relay (UDR) to the active radio netvi-ork controller. In the active radio network controller RNCl there is a macrodiversity controller (MDC). The active RNCl also includes a macrodiversity combination point (MDCP) for the uplink direction. The corresponding combination point for the downlink direction is located in the terminal TE. The active radio network controller RNCl also contains a set controller (SC). For each terminal there is in the active radio network controller RNCl a candidate set (CS) and, as a subset of the CS, an active set (AS). One or more radio network controllers (RNC2) that control base stations in the immediate vicinity (handover likely) of the base station set of the active radio network controller RNCl may control an external candidate set (ECS). The external candidate set ECS may include one or more base stations controlled by the radio network controller RN'C2, The radio network controller RNC2 includes an external candidate set controller (ECSC) to control the external candidate set. The anchor controller RNCO or the active RNCl (location selectable) includes a so-called set control function (SCF) that monitors tlie need for handover between radio network controllers, prepares the necessary external candidate set ECS and executes the handover. An anchor controller can be established in two alternative manners: * The radio network controller RNC through which the connection was originally set up is chosen the anchor controller. Then, in principle, all radio network controllers may function as the anchor. In practice, this alternative calls for logical RNC-to-RNC connection facilities between all radio network controllers RNC in the radio network GRAN. * Within a radio network GRAN, all anchors are always established in one and the same radio network contToUer, so-called master-RNC, which at the same time is probably the only radio network controller connected with the core network CN. The master-RNC includes the anchor-RKC functions (ARNCF). The master-RNC facilitates a star-like topology for the connections between radio network controllers. The examples illustrated by Figs. 13 and 14 are based on a situation where the anchor has been selected and one active RNC is connected with it which is not an anchoNRNC. The anchor controller RNCO shall have a logical communications connection with both the radio network controller RNCl and the RNC2. The physical realisation of the logical RNC-to-RNC communications connection between the radio network controllers RNCl and RNC2 may be a direct RNC1-RNC2 link or, optionally, the communications between the radio network controllers RNCl and RNC2 can be realised by relaying via the anchor controller RNCO. In Fig. 13 the set control function SCF is located in the anchor controller RNCO so that a logical connection between radio network controllers RNCl and RNC2 is not needed. Other logical RNC-to-RNC connections can be physically realised in the three manners described above (via CN, using RNC-to-RNC cable/radio link, or via base stations). A logical RNC-to-RNC communications connection is in principle independent of the physical implementation. E.g. in optimised routing, where tlie logical communications connection exists between the anchor controller and the active radio network controller, the physical connection can even be relayed via previous active radio network controllers if necessary. The anchor-RNF fiinction ARNCF comprises tasks as follows: - Setting up logical RNC-to-RNC connections between the anchor controller and the active radio network controller, - User data relay UDR, i.e. directing the downlink data to radio network controller RNC2 and receiving the uplinlc data from macrodiversity combination point MDCP-up/RNC2 of the radio network controller RNC2, and - Setting up, controlling and releasing a logical connection between core network CN and radio network. The user data relay UDR comprises tasks as follows: - Relaying traffic between a terminal TE and core network CN instead of base stations controlled by own radio network controller to another radio network controller according to instructions from the anchor-RNC fimction ARNCF. The user data relay controls the user data stream directly or controls the operation of the logical link control LLC. The logical link control LLC controls the radio connections between the radio network controller and a temainal. The tasks of the logical link control LLC include error detection, error correction and retransmission in error situations. In addition, the logical link control LLC comprises control for the necessary buffers and acknowledge windows. The logical link control unit LLC has a generalised meaning; it may terminate the corresponding LLC protocol of the terminal, but it can alternatively serve as an LLC relay. In an LLC relay function the locical link control unit may terminate the messages of the radio network in a normal maimer, but it relays the core network messages (core network data and signalling) further to a defined node of the core network CN. An example of this is relaying messages between a terminal and core network of the General Packet Radio Service GPRS. In this case the Serving GPRS Support Node (SGSN) would serve as a terminating unit. The logical link control LLC can be located such that it is always in the anchor controller. Then there is no need to transmit big LLC buffers within the radio network in connection with a handover of an active radio network controller. Alternatively, the logical link control may be located always in the active radio network contToller, in which case the LLC buffers have to be Uransferred in conjunction with a handover between radio network controllers. Possible transfer of the logical link control from a radio network controller to another is carried out under the control of the user data relay UDR in the anchor controller. The location of the logical link control in the active radio network controller is shown by dashed lines in Figs. 13 and 14, The user data relay LT)R carries out data relaying also in cases where the role of the logical link control is small, e.g. in the so-called minimimi mode, or when the logical link control has no role at all. Possible locations of the logical link conn-ol are also determined in part by the macrodiversity combining used. Radio network controller managers create or remove, depending on the internal implementation method, terminal-specific functions (e.g. ECSC, MDC and MDCP) in the radio network controller and direct the signalling messages to the correct function in the radio network controller. The macrodiversity combination point MDCP and macrodiversity controller MDC represent ordinary functions related to the macrodiversity implementation used. The user data relay UDR is related to inter-RNC communications within the radio network. The anchor-RNC function (ARNCF), which is active only during a handover, belongs to the disclosed anchor-based handover arrangement according to the invention. The set control function SCF, set controller SC and the external candidate set controller ECSC belong to the disclosed arrangement according to the invention that uses an external candidate set. In a macrodiversiiy implementation which comprises on the uplink transmission path only one transmission in the terminal, the macrodiversity combination point MDCP/up is located in the radio network controller. On the downlink transmission path with multiple transmissions (each base station having its own) the macrodiversity combination point MDCP/down is located in the terminal. The macrodiversity combination point MDCP and macrodiversity controller MDC perform the functions that belong to macrodiversity combining according to the macrodiversity implementation used. The functions add and remove base stations from the internal candidate set and from the active set. Furthermore, the macrodiversity controller MDC according to the invention shall be capable of - indicating to the set controller SC the completed additions or removals of base stations to and from the active set of base stations, - adding to/removing from the candidate set visible to the terminal the base stations added to/removed from the external candidate set, - producing for the set controller the necessary radio path quality reports comparable with the external candidate set controller ECSC, and - indicating on request of the set controller SC to the terminal that an entirely new active set (former external candidate set) has been taken into use. The set controller SC carries out tasks as follows: - Checks using the boundary base station list BBSL whether a base station added to/removed from the active set belongs to the so-called boundary base stations of a neighbour radio network controller. - Requests the set control function SCF to realise a creation/removal of an external candidate set in a neighbour radio network controller and to provide the necessary information such as the identity of the base station that triggered the request, the identity of the terminal, etc. - When the external candidate set changes, transmits via the macrodiversity controller MDC to the terminal the information needed by the terminal in the external candidate set measurement. - Provided that intense monitoring is used, produces and transmits information to the set control function SCF that is comparable with intense monitoring controlled by the external candidate set controller ECSC. - Com'eys to the macrodiversity controller MDC the radiotechnical parameters of the external base station set that is about to become active. The macrodiversity controller MDC sends them further to the terminal like the parameters it produced itself. - On request of the set control function SCF, terminates the operation of a terminal in its own radio network controller RNCl or, alternatively, converts the active set of its own radio network controller to the external candidate set of the new active radio network controller RNC2. The set control function SCF comprises tasks as follows: - On request of the set controller SC, allows/forbids, possibly negotiating with, say, the target radio network confroUer, the creation of an external candidate set ECS. - Requests that a neighbour radio network controller create an external candidate set for a certain terminal, transmitting the information (say, base station identity) produced by the active radio network controller to the neighbour radio network controller RNC2. -When creating or modifying an external base station set transmits to the set controller SC the data needed by the terminal in the measurement. - Receives the connection quality reports of the set controller SC and external candidate set controller and makes a handover decision based on them. - Decides on a handover to a neighbour radio network controller or on intense monitoring. - If intense monitoring is possible, requests the external candidate set controller ECSC to start intense monitoring. Requests from the macrodiversity controller the data required for intense monitoring and sends tliem to the extemal candidate set controller. Requests the macrodiversity controller to produce data comparable with the intense monitoring data produced by the extemal candidate set controller ECSC if said data differ from normal reference data. Receives the intense monitoring results from the extemal candidate set controller ECSC and compares them with the quality data received from the set conU-oller SC. - Indicates to the extemal candidate set conUroller ECSC that the handover has been completed and receives the radiotechnical parameters of the active extemal base station set of the extemal candidate set contiroller ECSC and sends them flirther to the set controller SC. - Indicates to the anchor-RNC function ARNCF that the handover has been completed between the two radio network controllers. - When the base station set of the radio network controller RNC2 has become the active set, requests the set controller SC/RNCl of the old radio network controller RNCl to terminate operation and to remove the rest of the fiinctions related to the terminal from the radio network controller RNCl or, alternatively, convert the radio network controller RNCl into an external candidate set controller for the radio net¬work contiroller RNC2. The external candidate set controller ECSC has tasks as follows: - When starting for a given terminal it creates for the base station BS/RNCl, which triggered the preparation, a suitable external candidate set ECS based e.g. on geo¬graphic and/or propagation technical location data and, when the external candidate set ECS exists, updates it constantly according to the base stations added to/removed from the active set. - Conveys to the set control flinction SCF the data required for the external candidate set ECS measurement at the terminal. - In intense monitoring, on the basis of the terminal-specific information produced by the set control function, sets up in the radio network controller RNC2 the fiinctions that are needed in the uplink quality sampling and reports the results of the sampling to the set control function SCF. - As handover starts, sends the radiotechnical parameters of the external base station set becoming active to tlie set conti:ol function SCF. Starts in the radio network controller RNC2 the uplink macrodiversity conti-oller MDC/RNC2 and the macro-diversity combination point MDCP-up/RNC2 needed in the active radio network controller, using the external candidate set as the initial state for the new active set. At the same time establishes the fixed and radio connections needed by the active set. Let us consider the execution of a handover between radio network controllers in the exemplary situation depicted by Figs. 13 and 14. Two phases can be discerned in the handover between radio network controllers: - inter-RNC handover preparation phase and - inter-RNC handover execution phase. The following example of the preparation phase assumes that the set control flinction SCF is in the anchor contiroller RNCO, so a connection between radio network controllers RNCl and RNC2 is not needed. The preparation phase is the same in the uplink and downlink directions. hi the situation depicted in Figs. 13 and 14 the handover preparation comprises the foUovving steps: First, the radio network controller RNCl adds a base station to the active set AS. The signalling flow diagram in Fig. 15 shows one method of adding a base station to the active set. Then the set controller SC/RNCl detects on the basis of tlie boundary base station list BBSL that a base station has been added to the active set which is located in the immediate vicinity of the base stations controlled by a neighbour radio network controller RNC2. The set controller SC/RNCl sends a message about this to the set control function SCF. If this is the first such base station, the set control function SCF requests that an external candidate set controller ECSC be started in the neighbour radio network controller RNC2. Next the radio network controller RNC2 starts the external candidate set controller ECSC for the terminal. Based e.g. on the geographic location data the external candidate set controller ECSC determines a suitable external candidate set ECS for the terminal and sends information about the base stations belonging to the external candidate set to the radio network controller RNCl via the set control function SCF. Alternatively, if there is a direct signalling connection between the radio network controllers RNCl and RNC2, this can be done direct to the set controller SC/RNCl. The set controller SC/RNCl adds the external candidate set ECS to the set of base stations to be measured at the terminal. This is done controlled by the macrodiversity controller MDC/RNCl as in the case of an internal candidate set. After that the terminal uses e.g. pilot signals to perform usual measurements for the base station set that includes the candidate set CS and the external candidate set ECS. In this example it is assumed that the terminal makes a decision or proposition for transferring base stations between tlie active set and the candidate set, and the transfer can be carried out by the macrodiversity combination point MDCP and macrodiversity controller MDC. The set controller SC/RNCl is informed about the transfer. When the macrodiversity controller MDC/RNCl detects the request of transferring a base station belonging to an external candidate set ECS to the active set, the request is transmitted to the set controller SC/RNCl to be further considered or to be executed. If the only boundary base station toward the radio network controller RNC2 is removed from the active set, the set controller SC/RNCl, having detected the situation, removes the extemal candidate set controller ECSC from the radio netv^'ork controller RNC2 by sending a removal request to the set control function SCF/RNCO, Fig. 16, The set control function SCF/RNC then conveys the request to the radio network controller RNC2 which removes the external candidate set controller ECSC. The procedure then starts over again. Otherwise the set controller (SC/RNCl) requests for external candidate set update in the radio network controller RNC2. If the set control function SCF finds that a base station/base stations controlled by the radio network controller RNC2 give(s) a better signal, the set control function SCF may alternatively order a handover between radio network controllers RNCl and RNC2 or only start optional intense monitoring in the radio network controller RNC2. In intense monitoring, a preprocess MDCP' like the macrodiversity combination point is set up in the radio network controller RNC2 for the uplink transmission path, and said preprocess once in a while receives data from the terminal but does not itself transmit data further but only tlie connection quality report to the set control function SCF. Having foimd on the basis of measurements or intense monitoring that a handover is necessary to base station(s) controlled by the radio network controller RNC2, the set control function SCF starts the execution phase of a handover between radio network controller RNCl and radio network controller RNC2. An inter-RNC handover can be carried out as follows: - The active set is completely transferred to the new radio network controller RNC2. Thus only one radio network controller is active at a time. In the handover execution phase the external candidate set ECS2 of the radio network controller RNC2 completely becomes the terminal's active set AS, and the active set ASl and candidate set CSl of the radio network controller RNCl are removed. Optionally, the active set AS of the radio network controller RNCl may remain as candidate set ECSl. This arrangement avoids the problem of RNC synchronisation found in hierarchic combining. - In hierarchic combining, each radio network controller has an active set of its own. All active radio network controllers perform their own combining for the data in the uplink direction. Final uplink combining can be carried out in radio network controller RNCO. Then it is not necessary to establish a macrodiversity controller proper MDC/RNCO in the radio network controller RNCO or functions equivalent to a macrodiversity combination point MDCP-up/RNCO, if the combmation points of the active radio network controllers are able to preprocess the final result for fixed transmission in such a manner tliat final combining is easy to perform in the radio network controller RNCO. Alternatively, one of the active radio network controllers may serve as a so-called combination anchor, combining the user data of the other active radio network controllers prior to the transmission to the radio network controller RNCO. The user data relay UDR/RNCO has to duplicate the downlink user data for the downlink connection combined in the terminal. Additionally, the base stations of the active sets of the different radio network controllers must be synchronised as required by the CDMA method used. Hierarchic combining may comprise several hierarchy levels. - A combination of the alternatives described above is used e.g. in such a manner that the downlink direction employs complete transfer of active set and the uplink direction employs hierarchic combining. Then in the downlink direction user data are transmitted via the previous active set until measurements show that the new base station set is better. Then the downlink data will be transmitted via the new set. By means of this solution, the advantages of hierarchic combining are retained in the uplink direction but data duphcating is avoided in the downlink direction. Following example of the execution phase of an inter-RNC handover is based on the complete transfer of the active set both in the uplink and m the downhnk directions (alternative 1), The execution phase example assumes that the set control function SCF is located in the anchor controller RNCO so that no logical RNC-to-RNC connection is needed between the radio network controllers RNCl and RNC2. The execution phase example is based on the use of macro diversity in a generic CDMA system. The example is illustrated by the message flow diagram in Fig. 17. In the example discussed here the handover execution comprises the following steps after the set control function (SCF) has made the handover decision. First, the anchor function ARNCF of the anchor controller RNCO sets up a logical RNC-to-RNC connection between the anchor controller RNCO and the new active radio network controller RNC2. Then the set control function SCF informs the radio network controller RNC2 about the execution of the handover. The external candidate set controller ECSC sends to the set control function SCF or, alternative¬ly, direct to the old set controller SC/RNCl the radiotechnical parameters of the active-to-be base station set to be further transmitted to the terminal. Internal operation of the radio network contToller RNC2 is mostly the same as in conjunction with the set-up of a normal call with the difference that the external candidate set is immediately made the final active set. Listead of an external candidate set, a set controller SC/RNC2, macrodiversity controller MDC/RNC2 and macrodiversity combination point MDCP/RNC2 are established for the uplink direction. Controlled by the radio network controller RNC2 it is reserved or created terminal-specific fixed bearers needed for user data transmission between the radio network controllers and the base stations in the active set as well as radio bearers between base stations and the terminal in manners used in the radio network unless such connections have already been completely created in intense monitoring of the preparation phase, On request of the set control function SCF the user data relay UDR in the anchor-RNC function ARNCF modifies its operation as follows. The user data relay UDR prepares to receive the uplink user data from the macrodiversity combination point MDCP-up/RNC2 of the radio network controller RNC2. The user data relay UDR directs the downlink user data also to the radio network controller RNC2. Next, the set control function SCF/RNC2 sends to the set controller SC/RNCl of the radio network controller RNCl the parameters (such as the time reference and the scrambling and/or spreading code used) of the pilot signals of the base stations in the active set of the radio network controller RNC2. The set controller SC/RNCl in the radio network controller RNCl sends to the terminal the parameters of the new active set. Then the macrodiversity combination point MDCP/RNC2 in the radio network controller RNC2 starts transmission with the new active set AS/RNC2. This is ac¬knowledged to the anchor-RNC function ARNCF via the set control function SCF. Finally, the anchor function ARNCF may request the radio network controller RNCl to remove the terminal's set controller SC/RNCl, macrodiversity conuroller MDC/RNCl and macrodiversity combination point MDCP/RNCl as well as to release the terminal-specific fixed bearers between the radio network controllers and base stations and possible remaining radio path reservations. Alternatively, the anchor contioller may request the radio network controller RNCl to turn the active set of the radio network controller RNCl into an external candidate set ECS. This having been acknowledged, the inter-RNC handover is completed. In the examples discussed above it is assumed that the frequency of the external candidate set ECS complies with re-use 1, typical of a CDMA system, so that the external candidate set has the same firequency as the candidate set proper. It is however possible to establish an external candidate set at another frequency. Then, the active set AS of only one candidate set can be in use. Even if macrodiversity combining were not an advantageous solution between different frequencies, this embodiment still facilitates the change from candidate set AS to new candidate set AS' in accordance with the principles set forth above. The present invention can be used in connection with a great number of applications. These include e.g. database search services, data downloading, video conferencing, "on demand" data purchases from a communications network, use of world wide web services in the Internet including web browsing etc. The embodiments discussed above are naturally exemplary and do not limit the invention. For example, the terminal may comprise a mobile station, portable terminal or a fixed terminal, such as the terminal of a cordless subscriber connection. Particularly it should be noted that the creation of an external candidate set for an inter-RNC handover can be carried out independently of whether data communic¬ations will be routed to the new active base station via another radio network controller, such as an anchor controller. The steps of the above-described method according to the invention can also be carried out in an order otlier than that given above and some steps may be skipped as unnecessary. Above it was discussed embodiments wherein the radio network employs the CDMA system. However, it should be noted that the present invention is in no way limited to the CDMA system but it can be utilised in other systems as well, such as the TDMA system, for example. List of abbreviations used in Figures and description CN Core Network GRAN Generic Radio Access Network TDIvLA. Time Division Multiple Access CDMA Code Division Multiple Access TE Terminal Equipment BS Base Station nBS new Base Station oBS old Base Station BSC Base Station Controller RNC Radio Network Controller nRNC new Radio Network Controller oRNC old Radio Network Controller aRNC anchor Radio Network Controller aRNCF anchor Radio Network Controller Function bRNC active Radio Network Controller which is not anchor RNC UDR User Data Relay CS Candidate Set AS Active Set ECS External Candidate Set ECSC External Candidate Set Controller MDC MacroDiversity Controller SC Set Controller SCF Set Control Function BBSL Boundary Base Station List MDCP MacroDiversity Combination Point RI Radiopath Information BI Bearer Information ID IDentity HO Handover ack aclcnowledge up uplink down downlink req request resp response WE CLAIM : 1. A method for controlling radio communication between a terminal (MS, TE) and a communications system (CN, GRAN) in a communications system wherein the communications connection between the system and the terminal is established via an active radio network controller (RNC) and active base station (BS), wherein the active radio network controller maintains a first list, "a candidate set", of base stations for selecting an active base station for the connection, characterised in that prior to handover between radio network controllers a second list, "external base station candidate set", is compiled in a neighbouring radio network controller of those base stations that would constitute the candidate set should said neighbouring radio network controller be made the active radio network controller, and the candidate set of the new active radio network controller is established on the basis of said eternal base station candidate set. 2. The method of claim 1, wherein the communications connection is directed to said active radio network controller via a second radio network controller (621-628). 3. The method of claim 1 or 2, wherein a handover between radio network controllers comprises a preparation phase and execution phase. 4. The method of claim 3, wherein for the connection there is a set of active base stations, and said preparation phase includes a step for adding a base station to the set of active base stations. 5. The method of claim 3 or 4, wherein for the connection there is a set of active base stations, and said execution phase includes steps for changing the active radio network controller and the active base station set. 6.' The method of claim 3 or 4, wherein said execution phase includes steps for keeping at least two radio network controllers and their base station sets active. 7. The method of claim 5, wherein the method comprises the step of changing the active radio network controller and active base station set completely. 8. The method of any one of the preceding claims, wherein for a handover in which a first radio network controller is active before the handover and a second radio network controller is active after the handover the method comprises the steps in which : before the handover a suitable external base station candidate set is determined for the terminal in the second radio network controller, and information about the base stations belonging to the external base station candidate set is transmitted to the first radio network controller. 9. The method of any one of the preceding claims, wherein the external base station candidate set is added to the set of base stations to be measured at the terminal. 10. A communication system comprising a radio network controller, said radio network controller (RNC) comprising means for routing communication to another radio network controller during a connection and means for compiling a first list, a "candidate set", of base stations for selecting an active base station for the connection when the radio network controller is active in the connection, characterised in that the radio network controller comprises means for creating a second list, "an external base station candidate set" of those base stations at would constitute the candidate set should the radio network controller be made the active radio network controller in the connection, and means for creating new candidate set based on said external base station candidate set. 11. The communication system of claim 10, comprises a first and a second Radio Network Controller (RNC) and base stations (BS) connected to the radio network controllers to provide a communications connection between the system (CN, GRAN) and a terminal (TE) connected with it, wherein an active radio network controller comprise means for compiling a first list, a "candidate set", of base stations for selecting an active base station for the connection, wherein it comprise means in a neighbouring radio network controller for creating a second list, "an external base station candidate set" of those base stations that would constitute the candidate set should said neighbouring radio network controller be made the active radio network controller and means in said neighbouring radio network controller for creating a new candidate set based on said external base station candidate set. 12. The communication system of claim 11, comprises means for directing the communications connection to an active radio network controller (bRNC, RNCl, RNC2) via another radio network controller (aRNC, RNCO). 13. The communication system of claim 10, comprises means for creating a boundary base station list of base stations which are located in the immediate vicinity of base stations controlled by a neighbour radio network controller , and means for requesting the neighbouring radio network controller to start creating an external base station candidate set when a base station in the boundary base station list becomes active. 14. The communication system of claim 10 or 13, comprise means for sending information about the base stations belonging to the external base station candidate set to an active radio network controller. 15. The communication system of claim 10, 13 or 14, comprise means for turning an external base station candidate set into a candidate set. 16. The communication system of any one of claims 10 to 15, comprises a logical link control unit (LLC) for controlling communication between radio network controller (RNC) and a terminal (TE). 17. The communication system of any one of claims 10 to 16, comprises a set controller (SC). 18. The communication system of any one of claims 10 to 17, comprises an external base station candidate set controller (ECSC). 19. The communication system of any one of claims 10 to 18, comprises a macrodiversity controller (MDC). 20. The communication system of any one of claims 10 to 19, comprises a macro diversity combination point (MDCP).

Documents:

100-che-2005 abstract duplicate.pdf

100-che-2005 abstract.jpg

100-che-2005 abstract.pdf

100-che-2005 claims duplicate.pdf

100-che-2005 claims.pdf

100-che-2005 correspondence-others.pdf

100-che-2005 correspondence-po.pdf

100-che-2005 description (complete) duplicate.pdf

100-che-2005 description (complete).pdf

100-che-2005 drawings duplicate.pdf

100-che-2005 drawings.pdf

100-che-2005 form-1.pdf

100-che-2005 form-18.pdf

100-che-2005 form-26.pdf

100-che-2005 form-3.pdf

100-che-2005 form-5.pdf