In the following section, we refer in particular to the GPRS (General Packet Ra- dio Service) wireless communication system, but the present invention is not limited thereto being the present invention applicable to any kind of wireless communication including other extensions of GPRS such as the evolution of GPRS radio access, i. e., GERAN (GSM Edge Radio Access Network), or the application of GPRS concepts in the 3GPP (Third Generation Partnership Project) network.

GPRS (General Packet Radio Service) network has being standardized at the time of filing the present application by ETSI (European Telecommunications Standards Insti- tute), having issued the following documents: [GSM 04.60], ETSI EN 301 349 V8.0.0,"Radio Link Control/Medium Access Control (RLC/MAC) protocol", Digital cellular telecommunications system (Phase 2+); General Packet Radio Service (GPRS); Mobile Station (MS)-Base Station System (BSS) interface; (version 8.0. 0 Release 1999).

In particular the present invention relates to a method for accessing the PRACH chan- nel (Packet Random Access CHannel) of the GPRS system in order to optimize the throughput on the PRACH, minimising the access delay and reducing the number of loss requests due to collisions.

The method of the invention does not require any change on the fixed part of the sys- tem side, and it has a low impact on the wireless communication device side.

During the last few years, we have assisted at the great success of mobile conversa- tional services. Mobile phone is now the most common device and it is almost excep- tional to find someone without at least one mobile.

Next challenge for mobile industry is to provide mobile networks for data communica- tions. In fact, in the next five years, the traffic load generated by packet mobile services is foreseen to equal that generated by mobile circuit-based services.

The efficient and optimised support of data packet services is one of the main objec- tives in standard bodies and for defining the specifications for the next generation of mobile networks such as 3GPP (3rd Generation Partnership Program), 3GPP2, IMT- 2000, IETF (Internet Engineering Task Force), and MWIF (Mobile Wireless Internet Fo- rum). At the same time, the mobile operators around all Europe are preparing, at the time of filing this application, the launch of GPRS (Generic Packet Radio Service) to provide mobile data service.

General Packet Radio Services (GPRS) offers an efficient utilisation of radio resources for packet services characterised by a discontinuous bit rate generation. The basic idea beyond GPRS is to use the radio time slots of GSM access unused from voice service to carry in a packet mode fashion asynchronous data. The allocation of channels is flexible : the network can allocate from 1 to 8 time slots and rates that can be theoreti- cally up to 160 kbit/s. Active users share same radio resources and up-and down channel may be reserved separately.

GPRS uses the same radio Base Station (BTS) as voice with HW and SW upgrades in the BSC (Base Station Controller), and a completely new core network. This choice allows leaving untouched the radio access elements that represent the main invest- ment for an operator. The new core network elements are the Serving GPRS Support Node (SGSN), a router responsible for terminals in a given region and the Gateway GPRS Support Node (GGSN), a router linked to an external data network (e. g., Inter- net) and responsible for routing packets to the appropriate SGSN.

As radio is a limited resource, the efficient utilisation of radio access is an important point and it can be bottleneck of the system. Thus, the good configuration of the control and data channels on the radio access is a critical issue.

Let assume that a mobile terminal has already a communication context with the GPRS network.

Each time the mobile has a packet to transmit, it has to open a second communication context with the radio access network and initiate a TBF (Temporary Block Flow) es- tablishment. This context is opened each time the mobile has to transmit packets to the network and it is released when the transmission of the packets has been completed.

The TBF establishment can be requested by an RLC (Radio Link Control) message called PCR (PACKET CHANNEL REQUEST) sent on the PRACH (Packet Random Access CHannel). PRACH is one of the Control CHannel (CCH) of the GPRS radio ac- cess.

The current specification defines in section 7.1.2.1.1 of [GSM 04.60] the algorithm to

access the PRACH and to initiate a TBF (Temporary Block Flow) establishment by the mobile station on PCCCH (Packet Common Control CHannel).

This algorithm consists of the main following steps: Step 1J When the mobile has an LLC (Link Layer Control) frame to transmit, but without having a TFB already allocated, it sends a PCR (PACKET CHANNEL REQUEST) message on the PRACH.

Step 2) The first attempt to send the PCR message is done in the first possible TDMA frame containing PRACH on PDCH (Packet Data Traffic CHannel) matching the mobile sta- tion's PCCCH_GROUP if the following test is passed. The test requires that the Per- sistency Level P (i) value is less or equal than the number R uniformly selected by the mobile in the interval [0,15], i. e, P (i) P MJ The P (i) is defined by the network, paged to the mobiles and can have four different values related to the four different priority classes i s [1-4]. The default value of P (i) is 0.

For instance, if we set (P [1] =0, P [2] =3, P [3] =7, P [4] =16), it means that a mobile requir- ing resources with the highest priority (1) will ever pass the test and always transmit in the first TDMA frame of the PRACH, the mobile with radio priority 2 and 3 will have re- spectively probability 3/4 and 1/2 to pass the test, while the mobile requiring a TFB with the lowest priority will never pass the test.

Step 3) If the test [1] is not successfully, the terminal has to wait. The value of TDMA frames of the PRACH it has to wait is obtained by extracting a sample in the range [S, S+T-1]. S and T values are determined by extraction from two sets of numbers defined in [GSM 04.60 Section 12.14]. Therefore, if the test [1] is not successful, the number of TDMA frames of PRACH it has to wait before it can send a request on the media varies be- tween 12 and 267.

Step 4) The mobile repeats the same procedure of step 3 for almost MAXRETRANS (the maximum number of retransmission) times to schedule retransmissions of the PCR message until it receives response from the network. Retransmissions are needed due to both collisions with PCRs of other mobiles and transmission errors on the channel.

MAXRETRANS vatue depends on the radio priority and it is defined in [GSM04. 60

Section 12.14].

Step 5) If either the maximum number of MAXRETRANS or the timer T3186 expires, the ac- cess procedure is aborted and the mobile starts a cell reselection.

The algorithm is efficient and works well when PRACH slots are uniformly distributed in the PDCH [see Fig. 2 case 1] and, thus, the entire PDCH is dedicated to PRACH channel.

That is not the case of other configurations where only a few PDTCH frames are de- voted to PRACH as shown in Fig. 2, case 2 and case 3.

In fact, let consider the PRACH configurations of Fig. 1 case 2 and case 3 and let as- sume that LLC arrival is a random process and, then, the access procedures are initi- ated randomly during a frame.

As TDMA channels for a PRACH block are grouped into four, if PCR messages are transmitted in the first TDMA slot, then, in the first TDMA slot of the PRACH Block there is a higher probability to have collisions due to PCR messages than in the other TDMA slots. This effect can be observed in Fig. 3 in case 2 where the PRCs related to the transmission requests of LLC frames 1 and 2 and in case 3 where the PRCs re- lated to the transmission requests of LLC frames 1,2 and 3 are transmitted in the first slot of the PRACH Block generating collision.

Background Art In the current ETSI specification, it is possible to decrease collisions on the first of the four slots in the PRACH Block by increasing the values of Persistency Level, P (i). In fact, the higher the P (i) value the higher the probability to reschedule PCR transmission randomly in the next slots of the PRACH.

However, this strategy leads to a higher access delay and it increases the probability to abort the access procedures due to the expiration of timer T3186.

The effect of this known strategy is therefore to augment the access delay to the PRACH channel and to increase the probability to abort the access procedure.

Object and Summary of the Invention Object of the present invention is to overcome the drawbacks of the approach dis- closed in the above mentioned ETSI specification aiming to decrease collisions on the first of the four slots in the PRACH Block and to define a method adapted to avoid colli- sions.

According to the invention, the collisions technical problem have been solved using the method disclosed in the preamble of the appended claim 1, addressed to a method

adapted to reduce collisions for every configuration of PRACH on the Multiframe (see Fig. 1). The claimed method brings to a significant gain in terms of load on the PRACH channel, the minimization of the access delay, and the reduction of the probability to terminate the access procedure due to expiration of time out.

As mentioned before, currently, the ETSI specification [section 7.1.2.1.1 of GSM 04.60] suggests to perform the first attempt to send the PCR message in the first possible TDMA frame containing PRACH on PDCH (Packet Data Traffic CHannel) matching the mobile station's PCCCH_GROUP.

According to the invention, instead to require the transmission in the first possible TDMA of the PRACH (STEP 2), is suggested to select randomly one of the TDMA slots that compose the first PRACH Block.

According to the above, considerable advantages are achieved with the present inven- tion when compared with method and systems of prior art and these advantages will be obvious by observing the graphs contained from Fig. 4 to Fig. 8 disclosed in details hereinafter.

Brief Description of the Drawing The invention, together with further objects and advantages thereof, may be under- stood with reference to the following description, taken in conjunction with the accom- panying drawings and in which: Fig. 1 is an illustration of the 52 Multiframe on one PDCH.

Fig. 2 is a first illustration of one PDCH of the 52 Multiframe with different con- figurations of PRACH.

Fig. 3 is a second illustration of one PDCH of the 52 Multiframe with different configurations of PRACH, example of LLC frames arrival and PCR requests on the PRACH transmitted according to the method of the invention.

Fig. 4 is a diagram showing the Served load (D) versus Effective load (G) for PRACH configuration of Fig. 2 case 3 and Priority=1.

Fig. 5 is a diagram showing the Served load (D) versus Effective load (G) for PRACH configuration of Fig. 2 case 2 and Priority=4.

Fig. 6 is a diagram showing the mean access delay (T) in [ms] versus offered load (O) for PRACH configuration of Fig. 2 case 3 and Priority=1.

. Fig. 7 is a diagram showing the mean access delay (T) in [ms] versus offered load (O) for PRACH configuration of Fig. 2 case 2 and Priority=4.

. Fig. 8 is a diagram showing the probability to abort the access procedure (p)

versus offered load (O) for PRACH configuration of Fig. 2 case 3 and Priority=4.

Detailed Description of a Preferred Embodiment of the Invention Fig. 1 shows the GPRS 52 Multiframe on a PDCH. Each GPRS 52 Multiframe carries 12 RLC Blocks. Each RLC block consists of 4 TDMA slots. (12X4 RLC slots+ 4 Idle frame = 52 Slots).

Fig. 2 shows a generic PDCH of the GPRS 52 Multiframe with different configurations of PRACH. In Case 1, PRACH slots are uniformly distributed in the PDCH and the en- tire PDCH is dedicated to PRACH channel. Case 2 and Case 3 with 6 and 2 PRACH blocks of the PDCH are dedicated to the PRACH channel.

Let consider the following performance metrics: O the offered load on the PRACH, which denotes the mean arrival of PCR messages in a Multiframe to the number of PRACH slots in the Multiframe, D the served load by the PRACH, which is the mean number of successfully transmitted PCR messages which do not encounter collisions in a Multi- frame to the number of PRACH slots in the Multiframe, G the effective load on the PRACH which includes the additional load gener- ated by the retransmission of PCR messages, T the access time to the PRACH [ms], i. e., the time interval between an LLC request and the successful access of the PCR message on the PRACH, p the probability to abort the access procedures due to collisions.

Two different configurations of PRACH shown in Fig. 2 have been taken into account: case 2 with 6 PRACH blocks for Multiframe, and case 3 with 2 PRACH blocks for Multiframe.

For the metrics defined in the previous list, we have compared in the following draw- ings the performances of the current algorithm (current) with the improvement of the present invention (proposal) with a homogenous traffic load generated randomly with: radio priority 1 (i. e., MAUX TRANSOM =1) and P (1) =1 and P (1) =5, radio priority 4 (i. e., MAX TRANSM =7) and P (1) =1 and P (4) =5.

The evaluation study has been obtained by simulation.

Fig. 4 is a diagram showing the Served load (D) versus Effective load (G) for PRACH configuration of Fig. 2 case 3 and Priority=1. Dotted lines refer to performances ob- tained with the current proposal with respective P (1) =1 and P (1) =5. Continuous lines refer to performances obtained with our proposal with respective P (1) =1 and P (1) =5.

Fig. 5 is a diagram showing the Served load (D) versus Effective load (G) for PRACH

configuration of Fig. 2 case 2 and Priority=4. Dotted line. s refer to performances ob- tained with the current proposal with respective P (4) =1 and P (4) =5. Continuous lines refer to performances obtained with the method of the invention with respective P (4) =1 and P (4) =5.

Fig. 4 and Fig. 5 show therefore served load (D) vs. the effective load (G) with the cur- rent method specified in [GSM 04.60] and with the method of the invention respectively case 2 of channel configuration (Fig. 5), and case 3 of PRACH configuration (Fig. 4).

As it can be noted, the current algorithm (dotted lines) obtains better performances in terms of served load with the higher value of Persistency Level.

The method of the invention provides always higher served load values with respect to that of the current specification, while the served load is not sensitive to Persistence Level values.

Regarding the configuration of the PRACH of course we achieve better performance with the lower number of PRACH blocks in the Multiframe, i. e., PRACH configuration case 3.

Fig. 6 is a diagram showing the mean access delay (T) in [ms] versus offered load (O) for PRACH configuration of Fig. 2 case 3 and Priority = 1. Dotted lines refer to per- formances obtained with the current proposal with respective P (1) =1 and P (1) =5. Con- tinuous lines refer to performances obtained with the method of the invention with re- spective P (1) =1 and P (1) =5.

Fig. 7 is a diagram showing the Mean access delay (T) in [ms] versus offered load (O) for PRACH configuration of Fig. 2 case 2 and Priority=4. Dotted lines refer to perform- ances obtained with the current proposal with respective P (4) =1 and P (4) =5. Continu- ous lines refer to performances obtained with the method of the invention with respec- tive P (4) =1 and P (4) =5.

Fig. 6 and Fig. 7 show therefore the access delay versus the offered load on the PRACH with the current method specified in [GSM 04. 60] and with our proposal with respectively 'case 2 of PRACH configuration (Fig. 7), and case 3 of PRACH configuration (Fig. 6).

Considering the access delay, the mean access delay values obtained using the method of the present invention (continuous line) are much better than that provided by the current method (dotted lines).

For both for case 3 (Fig. 6) and case 2 (Fig. 7), the claimed method achieves mean ac- cess delay values even than two times lower with respect to the current algorithm.

The original proposal has better performances in terms access delay with the low value of Persistency Level, but it produces lower values of offered load.

On the contrary, with the method of the invention it is possible to maximise both the throughput and minimise the access delay.

Fig. 8 is a diagram showing the probability to abort the access procedure (p) versus offered load (O) for PRACH configuration of Fig. 2 case 3 and Priority=4. Dotted lines refer to performances obtained with the current proposal with respective P (4) =1 and P (4) =5. Continuous lines refer to performances obtained with our proposal with re- spective P (4) =1 and P (4) =5.

Fig. 8 shows therefore the probability to abort the access procedures due to collisions versus the offered load with the current method specified in [GSM 04.60] (dotted lines) and with the method of the invention (continuous lines) with case 3 of PRACH configu- ration.

The original proposal obtains good performances in term probability to abort the access procedures due to collisions with the highest value of Persistency Level P=5 but this setting leads to high values of access time.

The method of the invention produces lower probability values to abort the access pro- cedures with respect to the values of the current method minimising both the access delay and maximise the served load.

Therefore, while a particular embodiment of the present invention has been shown and described, it should be understood that the present invention is not limited thereto since other embodiments may be made by those skilled in the art without departing from the above mentioned object. It is thus contemplated that the present invention en- compasses any and all such embodiments covered by the following claims.