SYNCHRONIZATION OF BASE STATIONS IN A CELLULAR TELECOMMUNICATION SYSTEM TECHNICAL FIELD OF THE INVENTION The present invention relates to a method for synchronizing a radio transmitter with other receivers in a network, in order to minimize the interference effect of other transmitters. Particularly, the invention relates to the synchronization of a network of base stations in a mobile communication system. DESCRIPTION OF THE RELATED ART In a digital cellular radio system operating in accordance with the TDMA principle, such as GMS, D-AMPS or PDC, radio messages are transmitted in frames from base stations, each frame includes a given number of time segments. The transmissions in the different time segments are generally intended to be received by different mobile radio receivers, and therefore it is necessary to ensure that the receiver is synchronized with the transmitter. The mobile transceivers are synchronized with their respective base stations by means of a signal from the base station. For example, in the D-AMPS of total speed, 3 mobile receivers share the same frequency channel, and therefore each channel is divided into three time segments, each with a duration of 6.7 ms, 3 segments forming a frame. The frames are repeated 50 times per second. Each time segment in the table is assigned a particular mobile receiver, until the call is released or until the mobile is transferred to another channel, for example, in another cell. In each time segment, 324 bits are transmitted, most of which are data bits, but 28 of which form a synchronization word. The standard published by the Electronics Association as publication IS136 of EIA / TIA, which specifies the D-AMPS system, defines 6 different synchronization words, but only 3 of them are used for a channel operating at full speed. Accordingly, a different synchronization word is assigned to each time segment in a frame, and the base station transmits the relevant synchronization word once during each time segment. The mobile receiver can recognize the transmissions provided for it by identifying the synchronization word, and similarly includes the same synchronization word in its own transmissions to the base station. The synchronization words in IS136 are chosen in such a way that there is a minimal correlation between them. Accordingly, there is very little chance that a receiver mistakenly identifies a synchronization word transmitted with a different synchronization word. However, there is a danger that a receiver will receive the expected synchronization word from an interfering transmitter operating on the same frequency, and misinterpret it as its expected synchronization word. In addition, there is the possibility that a receiver mistakenly interprets data sent by an interfering transmitter on the same frequency as its expected synchronization word. Attempts were made in the prior art to overcome these problems. One known possibility is simply to allow each base station tranceptor to select its own timing, which means that there is no synchronization between the base stations. In this situation, it is possible, even if unlikely, that an interfering transmitter is transmitting the same synchronization word with a signal level sufficient to cause an interference, and at a point of time sufficiently close to the expected time to cause the possibility of a wrong synchronization. A known alternative possibility is the synchronization of the entire network, such that each station is transmitting the same synchronization word at the same time. This increases the probability that the synchronization word is mistakenly received from an interfering transmitter, and its interpretation as if it were its own word of expected synchronization. The probability of a wrong identification in this way depends on the carrier-to-interference ratio (C / I), which refers to the relative signal levels of the transmissions from a desired transmitter, and from an interfering transmitter that operates on the same frequency. Since there is only a limited number of frequencies available for use in a system, it is necessary to reuse the frequencies. Frequency planning can optimize the frequency rejection distance, and therefore optimize the C / I ratio, but in general it can not guarantee that the C / I ratio is sufficiently high to avoid the possibility and interference through the wrong detection of the synchronization word from an interfering transmitter. It is an object of the present invention to increase the probability of correct synchronization by planning the use of synchronization words. COMPENDIUM OF THE INVENTION A network according to the present invention comprises several base stations, some of which use the same frequency. According to the invention, the base stations are synchronized in such a way that their time slices coincide and the network is planned in such a way that the neighboring base stations operating on the same frequency transmit different synchronization words during any segment of time. given time. The invention also relates to the method of coordinating the coordination positions of the air box positions of the base stations and to the base stations themselves. The advantage of the invention is that the risk of erroneous synchronization can be minimized, even under less favorable C / I conditions. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a schematic block diagram of a telecommunications network according to the invention. Figure 2 represents transmissions by a base station at a particular operating frequency. Figure 3 represents the coverage of an area by a cellular communication network. DETAILED DESCRIPTION OF PREFERRED MODALITIES Figure 1 represents a cellular communications network having a Mobile Services Switching Center (MSSC) 2, which is in communication with base stations 4, 6, 8, 10 (known as BSl, BS2, BS3 and BS4, respectively) and that controls said base stations. It will be noted that the network includes many more base stations, and only 4 base stations are presented here to facilitate the explanation. Figure 2 shows transmissions by a first base station BSl. In each frame, of a duration of 20 ms, 3 time segments are observed, each assigned to a different mobile receiver. Communications are a first mobile receiver in the CH1 channel are performed in the first time segment, communications with a second mobile receiver in a CH2 channel are made in the second time segment, and communications with a third mobile receiver in a third CH3 channel is performed in a third time segment. Each time segment consists of 324 bits, 28 of them form a sync word SYNC. Each time segment contains a different synchronization word in order to allow the mobile receivers to identify the transmissions that they should receive. Similarly, transmissions from the mobile receivers to the base stations also contain the synchronization words. Due to the limited availability of frequencies, each frequency must be reused by several base stations within the network. Accordingly, in FIG. 1, the base stations BSl, 'BS2, BS3 and BS4 all operate on the same frequency. Figure 3 shows a part of an area covered by a mobile telecommunication network. The area can be considered as covered by a large number of generally hexagonal cells, each of which contains a base station. (Those skilled in the art will know that, otherwise, the base stations may be located in cell boundaries, radiating different operating frequencies in the different adjacent cells). Figure 3 shows the Cl, C2, C3 and C4 cells, where the base stations BSl, BS2, BS3 and BS4 are located respectively. To reduce the possibility of interference between base stations transmitting on the same frequency, the network is planned in such a way that base stations using the same frequency are separated by the greatest possible distance. In this case, the C cells! And C2 are separated by the Minimum Frequency Rejection distance. Thus, a mobile receiver in the cell Cl will attempt to detect transmissions from the base station BSl, at the particular operating frequency of this base station, which contains a particular synchronization word. However, there is a danger that, on the contrary, the receiver will detect the signal transmitted from the base station BS2 in the cell C2 or from the base station BS3 in the cell C3 (or from any other station). of base that uses the same frequency of operation). The likelihood of this happening depends on the ambient reception conditions, but it also depends crucially on the distance D rejection of frequencies. In this context, the other base stations, transmitting signals that may interfere with the desired transmission at the predicted frequency are known as co-channel interferers. In the case where the mobile receiver mistakenly interprets data from a co-channel interfering device as the desired synchronization word, the receiver may lose the required timing, or may incorrectly start his demodulator, resulting in deterioration of the demodulator performance. In accordance with the invention, the MSSC 2 transmits control signals to all base stations, causing them to transmit their time slots simultaneously or at least within a pair of symbols between them. That is, the base stations share the same V air box clock ". Thus, each base station is transmitting one synchronization word more or less at the same time as the other base stations, or at least more or less at the same time as the other base stations operating on the same frequency, it will be noted that control signals from a central unit to the base stations can also be transmitted via satellite, if preferred. The base stations can also use GPS clock signals, in which case the base stations will need to be equipped with suitable receivers to detect signals from GPS satellites.In the modality described to date, the network uses three different words of synchronization, each one associated with a respective time segment The MSSC2 controls the base stations in such a way that, as far as possible, internal devices adjacent co-channel listeners are not transmitting the same synchronization word at the same time as the others. Therefore, in the network illustrated in FIG. 3, the base station BSl in the cell Cl can at a particular time be transmitting a first synchronization word, which will be followed in the next two time segments by a second synchronization word and a third word of synchronization. The base station BS2 in the cell C2 can at the same time be transmitting the second synchronization word that will be followed, in the next two time segments by the third synchronization word and the first synchronization word. A base station BS3 in cell C3 may be transmitting the third synchronization word at the same time, which will be followed in the next two time segments by the first synchronization word and the second synchronization word. Then, the base station BS4 in the cell C4 can, like the base station BSl, be transmitting the first synchronization word which will be followed in the next two time segments by the second synchronization word and the third synchronization word. This means that a mobile receiver in the Cl cell, which attempts to detect the transmission of the base station BS1, will now have very little probability of carrying out erroneous detection by receiving transmissions from the C2 or C3 cells, due to the very low correlation between the three words of synchronization. The closest cell from which it is likely to receive a synchronization signal that is erroneously detected as a synchronization signal from the base station BS1 is the base station BS4 in the C cell. As indicated in Figure 3, cell C4 is V3D of cell Cl. This distance can be considered as a minimum distance of time position refusal, by analogy with the minimum distance D of frequency rejection. Accordingly, the probability of erroneous synchronization caused by a mobile receiver detecting the synchronization word provided from the co-channel interfering device is greatly reduced. This means that, when planning the network, it is possible either to use a shorter frequency reuse distance, or to operate in worse C / I conditions.