Method for receiving radio signals via a radio channel

State of the art

p0001The invention is based on a method for receiving radio signals via a radio channel according to the genus of the independent patent claim.

p0002It is already known that time synchronization is carried out in a receiver in the case of radio signals which are divided into frames. In particular, a voltage-controlled oscillator is used in the receiver for generating a clock frequency in the receiver, in particular in the case of DAB (digital audio broadcasting). If the clock frequency of the voltage-controlled oscillator is unequal to the clock frequency used in the transmitter, then an error occurs in the time synchronization by the fact that the actual frame start is removed further and further from the receiver assumed frame start due to the frequency error. This is readjusted by adjusting the frequency of the voltage-controlled oscillator by changing the voltage of this oscillator.

p0003In <patcit id="pcit0001" dnum="EP872304A2"><text>EP-A2 872 304</text></patcit> A synchronization is disclosed in which a time difference between the receiver and the transmitter is determined. This is based on a difference between a reception time and a detection time. Furthermore, an average value of the deviation is formed. A correction signal is formed as a function of the deviations.

p0004Advantages of the invention

p0005In the case of a reception interruption, for example during passage through a tunnel, the time synchronization is not lost, so that a user of the radio receiver does not experience quality losses after a reception interruption.

p0006A measure from the past error values ​​is calculated to determine the correction of the time synchronization during the reception interruption. Thus the past deviations during the time synchronization are used to predict the deviation in the case of a reception interruption.

p0007It is advantageous that very large error values, which are calculated by the TFPR symbol, are discarded. In particular, error values ​​caused by a short-term transmitter change in, for example, a common-mode network, as is the case with DAB, are thereby eliminated. As a result, greatly deviating values ​​for the average value formation are excluded.

p0008Advantageous further developments and improvements of the method for receiving radio signals via a radio channel are described in the dependent claims.

p0009It is particularly advantageous that with the method according to the invention a fixed frequency oscillator can be used in the receiver which is extremely cheap and greatly simplifies the design of the radio receiver.

p0010Moreover, it is advantageous that upon receipt of DAB, the TFPR (Time Frequency Phase Reference) symbol is used to establish the time synchronization in the receiver. For this purpose, an impulse response of the radio channel is advantageously calculated by means of the TFPR symbol in order to calculate the deviation of the time synchronization, which is caused by the propagation time.

p0011In an advantageous further development of the invention, the average value of these error values ​​is calculated as the measure which is calculated for the past error values. Thus, short-term deviations of the error values ​​will have no appreciable influence on the correction of the time synchronization in the event of a reception loss.

p0012Furthermore, it is advantageous that by comparing successive error values ​​from the receiver it is recognized whether a trend occurs in the development of the error values ​​so that a prediction of the future deviations due to this trend becomes possible. This allows an adaptive control of the time synchronization in the event of a reception failure.

p0013In addition, it is advantageous that the occurrence of a transmitter change in a common-mode network is detected. In this way, advantageously only error values, which have been calculated for a transmitter used for the reception, are used for the average formation.

drawing

p0014Exemplary embodiments of the invention are illustrated in the drawing and explained in more detail in the following description. FIG. 1 shows a method for time synchronization, FIG. 2 shows a DAB frame, FIG. 3 shows two pulse responses, and FIG. 4 shows a method for adaptive time synchronization in the event of a reception failure.

DESCRIPTION OF THE EXEMPLARY EMBODIMENTS

p0015In the time synchronization of radio signals which are transmitted in frames, as is the case, for example, with digital broadcasting methods such as DAB, the receiver is synchronized to the beginning of the frame. This synchronization to the beginning of the frame is necessary so that the information transmitted at the beginning of the frame can be read by the receiver and the frame can be decoded at all. In particular, at DAB, the information at the start of the data includes data on the remaining data that is transmitted within this frame. If the receiver has a different frequency with which the receiver processes the frames compared to the transmitter, then the synchronization will move away from the start of the frame and the recipient will then be unable to process the information in the frame correctly. Therefore, a regulation of the time synchronization is necessary, in particular if the clock frequency can not be adjusted in the receiver.

p0016If the frequency is not adjusted in the receiver, a pointer to the data stream where the frame starts must be set in the receiver. The pointer is a date that is set by the software that controls the receiver. The pointer points to a memory address, in this case the memory address at which the beginning of the frame is stored.

p0017This pointer is set, for example, by a processor in a receiver. If an error occurs during the synchronization, this pointer no longer points to the beginning of the frame to be processed at the moment. Then this pointer must be moved according to an error value, so that the pointer points back to the beginning of the frame. The error value thus means here the time error in time synchronization.

p0018DAB is a high-bandwidth multi- frequency carrier transmission system. Within this digital terrestrial hearing radio system, the modulation method of orthogonal frequency division multiplex (OFDM) is used. OFDM is a modulation method in which the signal to be transmitted is distributed over a plurality of subcarriers, the signals distributed to these subcarriers not interfering with each other. This behavior is described as orthogonal.

p0019In transmission mode II at DAB, a frame with a duration of 24 ms is available. The sampling frequency is 2.048 MHz, with samples for the frame being 49152 samples. If there is a frequency difference between the transmitter and receiver of 0.005% at the sampling frequency, a clock drift of approximately 2.46 sampling periods per 24 ms frame is produced. The receiver must therefore shift the pointer to the beginning of the frame by 2.46 samples for a correct temporal synchronization with every 24 ms frame. In the case of a longer interruption of the reception, for example due to tunnel passages, a correct time synchronization is not possible at the same time. The method according to the invention avoids a new synchronization which is then required.

p0020In addition to DAB, other broadcasting methods such as DVB (digital video broadcasting) and DRM (digital radio Mondial) also have a frame structure, these broadcasting methods also having the modulation method OFDM. These methods differ from DAB on the one hand by the structure of the frame structure and on the other by the transmission frequency ranges and transmission rate. In the case of DVB and also DRM, test signals are used for determining the pulse response and thus for time synchronization.

p0021An oscillator, that is to say a circuit which generates oscillations which only emits oscillations with a frequency, is referred to as a fixed-frequency oscillator. The voltage-controlled oscillator, on the other hand, can be forced by the change of an applied voltage to give a vibration with a changed frequency. The fixed frequency oscillator is cheaper and easier to manufacture due to its simpler structure.

p0022FIG. 1 shows a method for time synchronization. This method runs here in a DAB receiver, this method being controlled by software running on a processor of the DAB receiver.

p0023In process step 1, the process is started. In method step 2, a check is made as to whether a reception of DAB broadcasting signals is currently in progress. This is determined by the fact that the signal power is observed. If signal power is detected above a predetermined threshold, a DAB broadcast signal is present, otherwise not.

p0024If a reception of DAB broadcasting signals is present, a comparison of the phase reference symbol, which is present in the received frame, is carried out in method step 3 with a phase reference symbol which is stored in the receiver. This comparison is performed to determine an impulse response of a radio channel.

p0025FIG. 2 shows a DAB frame for the DAB mode 1. The DAB frame has a width of 96 ms. Within the DAB frame, the synchronization channel 11 is first transmitted by the zero symbol 9 and the phase reference symbol 10. Then, information is transmitted in the fast-information channel 12 where the data allocated to the programs and data are to be found in the main service channel 13 and how the respective channel encoding looks. The main service channel 13 is transmitted with the remaining 72 OFDM symbols. The main service channel 13 contains the actual useful data, such as audio signals or additional data.

p0026By means of the phase reference symbol from the DAB frame, the pulse response of the radio channel is determined. The phase reference symbol is transmitted in each frame, that is, every 96 ms. This clock is hereinafter referred to as a frame clock. The pulse response is calculated so that the frequency-dependent transmission function H (f) is first to be determined. H (f) is determined by multiplying the received signal by the complex-conjugated send signal. The original phase reference symbol is used and conjugated as a complex conjugated send signal. By backtransformation into the time domain, the impulse response h (t) is obtained from the transfer function H (f). Prior to the determination of the transfer function H (f), the phase reference symbol was transformed by transformation from the time domain to the frequency domain.

p0027FIG. 3 shows two pulse responses of the radio channel. The peak value of the momentarily calculated pulse response 21 of the radio channel is compared with a further pulse response 20 as a reference, the deviation 14 from this reference giving the error value for the time synchronization.

p0028This reference is due to the occurrence of a peak value in the impulse response in the middle of the time values ​​used for the calculation of the impulse response. This is because, by means of a fast Fourier transformation (FFT), the conversion of a time signal into the frequency domain is carried out. A time window is used here for the FFT, which is applied over the values ​​which are transformed into the frequency domain by the time domain, thus the values ​​for the determination of the transfer function and consequently the impulse response are determined. The transfer function from the frequency domain is converted into the time domain by means of an inverse fast Fourier transform.

p0029If the receiver is correctly synchronized, then the time window for the FFT is located directly above the phase reference symbol, whereby the peak value of the pulse response will then occur in the middle of the time window. If the receiver is not correctly synchronized, then the peak value of the impulse response is also shifted correspondingly, whereby the error value for the time synchronization is calculated.

p0030If, moreover, several peak values ​​are present in the impulse response, it is necessary to have all these peak values ​​within the time window of the FFT and, for example, to calculate the time of greatest energy in the impulse response and to determine the deviation from the time center of the impulse response to the error value In time synchronization. One method to determine the placement of the FFT window is the strategy of maximizing the distance of the peaks from the edges of the time window.

p0031In method step 4, an average value of the error values ​​for 100 frames is calculated. For this purpose, the error values ​​of the last 100 frames are added and divided by 100. Furthermore, a larger or smaller number of frames can be used for the average value formation, as far as this seems necessary or is made possible by, for example, memory location resources in the receiver.

p0032In process step 5, the average value is stored. In method step 6, the time synchronization in the DAB receiver is corrected by the error value determined for the current frame. For this, the pointer which points to the frame start assumed incorrectly by the error value is now set to the correct frame start. In method step 7, the data contained in the current frame are processed in the receiver. With DAB, a decoding is performed here.

p0033If there is no reception of DAB broadcasting signals, for example in the case of tunnel runs, the pointer pointing to the beginning of the frame is corrected in method step 8 and thus the time synchronization is corrected by the average value which was last stored. In this case, the average value may be positive or negative, since mis-synchronization can be present in both directions since the frequency through the oscillator in the receiver can be either too large or too small compared to the frequency in the transmitter.

p0034At DAB, a DC wave network is available. The transmitters are therefore operated at a certain frequency. It can therefore happen that a very large error value occurs during the time synchronization in the case of a transmitter change, that is, when the signal of a transmitter is no longer receivable and only the signal of another transmitter can be received. If this change is only short-term, then the receiver discards this error value so as not to distort the average value of the error values. The receiver does not discard the error value, however, because a permanent change of the sender can occur. Therefore, the receiver determines further error values ​​until a predetermined number of error values ​​are reached. The receiver then determines the average value and checks whether the error values ​​deviate strongly from this average value. For this purpose, a predetermined threshold value is stored in the receiver. The error value is compared with this threshold value, if the error value lies above this, then the error value is discarded.

p0035If, however, the transmitter is changed continuously, then the receiver will reject the error values ​​which have arisen for the first transmitter, which is no longer used for reception, in order to form an average value of the error values ​​only for the transmitter now being used. For this purpose, the receiver forms two average values, one for the error values ​​before the error value, which deviates strongly from the average value, and one for the error values ​​which follow the error value, which greatly deviates. If the difference between the two average values ​​is below a predefined threshold value, then there is no permanent transmitter change, but if this is the case, a permanent transmitter change is present.

p0036Moreover, it is conceivable at DAB not to operate a common-mode network and to supply a region with a transmitter.

p0037FIG. 4 shows a method which is used for the adaptive maintenance of the time synchronization in the event of a reception failure.

p0038In method step 15, the error values ​​are calculated as described above. In method step 16, the current error value is compared with at least two past calculated error values. In method step 17, it is checked whether the difference between these error values ​​is based on a monotonically increasing or monotonically decreasing function, so that the statement can be made as to whether a trend is apparent from the difference between the error values, which is based, for example, on temperature effects. If such a trend is detected, the future error values ​​are estimated by an extrapolation of the function in method step 19 in order to correct the time synchronization by these error values ​​in the event of a reception failure.

p0039Extrapolation means here that the future error values ​​are calculated, for example, in such a way that the last calculated error value, still present as reception of broadcasting signals, for the first error value, if there is already no reception of broadcast signals, is changed by the value Which the penultimate error value had changed to the last error value for broadcast reception. So here is a simple linear extrapolation.

p0040If no trend can be read from the comparison of the error values, the time synchronization is corrected by the error value in the case of reception of broadcasting signals, or in the case of a reception interruption by the average value of the last error values.