JPH07107538A - Channel selection device, fixed radio station and radio transmission system - Google Patents

Abstract

PURPOSE: To improve traffic between a moving station and a central station by providing an adder for adding input signals by using a weighting coefficient and an evaluation device for deciding the weighting coefficient by the evaluation of reception signals. CONSTITUTION: This channel selection device 1 is provided with an input unit 8, and the input signals e1-en are supplied to the input unit by a digital form and are supplied further to an adder 2 and an evaluation device 3. The evaluation device 3 is provided with a signal processor 4 for taking out signal energy E(k), a pause detection means 5 for judging a first state 'radio communication' z(K)=F and a second state 'pause' z(K)=P, a measurement means 6 for a measuring channel quality Q(k) and a decision means 7 for deciding weighting coefficients G1-Gn. On the output side of the adder 2, reception signals (y) formed by the weighted sum of the input signals e1-en are obtained, and the reception signals (y) are transmitted through a D/A converter 9 to a loudspeaker 10.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a channel selection device for selecting one reception signal from a predetermined number of reception signals.

[0002]

2. Description of the Prior Art A channel selection device of the type mentioned at the outset can be used, for example, in a radio transmission system consisting of a plurality of receivers arranged at different locations in the reception area. Each of these receivers is provided for receiving radio signals transmitted from, for example, an aircraft. The radio signal can be received by a predetermined number of receivers and transmitted as a received signal to the central station. The receiving branches are integrated in the central station, that is to say that a certain number of received signals are converted into a single output signal. For this purpose, for example, the "best" receiving branch is selected.

US Patent Application No. 4246655
A device of the type mentioned at the outset is known from the publication.
At that time, this US patent application No. 4246655
FIG. 2 of the publication shows that a plurality of received signals are supplied to individual receivers. The output signals of these receivers are fed to their measuring means for measuring the signal-to-noise ratio. Furthermore, all output signals of these receivers are forwarded to a channel selector, which selects one receiving branch. In this case, the channel selection device is driven by the detection means for evaluating the output signal of the SN ratio measurement means.

[0004]

The object of the present invention is to improve the radio traffic between a mobile station and a central station in a device of the type mentioned at the outset.

[0005]

SUMMARY OF THE INVENTION According to the present invention, there is provided a channel selection device comprising: an adder for adding input signals using weighting factors; and an evaluation device for determining weighting factors by evaluating received signals. It is solved by having.

[0006]

The received signal consists of a radio signal transmitted as a radio call from a mobile station, for example provided in an aircraft. These radio calls are received by a plurality of receiving stations located at different places in the receiving area and transferred as a received signal to the central station via a plurality of receiving branch paths. The received signal can be transmitted as an analog signal to the central station and can be digitized in the channel selector or before being input to the channel selector. Alternatively, the received signals are already transmitted in digital form between the receiving station and the central station if they are digitally transmitted signals or if they are digitized at each receiving station. Therefore, the digital received signal can be used in any case. The received signal is first transmitted to the adder via the first signal path and then to the evaluation device via the second signal path. The evaluation device evaluates the received signal and from the result determines a weighting factor for the weighted addition by the adder. It only takes a short period, on the order of a few milliseconds, to determine the weighting factors, and thus a short response time to select the received signal. Such a short response time is extremely advantageous for wireless calls, for example in ground-to-air communications. Since the weighting factor can be a value between 0 and 1, no click sound is generated between the received signals when switching is performed. Since only the received signal itself is evaluated, signaling information signals, such as the type of signaling tones at the beginning and end of each radio call, can be omitted.

The first basis for the evaluation of the received signals is formed by the arrangement in which the evaluation device comprises a signal processor which extracts the signal energy from each received signal.

The effect of pulse-like impairments on channel selection, for example, can be reduced by retrofitting the signal processor with a median filter.

The first classification of the received signal into "sampled" signal components and signal components in which only channel "noise" is audible can be made with the following arrangement. That is, the evaluation device has a pause detecting means for evaluating the signal energy of the received signal, and the pause detecting means calculates whether the first state is "wireless call" by calculating the energy for a short period and determining the threshold value. It is provided to detect whether the state is "pause".

The erroneous classification of the "pause" state, for example when the speaker breathes, is avoided by providing a predetermined period after the end of the first state "radio call". During this period a "radio call" state is assigned, which is tied to the end of the detected "radio call" state.

The assignment of the weighting factors is substantially effected by the arrangement in which the evaluation device comprises measuring means for measuring the channel quality of the received signal. When the first state "wireless call" is detected, the measuring means measures the SN of the received signal.
It is used to estimate the channel quality in response to the ratio.

Continued use of the estimated channel quality of the preceding radio call is ensured as follows. That is, in order to determine the channel quality, when the "wireless call" state is detected, the maximum value of the preceding channel quality and the estimated SN ratio are used, and the second state "pause" is detected. In this case, the product of a predetermined coefficient and the preceding channel quality is used.

The cutting of the first syllable of a wireless call is avoided as follows. That is, the evaluation device has a deciding means for deciding the weighting coefficient, and the first received signal detected in the "wireless call" state is
It is prioritized when assigning weighting factors.

The present invention will now be described in detail with reference to the embodiments shown in the drawings.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The radio transmission system shown in FIG. 1 has a channel selection device 1, which is provided, for example, in a fixed radio station (= terrestrial base station) arranged on the ground. . The input signals e1 to en received by the receiving stations F1 to Fn are supplied to the channel selection device 1. Channel selection device 1
Further comprises an input unit 8 to this input unit, possibly in digital form, eg PCM.
The input signals e1 to en are supplied in the form. Input signal e
1 to en are supplied to the adder 2 and the evaluation device 3.
This evaluation device 3 comprises a signal processor 4 for extracting the signal energy E (k) and a first state "radio call" z (K).
= F and the second state "pause" z (K) = P, the pause detecting means 5, the measuring means 6 for measuring the channel quality Q (k), and the determining means 7 for determining the weighting factors G1 to Gn.
Have and. At the output side of the adder 2, the input signal e1
Received signal y formed by the weighted sum of
And the received signal y is transmitted to the speaker 10 via the digital / analog converter 9.

The transmission system shown in FIG. 1 is used to transmit radio signals between a pilot located at mobile station M and an operator located near speaker 10 of the terrestrial exchange. . In this case, the communication between pilot and operator should be as reliable and uninterrupted as possible. To ensure this connection, the sending / receiving station F1
~ Fn are distributed at various different locations over the transmission / reception reach.

The input signals e1 to en are based on a radio call transmitted from the mobile station M arranged in the aircraft, for example. The role of the channel selection device 1 is
Received signals e1 of all receiving stations F1 to Fn
en, e.g. to make available to the operator via the loudspeaker 10 the best channel, i.e. the best received signal e1-en in each case.

For such ground-to-air communication, the radio signals (= received signals) e1 to en received by all of the receiving stations F1 to Fn are sent to the channel selection device 1 of the commercial exchange, for example, by the channel multiplex transmission system. Transmitted in. Mobile station M and receiving stations F1 to Fn
The wireless connection between and changes over time as the aircraft moves. Moreover, the wireless connection between the mobile station M and the receiving stations F1-Fn is "scanned". That is, there is a perfect wireless connection only for a given period. During periods when there is no wireless connection, the transmission channel changes significantly over time, and therefore only noise can be detected in commercial exchanges. In the case of the transmission system shown in FIG. 1, the use of signaling tones to mark the beginning or end of a radio call is intentionally avoided. This is because this implies additional circuitry and cost. Received signal e
1 to en are transmitted in analog or digital form from the receiving stations F1 to Fn. Received signal e1
If ~ en are transmitted in analog form from the receiving stations F1 to Fn, the received signals are digitized at the input unit 8, so that the received signals e1 to en are
It can be taken out at the output of the input unit 8 in digital form, for example in PCM format. n
The individual reception signals e1 to en are not affected by the channel selection device 1, but are multiplied by the weighting factors G1 to Gn in the adder 2 for "channel combination" and added. Then, the weighted sum becomes audible through the speaker 10 as the selected reception signal y which is the output signal after the digital / analog conversion by the digital / analog converter. The main role of the channel selection device 1 is to obtain the weighting factors G1 to Gn in response to the received signals e1 to en. The evaluation device 3 is used for this purpose. In this case, the signal energy E (k) for each of the received signals e1 to en is measured in the signal processor 4 in the first step. On the basis of the measured signal energy E (k), the pause detection means 5 makes a classification into "radio call / pause" in order to detect the states z (k) = F and z (k) = P. When the first state "radio call" F occurs, the channel quality Q (k) is determined in the measuring means 6 and from this channel quality a weighting factor G for the individual received signals e1-en is obtained.
1 to Gn are derived. At that time, the weighting factors G1 to G
n varies between 0 and 1 which allows a smooth transition from the "off" state (= 0) to the "on" state (= 1) and vice versa. Therefore, switching noises such as click sounds do not occur in the transition between the individual input signals.

FIG. 2 shows an embodiment of the evaluation device 3 which can be used in the channel selection device shown in FIG. Here too, the reference numbers already used in connection with FIG. 1 are used. Next, referring to FIG. 2, the evaluation device 3
The individual steps of are described in detail. In this case, the individual channel indicating codes for distinguishing the different input signals e1 to en are not described. The reason is that the received signals e1 to en from all channels are processed in the same way and are processed independently of the other channels.
The channel indication code is not used until the weighting factors G1 to Gn are determined.

In the signal processor 4, the unit 4a first determines the signal energy E (= short-term energy value E (k)) according to the following equation:

[0021]

[Equation 1]

In this case, each signal sample value x (i) of one given channel is used only once for energy measurement. One energy value E (k) is thus obtained for every L sample values, and the signal processing on this basis is correspondingly repeated after every L sampling intervals. In practice, 4 ms was selected for the average period. This is 8 kHz
This corresponds to L = 32 sample values for a sampling rate of. This guarantees that this energy measurement can also detect very short pauses in the radio call, while at the same time this averaging period is not too short in order to have a reliable measurement.

By performing median filtering using the median filter 4b after measuring the signal energy E (k), it is possible to reduce the adverse effects of pulse-like disturbances on the selection of the received signals e1 to en. For this purpose, the last N energy values at the discrete time k are sorted according to size. In this case, the median E _m (k) = E
_{It is (n-1) / 2} .

The first state "radio call" F or z (k)
= F and the second state "pause" P, i.e. z (k) = P, a pause detection means 5 is provided. The state determination by the pause detection means 5 is substantially the short-term energy E (k) or the modified short-term energy E _m (k).
Based on a non-linear calculation of the value derived from and subsequent threshold determination. As input values for the pause detection means 5, the signal energies E (k) and _Em (k) and the last taken state z (k-1) are used. The state immediately before is supplied from the memory 5a to the first switch S1 on the input side and the second switch S2 on the output side of the pause detecting means 5. What is important is that in the "pause" state, whether the state z (k-1) taken immediately before is held or the state is shifted to another state is determined according to the determination criteria _Kp1 and _Kp2 . In the "wireless call" state, the judgment is made according to the judgment standard K _F. In other words, in the "pause" state, the determination is made according to a different criterion from that in the "wireless call" state. In resting detection means, depending on the previous state z (k-1), criteria for signal energy E to be evaluated at that time (k) and E _m (k) is obtained. In this case, the criterion for determining the second state "pause" can be synthesized from two separate criteria K _p1 , K _p2 , which are determined by the first state "radio call". If it is set to 1 and the second state "pause" is set to 0, it can be considered as an OR connection.

Based on these criteria K _p1 and K _p2 ,
The current short-term energy E (k) is compared with the previous energy value. In the criterion K _p1 , the preceding short-term energy E (k−1) is used as it is for comparison, and therefore, for the criterion K _p1 (k),

[0026]

[Equation 2]

The above is true.

On the other hand, with respect to the criterion K _p2 ,
The comparison with the average value of the preceding plurality of short-term energy values is performed as follows. That is,

[0029]

[Equation 3]

In this case, α is typically on the order of 0.1.

Subsequent threshold determination SE _KP1 , SE _KP2 , SE
The threshold for _KF can be set on a case-by-case basis and tailored to the short term energy of the channel noise. In simulations, a threshold value of the order of, for example, 10 −3 was found to be suitable. The basic technical idea regarding the selection of the criterion for the pause detecting means 5 is that a rapid energy change, especially an increase in energy can be detected reliably and promptly. In this case, the energy value E (k) for the channel noise is the second state "
It is assumed that the energy value for "rest" fluctuates significantly only slightly. The first state "radio call"
As a criterion for determining F, the average energy change is checked when in the "wireless call" state, ie z (k) = F. In this case, for example, the calculation is performed according to the following formula. That is,

[0032]

[Equation 4]

It has been found that it is suitable for selecting the judgment threshold value SE _KF that the short-term energy E (k
-1) is to select a threshold value equal to that.

The measuring means 6 of the evaluation device 3 receives the received signal e1.
Is for determining the channel quality Q (k) of ~ en. In this case, the estimation of the channel quality Q is the first state "
Only when the "wireless call" is detected, each received signal e1
It is performed according to the SN ratio SNR of en.

It has been found to be advantageous that the channel quality Q is reached when the first condition "radio call" F is detected.
(K) is obtained from the maximum value of the preceding channel quality Q (k−1) and the estimated SNR SNR. Second
If the state "pause" P is detected, the channel quality Q
(K) is a predetermined coefficient τ and the preceding channel quality Q (k-
It is calculated based on the product of 1). Therefore, the following expression holds for the definition of the channel quality Q (k). That is,

[0036]

[Equation 5]

In this case, the predetermined coefficient τ is slightly smaller than 1.0, for example τ = 0.999, and Q
(0) = 0 is selected. So in a wireless call,
The maximum value of the preceding channel quality Q (k-1) and the SNR value estimated at that time are used. By defining the channel quality Q (k) according to Equation 5, the effect of the estimated SNR value decreases as the SNR value decreases, while at the same time replacing the SNR value with the current value more quickly in the next radio call. Will be able to. Equation 5
Due to the definition of the channel quality Q (k) based on, this quality is not calculated anew at each radio call F, but as much information as possible is included in the current estimate from the previous radio call. Become so. This eliminates the drawback that the current SNR estimate is not yet very reliable at the start of a wireless call. On the other hand, the channel quality Q defined in equation 5
What is further assured by the definition of (k) is that in the case of a channel failure that was presumed to be the best during the last wireless call but failed during that time, the wireless call is lost. Is. Moreover, what can be obtained by the definition of the channel quality Q (k) defined in Expression 5 is
The first part of the stationary process has no effect on the channel quality when trying to obtain the SNR estimate, and S
If the NR estimate is sufficiently stationary, the channel quality Q
(K) is to follow the current SNR estimate. In the idle period, the channel quality Q (k) corresponds to the exponentially decreasing weighted SNR estimate of the preceding radio call.

In FIG. 2, Equation 5 is schematically shown as a block circuit diagram. Switches S3 and S4 are provided on the output side and the input side of the means 6, respectively. These switches are switchable between switch position P for z (k) = P and switch position F for z (k) = F, depending on the detected state z (k). The branch path for the switch position P has a multiplier 6b, which is
The preceding channel quality Q (k-1) is multiplied by the factor τ.
The branch path for the switch position F has a calculating means 6c which determines the SNR estimate in response to the energy value E (k) according to the instructions given below as equation 6. The calculating means 6d calculates the equation 5 for z (k) = F based on the SNR value and the preceding channel quality Q (k-1).
The current channel quality Q (k) is calculated according to

Next, how the SNR estimation value can be obtained based on the signal energy E will be described. For this purpose, the following definitions have been found to be advantageous for the SNR estimate.

[0040]

[Equation 6]

In this case, kεF means that the estimation is performed only during a radio call, that is, the estimation is performed only when z (k) = F. Of course, in order to find the extremum, the criterion defined in Eq. 6 requires a certain settling time. In practice, the first signal energy after the detection of a radio call F is used both as a maximum and as a minimum, on the basis of which all subsequent signal energies they represent a new maximum or minimum. Be inspected. Only when the maximum or minimum value no longer changes or changes only very slightly can this be called a reliable measurement.

Channel quality Q obtained based on equation (5)
It has been found to be advantageous to use a modified channel quality Q'instead of This modified channel quality is
It is the channel quality Q (k) according to equation 2 multiplied by the low pass filtered energy value. For this reason, a multiplier 6e is connected in front of the output side of the measuring means 6. The multiplier 6e solves two problems at the same time. That is, as the first, various received signals e1 to e1
A weak auditory impression resulting from different input levels of en is avoided and, second, the first syllable of the wireless call is not cut. Overall, a "first come first serve" approach is obtained, ie the first scanned channel receives the highest priority in assigning weighting factors unless another channel with better quality is detected.

Based on the results of the signal processor 4, the pause detecting means 5 and the measuring means 6 for measuring the channel quality, the determining means 7 determines the weighting factors G1 to Gn. In determining the weighting factors for the individual received signals e1 to en, the following are taken into account: Unless a radio call is received, an equal weighting G1 for all channels.
~ Gn = 1 / N is assigned. In this case, N is the number of reception channels. If different quality values are sought for different channels, the best channel weighting is

[0044]

[Equation 7]

Is increased according to In this case, η is a constant and defines an exponential rise rate. ε _s
Is an initial value of approximately zero and is valid when the preceding weighting factor was approximately zero or equal to zero. In contrast, the weighting of the channels with the smallest qualities Q (k) to Q '(k) is reduced as follows. That is,

[0046]

[Equation 8]

In this case, η is an exponentially decreasing time constant, and ε _s causes the weighting to reach zero in the last short period. Since the weighting factors G1 to Gn are determined in this way, the following requirements are sufficiently satisfied. When at least one radio path is scanned, at least one channel should be audible so that the call is not lost. Moreover, only one channel should be audible to avoid reverberation effects due to the relative delays of the various channels. The transition between channels should be free of switching noise, eg clicks, but should nevertheless be done as quickly as possible. Furthermore, all channels must be treated equally unless quality information is available on either channel.

In this case, the first requirement is already satisfied by the definition of Q '(k). The limit to one audible channel is met by the classification of the signal, as done by reducing the weighting factors for the worst channel. The smooth transition is achieved by a nearly exponential increase and decrease of the weighting factors G1 to Gn.

When the weighting factors G1 to Gn are determined,
In the adder 2, the individual reception signals e1 to en are weighted by the weighting factors G1 to Gn and added according to the following Expression 9.

[0050]

[Equation 9]

[0051]

According to the present invention, one out of a predetermined number of received signals can be obtained.
The radio traffic between the mobile station and the central station is improved in a channel selection device which selects one received signal.

[Brief description of drawings]

FIG. 1 is a diagram showing a wireless transmission system including a channel selection device.

FIG. 2 is a diagram showing an embodiment of an evaluation device.

[Explanation of symbols]

 1 channel selection device 2 adder 3 evaluation device 4 signal processor 5 pause detection means 6 channel quality measurement means 7 weighting coefficient determination means 8 input unit M mobile stations F1 to Fn reception stations

Claims (10)

[Claims] 1. From a predetermined number of received signals (e1 to en), 1
In a channel selection device (1) that selects one reception signal (e1 to en), the channel selection device (1) is configured to use weighting factors (G1 to
Gn) has an adder (2) that adds input signals (e1 to en) and an evaluation device (3) that determines weighting coefficients (G1 to Gn) by evaluating received signals (e1 to en). A channel selection device characterized by the above. 2. The evaluation device (3) is adapted to receive each received signal (e
Channel selection device according to claim 1, comprising a signal processor (4) for extracting signal energy (E) for 1 to en). 3. The channel selection device according to claim 2, wherein the signal processor (4) comprises a median filter (4b). 4. The evaluation device (3) comprises a received signal (e1).
~ En) has a pause detection means (5) for evaluating the signal energy (E) of the first state "radio call" (F) from the short-term energy calculation and threshold determination. 4. The channel selection device according to claim 1, wherein the channel selection device is provided for determining the second state “pause” (P). 5. The detected first condition "wireless call"
A predetermined period is provided after the end of (F), and the "wireless call" state (F) is assigned to the period, and the period is
The channel selection device according to claim 4, wherein the channel selection device is connected at the end of the detected first state "wireless call" (F). 6. The evaluation device (3) comprises a received signal (e1).
To en) for measuring the channel quality (Q) of the received signal (6) during the detected first state "wireless call" (F). e1-e
Device according to any one of claims 1 to 5, provided for estimating channel quality (Q) according to the signal-to-noise ratio of n). To determine the 7. A channel quality (Q _k), "radio communication" state (F) is the maximum value and the estimated SN ratio of the channel quality preceding when detected (Q _k-1) Channel selection according to claim 6, characterized in that the product of a predetermined coefficient and the preceding channel quality (Q _k-1 ) is used when the second state "pause" (P) is detected. apparatus. 8. The evaluation device (3) comprises a weighting coefficient (G1.
Gn) is determined, and when the weighting factors (G1 to Gn) are assigned, the received signals (e1 to en) in which the "wireless call" state (F) is detected are prioritized first. The channel selection device according to any one of claims 1 to 7, which is performed. 9. Fixed radio station for a mobile radio system, characterized in that it comprises a channel selection device (1) according to any one of claims 1-8. 10. The wireless transmission system according to claim 1.
10. A radio transmission system comprising the channel selection device (1) according to any one of items 1 to 9.