EP3332586A1

EP3332586A1 - Method for controlling the transmission power

Info

Publication number: EP3332586A1
Application number: EP16751541.0A
Authority: EP
Inventors: Francisco Lázaro Blasco
Original assignee: Deutsches Zentrum fuer Luft und Raumfahrt eV
Current assignee: Deutsches Zentrum fuer Luft und Raumfahrt eV
Priority date: 2015-08-07
Filing date: 2016-08-03
Publication date: 2018-06-13
Also published as: WO2017025413A1; US20180234926A1; FR3039945A1; US10225805B2; FR3039945B1; DE102015215177B3

Abstract

The invention relates to a method for determining the power control parameters in order to control the power, in particular the energy per symbol, with which signals are transmitted from transmitters of a group of multiple transmitters to a receiver assigned to said group in packets according to a multiplex specification in a communication system. Each transmitter of a group determines the transmitter transmission power randomly such that the signal-to-noise ratio of the receiver lies between a minimum and a maximum value. The minimum and maximum signal-to-noise ratio values that determine to which group a transmitter belongs are defined in a first table, and the probability value used for the transmitter is defined in a second table in order to determine the transmission power of the transmitter such that the transmitter signal-to-noise ratio lies within the total range of permissible signal-to-noise ratios or within the signal-to-noise ratio that is permissible for the group to which the transmitter belongs. New values for the first and second tables are defined such that the maximum load in packets per seconds which can be accepted by the communication channel while maintaining a desired threshold packet error rate is maximized and/or the distance from the histogram of received signal-to-noise ratios in dB to a uniform histogram of signal-to-noise ratios in dB is minimized.

Description

Method for controlling the transmission power

The invention relates to a method for controlling the transmission power (energy per symbol), with which in a communication system signals from transmitters of a group of multiple transmitters to a receiver assigned to this group of packets in accordance with a multiplex specification, in particular random access specification, transmitted.

With this invention, a new technique is proposed with which power control can be performed when using a random access technique to access a common transmission medium.

Random access is a technique for transmitting information over a transmission medium in which multiple terminals share the transmission medium. In random access techniques, there is no central control unit controlling access to the transmission medium.

An example of a random access technique is the Aloha protocol, in which each subscriber sends his data packets at all times and asynchronously. If more than one participant simultaneously transmits, the data packets collide and may be lost.

There are other random access techniques that achieve higher throughput. One technique that has received much attention is the one with Spread Spectrum Aloha (SSA) with Successive Interference Cancellation (SIC) at the receiver. When this technique is used similar to Aloha, the terminals send their packets at any time and in an asynchronous manner using spread spectrum techniques. Essentially, a collision of the packets occurs at the receiver. Due to the However, if spreader techniques are used, packets can also be decoded if they suffer collisions. A standard SSA Em pfänger attempt to decode each packet to a time in which he treats the rest of the packages as I nterfe ^¬ ence. In a more sophisticated receiver, the SIC is used at the receiver. W ith other words, stores the received Em pfänger Wel lenform during a time window w ith a length T. Within this window begins Em pfänger w ith the decoding of the pact with the highest ^E * / ^(N o + ^I). If the packet is successfully decoded, the receiver reconstructs the waveform from that packet and deletes it in its window. This will cancel the interference generated for all other packets. The receiver searches among those packages that are still present in his window, the package with the highest energy, decodes it and raises the interference, and the process is repeated until no white ^¬ more advanced packages are no longer present.

If SSA w ith SIC is used plays performance w ith covering different ^¬ nen packages em be pfangen, a central Rol le in the decoding process. In fact, the SIC can if all packets are m pfangen it the same power em, the load with respect to the standard decoding not raised stabili ^¬ hen. However, if the performance with which the packets are picked up follows the correct distribution, the SIC at the receiver can significantly increase the throughput of the SSA.

The meaning of the power distribution of the packets is illustrated below by means of a simple example. Consider a slot-based SSA scheme in which a fixed number of subscribers M in each time slot sends a packet. For all packages of the same over ^¬ transfer mode is used (Physical- and link layer transmission parameters: modulation, coding, etc.). Let ^ ⁽ l be defined as the m necessary E / ( ^N o + l) _f that must have a burst to decode properly (Es = P / fs, where fs is the used rate symbol). P (i) denotes the power with which the i-th packet was received For the sake of simplicity, the subscribers are subscribed in descending order according to their power P such that P (1)> P (2)> ... P (M-1)> P (M). The effective E, / (N ₀ + I) for the packet "i" in decoding is defined as:

N _Q + / (>;) where

is the interference of all subscribers for whom no decoding has yet been carried out.

Assume that whenever ^ ( ^? ) ^:> ^ ^{V (} '<1, the packet "i" is decoded with a probability of 1. If <Ir qj _S , the packet goes "i" and all subsequent ones Packets, ie the packets i, i + l ... M lost.

To illustrate the effect of power distribution on the performance of the system, two examples are given below. Fig. 1 shows in the upper part the course of Ύ over E / N (j _n <-] B) fQ _r <-] j _e different participants. The black horizontal boundary line is ^re ^ ^(l = -2.7 dB, the graph on ES / Q represents In this case, since ^• 7 (2)> req gj Vi |.. ^ All packets are decoded The lower part Fig. 1 shows the histogram of the ratio E / ^Q (in dB) Fig. 2 shows the same diagram for an overloaded system in which some packets are lost Specifically, all subscribers with Es / Nu less than 17 dB can The number of participants is the same in the two examples of Figures 1 and 2, with the only difference being in the power distribution of the packets. A good power distribution leads to a curve of 1 over E _s / No (j _n ^ ß), where 7 (> Ireq i _{. In the} ideal case, 7 (0 = Ireq Vi _# is true if a uniform distribution is used (in dB).

The following describes a power control mechanism disclosed in [3], which makes it possible to control the power distribution of the received packets in a wide variety of situations.

It is known in the art that the power distribution of the packets at the receiver is of central importance for SSA with SIC. In [1] the authors point out that if the performance of the packets follows a lognormal distribution, the throughput can be increased if the standard deviation of the lognormal deviation increases. For S-MIM [2], an ETSI standard, SSA with SIC is used at the receiver for mobile satellite communications. To improve the power distribution of the incoming packets, a uniform distributed power backoff in dB can be applied to the terminals. A terminal sets its transmission power on:

P = L + N _SAT + K + R _{r (md} (IBni)

L is the estimated attenuation suffered by the terminal in the backward link

• ^ SAT is the noise and interference power level at the receiver.

These parameters are calculated by the receiver and sent to all terminals via the forward link.

• K is defined as C / (N0 + I0) | T - GS, where C / (N0 + I0) | T is the target value for the desired ratio of C / (N0 + 10) at the satellite transponder input and GS Satellite antenna gain is in the edge area of the radiation coverage of the antenna on the ground, and • Rrand is a random value that is uniformly distributed between 0 and -mct. The parameter R-rnax is also calculated by the receiver and transmitted via the forward link to all terminals.

In the case where the required output P exceeds the endpoint capacity over ^¬, the terminal is not transmitting. S-MI M is a mobile satellite communication system, which is why after some time the terminal moves, the distance can be attenuated and the terminal can send.

In the following, several examples are given where the SSA with SIC is used at the receiver in satellite communication operations, whereby the technique used according to [2] can be improved.

In the first example assume a satellite beam in the backward link where a number q of subscribers is damped by rain. Fig. 3 shows the histogram of (in d B) for all packages using the power control technique described in [2]. From Fig. 3 it is obvious that the distribution of Es / N {) (j _nm jt, -] _er , -jj _e p _a ≤ ete is by far not a uniform distribution, there is a higher concentration of packets with low values of ^ / ^ιΛ, Ί) (in d B). The reason for this is that it is impossible for the proportion q of subscribers experiencing attenuation by rain to have high values of (in dB).

• The second example assumes a satellite beam in the reverse link, where not all of the users use the same transmission mode (physical and link layer configuration). It is angenom men that two transmission modes are used, which use the same modulation and coding scheme nd be ^¬ and occupying the same transmission bandwidth. However, the USAGE ^¬ finished spreading factor is different for the two modes. In mode 0, a spreading factor 256 is used while in mode 1 a spreading factor 64 is used. It is further assumed that all terminals have the same maximum transmission power. When using the power control according to [2], one obtains the one shown in FIG. 4 shows a histogram of E ^ / ^NQ (in dB). Again, the distribution of E _S / N ⁽ ) (j _n , -jβ) is not uniform. Again, there is a higher concentration of low value packets (j _n, j) T _he reason for this is in the fact that only participants can achieve the higher range of E _b / No (j _n, j) from the Mode 0 out.

The technique described in [2] is suitable for such mobile satellite communication operations where fade events are caused by the way to the satellite, e.g. B. is blocked by a building. As the participants move, fade events are generally short-lived. In other words, after a short time, there is a likelihood that the terminal will again be in an area with good propagation conditions. However, the technique described in [2] is not suitable for fixed satellite communications because the fade event can be long. Fade events are usually caused by rain, which can last for minutes or even hours. If the technique described in [2] is used, subscribers with "poor" propagation conditions may have to wait a long time before being able to transmit. On the other hand, their propagation conditions would be (still) good enough to achieve error-free transmission. In the following, a method for controlling the transmission power (energy per symbol) according to [3] is described with which signals from transmitters of a group of several transmitters to a receiver assigned to this group are transmitted packet by packet in accordance with a multiplex specification, in particular random access specification, wherein in the known method

the transmission powers with which the transmitters of the group transmit are within a total power range, and assigning to each transmitter a parameter indicating the transmission power with which the transmitter concerned is transmitting,

wherein the parameter is calculated on the basis of a random number and on the basis of probability values for the transmitters of the group within different specifiable power segments lying within the overall line area.

In this known method is

a first table is provided in which different transmit power segments of the total transmit power range each defined by a lower limit and an upper limit are provided, a second table containing for each transmit power segment a statistical probability value indicating how many transmitters are included within the one transmit power transmission segment,

wherein each probability value of the second table is assigned to a different transmission power segment, whereby the two tables define the expectation of how many transmitters of the group transmit signals with a transmission power lying within the respective transmission power segment,

a random number is provided for each transmitter (either by the transmitter itself or from outside), calculated from the probability values to be assigned to the respective transmitters as their associated parameters, and to each transmitter being assigned the transmit power segment within which the transmit power with which that transmitter is transmitting , and

the size of the transmission power with which the transmitter in question, within the transmission power segment associated with this transmitter, is selected by another calculation based on a random number.

According to an advantageous embodiment of the known method, it is provided that each transmitter has a maximum transmission power, that each Transmitter is assigned to the transmission power segment within which its maximum transmission power lies, that for each transmitter a random number calculated considering a uniform distribution to a predeterminable number space is provided (either from the transmitter itself or from outside), with the aid of the probability value for the latter determines whether the transmission power of the transmitter is between the lower limit of the relevant transmission power segment and a maximum transmission power or between the lower limit of the total transmission power range and its maximum transmission power.

Furthermore, it may be expedient in the known method that each transmitter has a maximum transmission power, that each transmitter is assigned to the transmission power segment within which its maximum transmission power lies, that a random number calculated taking into account a uniform distribution to a predeterminable number space is provided for each transmitter, which may be statistically equally distributed within the number space and in particular between zero and one, and that the transmitter concerned

if its random number is less than the probability value for the transmission power segment to which the transmitter is assigned transmits at a transmission power lying between the lower limit of the transmission power segment and the maximum transmission power of the transmitter, and

if its random number is less than the probability value for the transmission power segment to which the transmitter is assigned transmits at a transmission power that is between the lower limit of the total transmission power range and the maximum transmission power of the transmitter.

Advantageously, in the known method, the transmitters can transmit in different modes, in particular with different transmission rates (bit rates), wherein per mode, the transmission powers of all transmitters are controlled as described above. According to an advantageous embodiment of the known method, it is provided that before transmission of a signal from the transmitter to the receiver and / or at regular or irregular intervals, the control of the transmission power of the transmitter as described above, takes place.

In a preferred development of the known method, it is provided that a transmitter of the group, in the event that it is assigned a transmission power lying within a transmission power segment, which exceeds its allowable maximum transmission power, transmits with a transmission power between a predetermined minimum value and its maximum transmission power, namely viewed on a logarithmic scale substantially evenly distributed when the maximum transmission power is less than the lower limit of the relevant transmission power segment.

This known power control scheme is for the reverse link of a communication system, where several terminals communicate with a communication node by means of a random access scheme. There are no defaults for the random access scheme; this might or might not be timed, use a spread or not, and it might or may not use replicas. The communication node may or may not use a use interference cancellation or any other type of multi-party detection.

The terminals may use various physical and link layer (communication) configurations to transmit their data.

Assume that there is a forward link through which the communication node can send broadcast signals to the terminals.

The communication node sends two signaling tables, namely Table 1 and Table 2 to the terminals. In Table 1, upper and lower E / I ^ o values are given in dB for respective transmit power segments. The probability values in Table 2 assume values between 0 and 1 and define the probabilities of how many transmitters transmit with transmit power within the respective segments.

The terminals can the ratio of E _b / 1 \ ^Q _au f cj _he appreciated receiving end as a function of their transmission power. The terminals can calculate this estimate using an open-loop or closed-loop mechanism.

The terminals then use these tables in the following manner to calculate their transmission power.

If a terminal wants to transmit with a transmission mode "i", it will use the ith row in Table 1 and Table 2.

• The terminal estimates the maximum ratio _ι dj _e it can reach at the communication node by using its maximum transmission power. With B is the maximum ratio jn dB, which can reach the terminal. The terminal generates a pseudorandom number t uniformly distributed between 0 and 1.

The terminal determines what the largest n, n_max is for the B> If n_max is equal to u, the terminal sets n_max to u-1.

If t <Pi, _n _max, the terminal A sets _min = i, n_max and A _max = l. Otherwise, the terminal sets A _min / No _{u nc} j

The terminal then calculates "its" Es / JSlu ratio at the receiver as follows:

case

ity (the terminal does not send).

Otherwise

If B <Amm

uniformly distributed between S / A ^Q ii and B

Otherwise

If B <A _max

-E ^' .s / NO uniformly distributed between A _min and B

Otherwise

uniformly distributed between A _min and A _max

The End

As mentioned above, all E _s / ^ o values are given in dB.

It should be noted that the number of modes can be arbitrary (it can be one or more). The number u of columns in the table can be fixed or variable. The communication node may decide to dynamically increase or reduce the number u of columns. It should be noted that even if u = 2 and m = 1, the proposed scheme is not identical to [2]. In the proposed technique, terminals send each time they estimate that they can satisfy E _s / ^ o> - £ noi. The known method is characterized by the following properties:

• The communication node sends a table with u Values, which defines (u-1) - ^ - ^ o segments for each transmit mode.

• The communication node sends a table with (u-1) probability values of the use of each of the ^ No segments.

• According to the table with the probability values, the terminals arbitrarily choose a ^^ by uniformly choosing their ^ no.

• Send terminals whenever they can achieve a ^ No greater than E «f ^N.

example 1

Consider the reverse link of a satellite communication system and here one of the rays of the backward link. The terminals send via SSA and the receiver uses SIC. All terminals use the same transmission mode as e.g. :

• BPSK modulation

• Code rate = 1/3

• Spreading factor 256.

• 2500 participants

The link of a device in the middle of the beam is 29 dB.

The link budget loss due to the position of a terminal in the beam is ^{^} h and follows a uniform distribution (-6.0 dB). The terminal estimate of ^ ^' h, Li, has a Gaussian distribution in dB with a mean ^ f, and a standard deviation of 0.5 dB. A share of 25% of the terminals undergoes rain damping. The rain attenuation L _r has a Gaussian distribution in dB with a mean value of -10 dB and a standard deviation of 1 dB. The terminal estimate of the

Rain damping L, has a Gaussian distribution in dB and has a mean Lr and a standard deviation of 1 dB.

Fig. 5 is a graph of Fig. 1 above ^ ^ o for the power control (broken line) described in [2] and the power control proposed by the known method (solid line) using the following signaling tables:

Fashion Es / No_i Es / No_2 Es / No_3 Es / No_

1 -0.8 1, 2 22.2 31.2

Mode p_i p_2 p_3

1 0.0 0.13 0.1

It can be seen, as in the power control described in [2], for a range of values of the value 1 under ^{! rr (I} fall. If appropriation of the known method, however, is true />> ^{re (l} throughout the range of £ * / No, except for very low levels of Under the same conditions, the power control technique known here allows an increase in throughput of 10% compared to the technique described in [2].

For a better explanation of the known method, some details of the receiver processing operation are given. Among the e.g. 2500 devices per power segment that send:

• Average (1 -0.0) x 2500 = 2500 terminals choose the Es / No segment from -0.8 to 14.2,

• average (1 - 0.13) x 2500 = 2175 terminals choose the Es / No segment from 0.8 to 22.2, • choose on average (1 - 0.10) x 2500 = 2250 terminals the Es / No segment from 0.8 to 31.2.

Assume that a terminal estimates its maximum Es / No, B to be 33 dB. This terminal would generate a uniformly distributed random number t between 0 and 1.

If t <0.1, the terminal selects its transmission power such that Es / No is evenly distributed at the communication node between Es / No_3 = 22.2 dB and Es / No_3 = 31.2 dB.

If t> 0.1, the terminal selects its transmission power such that Es / No is evenly distributed at the communication node between Es / No_l = -0.8 dB and Es / No_3 = 31.2 dB.

Assume that a second terminal judges its maximum Es / No, B to be 25 dB. This terminal would generate a uniformly distributed random number t between 0 and 1.

If t <0.1, the terminal selects its transmission power such that Es / No is uniformly distributed at the communication node between Es / No_3 = 22.2 dB and B = 25 dB.

If t> 0.1, the terminal selects its transmission power such that Es / No is evenly distributed at the communication node between Es / No_l = -0.8 dB and B = 25 dB.

Assume that a third terminal judges its maximum Es / No, B to be -3 dB. This terminal would not transmit at all because its maximum estimated Es / No is less than Es / No.

For example, to generate a transmission power such that Es / No is uniformly distributed at the communication node between ZI dB and Z2 dB (ZKZ2), the terminal may generate a uniformly distributed random number h between 0 and 1. Es / No is then calculated as Es / No = ZI + (Z2-Z1) x h. As another example, let the terminal estimate the ratio Es / No at the communication node.

The terminal sends a message with a transmission power Pa in dB over the random access channel.

• The communication node responds to this message by specifying the Es / No = "Es / No a" in dB at which the packet was received.

• The terminal receives this message and "knows" that the power Pa in dB produces the Es / No a in dB. Now, the terminal can calculate the Es / No generated by a transmission power. For example, using the transmission power Pa-3 in dB, an Es / No = Es / No a-3 dB is calculated at the communication node.

With this method, the terminal can also calculate B, the maximum Es / No given at the communication node.

Example 2

In this second example, consider a similar case as in Example 1, where there are no rain-attenuated terminals and there are two modes of transmission that use the same modulation and coding and occupy the same bandwidth, but mode 1 has a spreading factor of 256 and mode 2 has a spreading factor 64. Assuming a system with 1200 participants with the mode 1 and 300 participants with the mode 2.

Fig. 6 is a diagram of Fig. 1 for Fig. 1 for the power control (broken line) described in [2] and the power control proposed by the known method (solid line) using the following signaling tables:

Mode Es / No_i Es / No_2 Es / No_3 Es / No_4

1 -o, 8, 2.2, 17.2, 31.2 2 -0.8 2.2 17.2 31.2

Mode P-i P-2 P-3

1 0.05 0.6 0.25

2 0 0 0

It can be seen how in the described in [2] Power control for a range of values of E ^ / N ^Q f _ur, -JJ _e falls two transmission modes of the value ^as' reg. When using the known method, however, ^'E Q NQ _{u nc} j fQ _r <applies throughout the range of E, / -] j _e two Übertagungsmodi. Under the same conditions, therefore, the known power control technique allows an increase in throughput of 10% compared to the technique described in [2].

As apparent from the above, with the method of [3], the transmission power control can be remarkably improved over the standard of [2]. For this purpose, the parameters which must be communicated to the transmitters are defined in [3].

Considerations are made according to the invention as to how particularly suitable values for these parameters can be calculated / defined.

The object of the invention is thus to further improve the known method according to [3].

To solve this problem, a method according to claim 1 is proposed by the invention. Advantageous developments of the invention are the subject of the respective subclaims.

In [3] are defined a number of parameters that must be communicated to the transmitters, and the way in which the transmitters their transmission power must calculate using the parameters. If the parameters "clean" are defined, the 1 vs curve can be controlled.

In [3] it is not explained how suitable parameters can be obtained. In a real system (eg, a satellite node), the receiver must somehow compute the power control parameters and communicate them to the terminals.

From [3] it was realized that in order to get good uplink performance from many terminals to a satellite, care should be taken that each should send a certain number of terminals in different power segments (eg 60% in power segment A, 11%). in power segment B, etc.). With the invention, the adaptation / definition of the boundaries of the individual power segments with which the intended total power range in which the terminals are transmitting is covered, and the definition of the respective number of terminals which should transmit in the respective power segment. By "terminals" are thus meant the transmitters of a group consisting of several transmitters which transmit their signals to one of these group of transmitters associated receivers (for example, satellites) in packets in accordance with a multiplex specification, in particular a random access specification.

In this invention, a method is described which enables the calculation of the values of the power control parameters defined in [3]. This allows optimizing the power distribution of the received packets in a wide variety of situations.

The transmission / reception scenario for implementing the invention is briefly summarized as follows.

A satellite assigned as a receiver to a group of transmitters sends one and the same signal to the transmitters of its group. The transmitters then return information to the satellite, namely packet according to a multiplex specification method. To ensure that all transmitters successfully send their information packets to the satellite, it is necessary to distribute the stations as evenly as possible over the ratio Es / No.

The invention will be described in more detail with reference to the drawing. In this case, Figs. 1 to 6 on the prior art according to [3], wherein

Fig. 1 shows γ over Es / No and the histogram Es / No for a system in which all transmitters are decoded,

Fig. 2 shows γ over Es / No and the histogram Es / No for a system in which some packets are lost,

Fig. 3 shows the histogram for Es / No for a setting in which some stations

Experienced rain damping, shows

Fig. 4 shows a histogram of Es / No for a multi-mode setting;

Fig. 5 shows γ over Es / No for Example 1, with the broken line γ at

Application of the method according to [2] and the solid line γ represents when the method according to [3] is used, and

Fig. 6 shows γ over Es / No for a two-mode system, with the upper part of this figure corresponding to Mode 1 and the lower part to Mode 2, while the other figures show:

7 shows a block diagram for the general description of the invention, FIG. 8 shows a first method 1 according to the invention, FIG. FIG. 9 shows a second method 2 according to the invention implemented as a steep drop gradient method. FIG.

10 is a third method 3 of the invention implemented to maximize the maximum achievable load as a steep slope gradient method;

FIG. 11 shows the normalized histogram of the ratio Es / No with old and new power control parameters and FIG

Figure 12 γ vs. Es / No with old and new power control parameters for a load of 150 packets per timeslot.

With this invention, a method is proposed which allows the calculation of "good" values for the power control parameters defined in [3]. This method can be used in the reverse link of a communication system in which several transmitters communicate by means of a random access scheme with a communication node (also called hubs). There are no defaults for the random access scheme; this could or may not be slots, or using a spread, and it might or may not use replicas. The communication node may or may not use interference cancellation or any other type of multi-party detection.

The transmitters can use various physical and link-layer configurations (transmit modes) to transmit their data.

Assume that there is a forward link through which the communication node can send broadcast signals to the terminals. The signals sent to the terminals are defined in [3] and shown in Tables 1 and 2 already mentioned. As described above in connection with [3], each transmitter must first define in which segment it is located. The station estimates its maximum achievable A, and assigns itself to the segment "I" if, and only if + l. In this case, it can be stated that the transmitter is listening to the segment "I". Subsequently, a pseudorandom number t is generated which is uniformly distributed between 0 and 1. If t> p ₀ , i, the transmitter randomizes its transmission power to be at the receiver (in dB) uniformly between

and A is distributed. Otherwise, the transmitter randomizes its transmission power to be that at the receiver (in dB) uniformly between and A is distributed.

All histograms used in connection with this invention are listed in dB. This means that the x-axis of the histogram has figures in dB.

The method described with this invention is based on the below explained assumption that the communication node knows the parameters in Table 1 and Table 2.

The block diagram of the process flow is shown in FIG. The procedure requires three inputs:

the power control parameters in Tables 1 and 2;

a histogram of the received bursts;

a model that exists in the presence of the Histogram an assessment of the Histogram at the receiver. With a _m receiver is meant when stations use their maximum transmission power.

Optionally, one can also supply as input a previously available estimate of the ^E s / ^N o _m ax histogram. This is z. For example, if the invention is used to track the ^E s / ^N o _m ax histogram over time, using the estimate at time "t" as the starting point for determining the ^ s / ^ 'o _max histogram at time "t + τ".

When using these inputs arise in the inventive

Procedure as output:

a sentence "optimized power control parameter". The communication node can then transmit these parameters to the transmitters using the parameters for the power control defined in [3]. The purpose of the method defined in this invention is that by using the new parameters, the ^ / ^ 'o distribution induced at the receiver will be "good" (high throughput and low burst loss rate).

an estimate of the maximum load that can be supported by the new set of "optimized power control parameters". With this estimate, the communication node can determine if the system is near overloading or not.

The method according to the invention works as follows.

1. The burst decoder provides a histogram of the the received bursts.

2. The histogram of ^ / ^ o is filtered. A low-pass filter (in the middle range) is preferred.

3. If an earlier Histogram exists, this is selected.

Otherwise, as discussed below, the procedure 1 applied to the ge ^¬ filtered E / ^ o Histogram to an initial estimate of the ^'o _maa - to obtain -Histogramms.

4. The method 2 is performed on the initial estimate of be ^¬ Wundt. This gives a refined estimate of the Mo _{TO £ tr} histogram.

5. The method 3 is based on the final estimate of be ^¬ Wundt. This method provides at its output a set of control parameters and an estimate of the maximum load that can be achieved using these power control parameters.

The following is a detailed description of methods 1, 2 and 3. Method 1

The block diagram of this method is shown in FIG. 8. This method provides an estimate of the histogram of 'o _m a. _r based on the filtered histogram.

A "custom filter" should be used whose impulse response is preferably in the shape of the left side of a bell-shaped curve.

1. Calculate the FFT of the filtered ^ Λ Α Ο histogram

2. Calculate the FFT of the custom filter

3. Calculate X = FFT (filtered ^ 7 ^ 0 histogram) / FFT (user-defined filter).

4. Calculate Y = IFFT (X)

5. Calculate Z = abs (Y)

6. The / ^Λ 'ο " _{ω: Ι} histogram is a filtered version of Z. The filter is preferably a low-pass filter.

Method 2

In this method, a "power control application" function is used which provides an estimate of the histogram after power control based on the ⁽ -histogram An example of this function is set out below. Method 2 is a numerical optimization method which attempts to find the Es / Nomax that induces a histogram closest to the E / N histogram obtained from the burst demodulator (see Figure 7). The metric to be minimized is the Euclidean distance between the histograms (estimated one). At this point many optimization methods can be used, e.g. For example, high-waste gradient methods, simulated annealing, and genetic methods.

The following is an example of the way in which the method could be implemented as a high-waste gradient method:

1. Apply power control to the ^Λ ο Jia Histogram. This gives an estimate of the ^ No histogram.

2. Calculate the distance between the estimated one obtained in step 1

Histogram and the I / O received from the burst decoder ^! o histogram (see Fig. 7). The distance can be the Euclidean distance between the histograms.

3. If the distance is below a specified target, the procedure ends, at its output the estimate of the - Output histogram. Otherwise, the / ^Λ Ό " _{ί £ ΰ} histogram is modified and the procedure moves to step 1.

4. The process is repeated until the distance is below the specified target or until a maximum number of iterations has been performed. Application of the power control function

It is assumed that the power control parameters listed in Table 1 define m segments for mode 0 and that the histogram is defined at n points X_l, X_2, ..., X_n. Reference is made to the values of the histogram as Y_l, Y_2, ..., Y_n and to the values of the estimated Histogram as follows:

Define an array of n elements, Z_l, Z_2, Z_n, and set all elements to 0

For each histogram bar X_j

Determine to which segment the point belongs. Point X_j belongs to segment "I" if and only if + l. Assuming that X_j belongs to segment i, the following is done:

Find the point X_u that is closest to z and set A0 = u.

Find the point X_u that is closest, and set AI = u.

Find the point X_u closest to lies, and set A2 = u.

For k = A0 to k = AI, with increasing k

Z_k = Z_k + (l-p_i) * Xj / (A1-A0 + 1)

For k = Al to k = A2, with increasing k

Z_k = Z_k + (p_i) * Xj / (A2-A1 + 1)

After terminating the loop over all histogram bars, the estimated ^ / Λ'o histogram is obtained by filtering Z with a filter whose impulse response corresponds to the left half of a bell-shaped curve as shown in FIG. Method 3

Method 3 is a numerical optimization method that receives as input the estimated ^E s / ^N o _m ax histogram. The method attempts to find the set of power control parameters that maximizes one of these two metrics:

the estimated maximum load (in packets / timeslot or packets / second)

the distance to a desired one Histogram.

At each iteration, the procedure performs:

Generate a new set of power control parameters Run the Power Control Application function using the new set of power control parameters and the estimated power control parameters -Histogramms. This function gives an estimate of the E »/ ^ o histogram.

Optimization metric. There are two options:

Calculate the distance from the estimated ^ / ^ ^' o histogram. The

Distance must be minimized.

Estimate the maximum load that can be achieved using the ^ ^ o histogram. The load has to be maximized.

If the E »/ No histogram reaches the target distance or target load, the procedure is stopped.

Otherwise, a new set of power control parameters will be generated.

Various numerical methods can be used, e.g. High slope / rise gradient methods, simulated "cooling" and genetic engineering.

In Fig. 10, an example of a high slope gradient method used to maximize the maximum achievable load is given. If the target load is not reached after a maximum number of iterations, the procedure is stopped.

The invention can be defined by the following properties:

The invention proposes a method which receives as input the E, / N () histogram of received bursts (packets) and provides as output a set of optimized power control parameters.

The procedure can be divided into two steps:

- Step 1. This covers procedures 1 and 2. Based on the ^ Λ / Λ'Ο histogram and knowing the set of power control parameters used by the transmitters, the Histogram estimated.

- Step 2. This covers the procedure 3. Using the estimated / ^Λ _'ο ίΤΜ! 2 · -Histogramms calculated set of optimized performance ^¬ control parameters.

The set of optimized power control parameters is obtained by:

Estimate the ^ / ^ o histogram using the power control parameters and minimize the distance to a desired histogram (e.g., a uniformly distributed histogram in dB).

Estimating the maximum achievable load based on the estimated EX / JSIQ histogram induced by the power control parameters.

example 1

It is the reverse link of a satellite communication system and in particular ^¬ sondere one of the beams of the reverse links considered. The transmitters use spread spectrum Aloha (SSA) with time slots, and the receiver uses SIC. Use all channels the same Übertragungsmo ^¬ dus such as: BPSK modulation

Code rate = 1/3

Spreading factor 32

Load 150 packets / timeslot

The link margin for a transmitter positioned in the middle of the beam is 20 dB, and the required for decoding Treg ₌ _ ₂ . ₇ dB.

The link budget loss due to the position of a transmitter in the beam is and follows a uniform distribution (-6.0 dB). The transmitter estimate of ^ b, L ^} > has a Gaussian distribution in dB with a mean ^ b and a standard deviation of 0.5 dB.

A share of 25% of the transmitters undergoes rain damping. The rain attenuation L _r has a Gaussian distribution in dB with a mean value of -10 dB and a standard deviation of 1 dB. The transmitter estimate of rain damping L _r has a Gaussian distribution in dB and has an average and a standard deviation of 1 dB.

Mode Es / No_i Es / No_2

1 0 22

Mode p_i

i o.o

In this example, the power control parameters are optimized to maximize the maximum achievable load.

The set of optimized power control parameters is: Mode Es / No_i Es / No_2 Es / No_3 Es / No_4 Es / No_s Es / No_6 Es / No_ ^

1 -1 8 11 14 17 22

Mode p_i p_2 p_3 _ p_5 p_6

1 0.0 0.0372 0-9877 0.4170 0.1715 0.6617

The set of optimized power control parameters is referred to herein as "new" power control parameters.

In Fig. 11 it can be seen that the ^ VNo histogram with the new power control parameters is closer to a uniform distribution than the histogram with the old power control parameters. In FIG. 12, the curve of ^ V ^ vs of of _ur ^ _ea | t _{s u n c} j new power control parameters is shown. It can be seen how in the new power control parameters the curve from Ύ to ^ is "flatter" and farther away from T ^ e ^.

For the "old" power control parameters, the maximum load was 185 packets / time slot, while for the new power control parameters the maximum load is 245 packets / time slot. Thus, in this particular case, the power control method allows for a 32% increase in load.

The invention can be used in wireless communication systems, such. In satellite communication systems and mobile communication systems. DIRECTORY OF ABBREVIATIONS

IFFT Inverse Fast Fourier Transformation

FFT Fast Fourier Transformation

SIC Successive Interference Cancellation

SSA Spread Spectrum Aloha

BIBLIOGRAPHY

R. De Gaudenzi, 0. Del Rio Herrero, "Advances in Random Access Protocols for Satellite Networks", 2009 International Workshop on Satellite and Space Communications, IWSSC 2009, Siena, Italy

ETSI TS 102 721-3 VI .2.1 "Satellite Earth Stations and Systems; Air Interface for S-band Mobile Interactive Multimedia (S-MIM); Part 3: Physical Layer Specification, Return Link Asynchronous Access."

EP 2 861 024 A1 (corresponding to DE 10 2013 221 866 A1)

Claims

A method for determining the power control parameters for controlling the power with which, in a communication system, signals from transmitters of a group of multiple transmitters to a receiver associated with that group are transmitted packet by packet in accordance with a multiplexing constraint

a. the stations are divided into one or more groups, b. the transmission power used by the transmitters in a group is determined such that the signal / interference ratio is between a minimum and a maximum value, c. each transmitter of a group determines its transmission power randomly such that its signal-to-noise ratio at the receiver is between a minimum and a maximum value, i. the method procedure for randomly calculating the transmission power depends on a probability value such that, as a function of this probability value, the signal-to-noise ratio achieved at the receiver is within the overall range of permissible signal-to-noise ratios or within the signal-to-noise ratio, which is allowed for the group to which the broadcaster belongs

e. wherein a first table defines the minimum and maximum signal-to-noise ratios that determine which group a transmitter belongs to,

f. wherein a second table defines the probability value used for the transmitter to determine its transmission power such that its signal-to-noise ratio is within the overall range of allowable signal-to-noise ratios or within the signal-to-noise ratio that is appropriate for the group is allowed to listen to which of the stations,

characterized,

G. new values for the first and second tables are defined such that H. maximizes the maximum load in packets per second that the communication channel can accept while maintaining a desired threshold packet error rate, and

i. the distance of a histogram from received signal-to-noise ratios in dB to a uniform histogram is minimized from an equal distribution of signal-to-noise ratios in dB, and

k. that defining new values for the first and second tables is performed in two steps, wherein

I. in a first step, on the basis of the signal-to-noise ratio histogram in dB at the receiver and the previous values of the tables, the histogram of the maximum signal-to-noise ratio in dB which the transmitters would reach is estimated, if they would send with their maximum available transmit power, and

m. in a second step, using as input the estimation of the maximum signal-to-noise ratio histogram in dB which the transmitters would reach if they were transmitting with their maximum available transmit power, new values for the first and second second tables are calculated.

2. The method according to claim 1, characterized in that the first step comprises the following substeps:

la. Apply a low-pass filter to the histogram of the signal

/ Noise ratio in dB,

lb. Making an initial estimate of the maximum signal-to-noise ratio histogram in dB that the transmitters would reach if they were at their maximum available

Would send transmission power,

lc. Using a model to estimate the histogram of the

Signal-to-noise ratio at the receiver based on the togramms of the maximum signal-to-noise ratio in dB that the transmitters would reach if they were transmitting at their maximum available transmit power,

iteratively refining the estimate of the maximum signal-to-noise ratio histogram that the transmitters would reach if they were transmitting at their maximum available transmit power, using a numerical optimization method that attempts to obtain the histogram of the signal / Noise ratio at the receiver based on the maximum signal-to-noise ratio histogram in decibels that the transmitters would reach if they were transmitting at their maximum available transmit power, using the model of substep 1c); the distance to the receiver's low-pass filtered histogram of the signal-to-noise ratio is minimized in decibels.

3. The method according to claim 2, characterized in that the second step comprises the following substeps:

2a. as input, using the output of the first substep, which is an estimate of the maximum signal-to-noise ratio histogram that would reach the transmitters if they were transmitting at their maximum available transmit power,

2 B. Using a model to estimate the signal-to-noise ratio histogram based on the maximum signal-to-noise ratio histogram in decibels that the transmitters would reach if they were transmitting at their maximum available transmit power,

2c. Calculate the new values for the first and second tables that optimize the metric defined in step h) and / or i).

4. The method according to any one of claims 1 to 3, characterized in that in the sub-step ld) one of the following numerical methods is used:

a. a gradient method

b. simulated "cooling off"

c. an evolutionary algorithm

d. a convex optimization method.

5. The method according to any one of claims 1 to 4, characterized in that in the sub-step 2c) one of the following numerical methods is used:

a. a gradient method

b. simulated "cooling off"

c. an evolutionary algorithm

d. a convex optimization method.

A method according to any one of claims 1 to 5, characterized in that the model for estimating the histogram of the signal-to-noise ratio at the receiver based on the histogram of the maximum signal-to-noise ratio in decibels that the transmitters would achieve if they were to transmit at their maximum available transmit power, carry out the following substeps:

a. Assume that the power control parameters in the first table define m signal / noise segments and that the histogram is defined at n points X_l, X_2, ..., X_n, where the value of the histogram as these n points is Y_l, Y_2, ..., Y_n is specified,

b. Denoting the "m-l" values in the second table by p_l, P_2, .. p_m-l,

c. Defining an array with n elements Z_l, Z_2, Z_n, and setting all elements to 0,

d. for each histogram value X_j, j from 1 to n: i. Determining to which segment the point belongs, the point X_j listening to the segment "i" if and only if X_j is between the i-1 th and the i th entry in the first table,

ii. after the segment has been determined and it is assumed that X_j belongs to segment i:

1. Find the point X_u closer to the first entry in the first table and define A0 = u.

2. Find the point X_u that is closer to the "i-1" entry in the first table and define Al = u.

3. Find the point X_u that is closer to the "i" th entry in the first table and define Al = u.

iii. for k = A0 to k = AI, with increasing k

1. Z_k = Z_k + (l-p_i) * Xj / (A1-A0 + 1)

iv. for k = Al to k = A2, with increasing k

1. Z_k = Z_k + (p_i) * Xj / (A2-A1 + 1)

e. after ending the loop over the histogram values X_j, obtaining the estimated histogram of the signal-to-noise ratio at the receiver by filtering the array Z by means of a low-pass filter.

A method according to claim 2 or any one of claims 3 to 6 when dependent on claim 2, characterized in that the impulse response of the low pass filter corresponds to the left half of a bell-shaped curve.

A method according to claim 2 or any one of claims 3 to 6 when dependent on claim 2, characterized in that the initial estimate of the histogram of the maximum signal-to-noise ratio in dB that the transmitters would achieve if they were to transmit at their maximum available transmit power, corresponds to an output of the first step from the past that has been stored.

9. The method according to any one of the preceding claims, characterized in that

the transmission powers with which the transmitters of the group transmit are within a total transmission power range, and

a parameter is assigned to each transmitter, which indicates how large the transmission power with which the respective transmitter transmits,

wherein the parameter is calculated on the basis of a random number and on the basis of probability values for how many transmitters of the group transmit with a transmission power which lie within different, specifiable and also within the total transmission power range lying transmit power segments.

10. The method according to any one of the preceding claims, characterized in that

a first table is provided in which different transmit power segments of the total transmit power range are defined, each defined by a lower limit and an upper limit,

a second table is provided in which a statistical probability value is included for each transmission power segment, which indicates how many transmitters transmit with a transmission power lying within the relevant transmission power segment,

a random number is provided for each transmitter, either by the transmitter itself or from outside the transmitter, on the basis of which the probability values to be assigned to the respective transmitters are calculated as their assigned parameters and thus to each transmitter the transmission power segment is assigned, within which lies the transmission power with which the transmitter concerned transmits, and

- The size of the transmission power with which the transmitter in question, within the transmission power associated with this segment is selected by another based on another random number calculation.

11. The method according to any one of the preceding claims, characterized in that

each transmitter has a maximum transmission power, each transmitter being assigned to the transmission power segment within which its maximum transmission power lies, with a random number computed taking into account a uniform distribution to a predeterminable number space, either from the transmitter itself or from outside it the transmitter, based on which it is determined with the aid of the probability value for the respective transmitter transmitter power segment, whether the transmission power of the transmitter between the lower limit of the relevant transmission power segment and a maximum transmission power or between the lower limit of the total transmission power range and its maximum transmission power.

12. The method according to any one of the preceding claims, characterized in that

each transmitter has a maximum transmission power, each transmitter being assigned to the transmission power segment within which its maximum transmission power lies, providing for each transmitter a random number calculated taking into account a uniform distribution to a predeterminable number space, statistically equally distributed within the number space or between zero and one may lie, and if the random number thereof is smaller than the probability value for the transmission power segment to which the transmitter assigned and transmits at a transmission power lying between the lower limit of the transmission power segment and the maximum transmission power of the transmitter, and

13. The method according to any one of the preceding claims, characterized in that

- The transmitters can transmit in different modes, wherein per mode control of the transmission power of all transmitters, as described above, takes place.

14. The method according to any one of the preceding claims, characterized in that

- The transmitter can transmit at different rates, with each mode, a different control of the transmission power of all transmitters, as described above, takes place.

15. The method according to any one of the preceding claims, characterized in that

- Before transmitting a signal from the transmitter to the receiver and / or at regular or irregular intervals, the control of the transmission power of the transmitter as stated above takes place.

16. The method according to any one of the preceding claims, characterized in that

the multiplex specification, according to which in the communication system signals from transmitters of a group of several transmitters to a group assigned to this group are transmitted in packets, is a random access specification.