Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B17/00—Monitoring; Testing
  • H04B17/0082—Monitoring; Testing using service channels; using auxiliary channels
  • H04B17/0087—Monitoring; Testing using service channels; using auxiliary channels using auxiliary channels or channel simulators
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B17/00—Monitoring; Testing
  • H04B17/30—Monitoring; Testing of propagation channels
  • H04B17/309—Measuring or estimating channel quality parameters

Definitions

• the invention relates to a method for generating noise in connection with the simulation of a radio channel and to a device implementing the method.
• Radio systems One significant problem in radio systems is the rapid variation of the properties of a radio channel with time. This relates especially to mobile systems, in which at least one of the participants in a connection is of- ten mobile. The attenuation and impulse response of the radio channel then vary within a wide phase and amplitude range even thousands of times per second. The phenomenon is random by nature, so mathematically it can be described by statistical means. The phenomenon complicates the design of radio connections and the used devices. [0003] There are many reasons for the variation in a radio channel.
• the signal When transmitting a radio frequency signal from a transmitter to a receiver in a radio channel, the signal propagates along one or more paths, in each of which the phase and amplitude of the signal vary, which causes fades of different lengths and strengths in the signal. In addition, noise and interference from other transmitters also disturbs the radio connection.
• a radio channel can be tested either under actual conditions or using a simulator that simulates the actual conditions. Tests conducted in actual conditions are difficult, because tests taking place outdoors, for instance, are affected for example by the weather and season that change all the time. Even measurements taken in the same place produce a different result at different times. In addition, a test conducted in one environment, such as a city, does not fully apply to a test conducted in another city. It is also usually not possible to test the worst possible situation under actual conditions.
• the testing of radio devices is usually done by ra- dio channel simulation.
• the simulation can be performed by means of a channel simulator, for instance.
• an external, analogue noise source is connected to the channel simulator.
• Channel simulation can also be performed in non-real time by using software in a computer, for instance, in which case noise can be generated by what is called the Box-Muller method.
• the Box-Muller method is characterized by the following equations (1 ) and (2):
• the figures x1 and x2 refer to in- read random numbers within a semi-open value range 0 to 1.
• the conversion produces as output two random numbers y1 and y2 that are Gaussian- distributed and the average of the distribution is 0, standard deviation 1.
• the formed random numbers y1 and y2 are used as real and imaginary components of a noise signal, for instance.
• a method for generating a noise signal in a channel simulator comprising: generating at least one random number, forming a component number containing at least one component, a numerical value associated with forming the at least one component of the component number being read from memory by means of the generated random number, and producing a noise signal by utilising the formed component number and summing the generated noise signal with an information signal in the radio channel of the channel simulator.
• the invention also relates to an arrangement comprising a noise generator.
• the noise generator comprises means for generating at least one random number, memory means for managing numerical values, means for forming a component number containing at least one component, the means being configured to read a numerical value related to the forming of the at least one component of the component number from the memory means by utilizing the generated random number, and means for producing a noise sig- nal in real time by utilising the formed component number, and means for summing the produced noise with an information signal to be transmitted in the channel.
• the invention is based on the idea that during conversion to a component number in the noise generator, the numerical value corresponding to at least one component is read from the memory means.
• the component number refers herein for instance to a complex number containing real and imaginary components.
• the component number can, however, also be a one-component number that only comprises a real component.
• the memory in turn refers to a table, database or register implemented by software or in some other corresponding manner, or a hardware solution suited for storing and reading numerical values.
• the formed component number is utilised when generating the noise signal.
• the noise signal is generated by summing several partial noise signals.
• Each partial noise signal is formed on the basis of the generated random number by turning the random number into a component number and using it to generate the noise signal.
• the distribution describing the component numbers can be extended by forming several partial noise signals.
• the crest factor (CF) of the distribution improves by using more partial noise signals in the distribution.
• the method is preferably implemented by means of two separate memory means, such as tables, with one of the memory means stor- ing the first part of the component number that is used as the amplitude value of the noise signal component.
• the second memory means correspondingly stores the second part of the component number that is used as the phase value of the noise signal component.
• the stored information is information calculated by the Box-Muller calculation method.
• the inven- tion is naturally not restricted to the calculation method, but the information read from the memory can also be formed in some other manner.
• the random number read in to the conversion is used as a pointer to indicate the correct location in the memory means.
• the information in the memory means can be arranged in or- der of magnitude or not.
• the arrangement for simulating a channel is preferably a channel simulator in which the simulation is done in real time.
• Real devices such as mobile stations, can be connected to the channel simulator for testing purposes.
• the inventive solution is implemented as hardware in the channel simulator to meet the requirement for real time.
• the noise generator of the invention can also be used as a stand-alone noise generator.
• the noise generator can also be used in generating a pseudo-noise signal, in signal generators and testing devices.
• One preferred embodiment measures the transmission power of the actual information signal and weights the generated noise signal by a weighting coefficient so as to obtain the desired signal-to-interference ratio of the noise signal and the information signal.
• the signal-to-interference ratio is preferably kept constant, but it can naturally also be varied with time in a desired manner.
• the weighting of the noise signal provides the advantage that the desired signal-to-interference ratio can be maintained even though changes occur in the power level of the information signal.
• a random number obtained from an M sequencer is used in forming the component number.
• the M sequencer preferably comprises several parallel lines that are at least partly de- layed in relation to each other. By using several parallel lines, the M sequencer is capable of producing the random number by means of one seed number at a symbol rate that at its maximum is the clock frequency of the logic. By means of the feedback nodes of the M sequences, the M sequencer can produce substantially uniformly distributed random numbers when the feedback con- nections are appropriately selected.
• the means implementing the method of the invention can be implemented for instance as ASIC (applica- tion-specific integrated circuit) or FPGA (field programmable gate array).
• the invention provides several advantages. When the values of the different components in the component number are read from memory instead of complex calculation, savings are obtained in the use of the processing capacity of the channel simulator.
• the processing capacity saved by the method can be allocated to other tasks of the channel simulator.
• the channel simulator used in simulation can also be built simpler, since fewer components requiring speed are needed in the device. It is clear that this, too, saves costs.
• FIG 1 shows one preferred embodiment of the method of the invention
• Figure 2 is a block diagram illustrating a channel simulator
• Figure 3 shows one embodiment of the mutual weighting of the information signal and noise signal in the channel simulator
• Figure 4 shows one preferred embodiment of a hardware solution of two-component number conversion
• Figure 5 illustrates the generation of a noise signal from partial noise signals
• Figure 6 shows a preferred embodiment for forming random numbers
• Figure 7 shows a preferred embodiment of an apparatus for forming random numbers.
• Figure 1 shows a preferred embodiment of the invention.
• the starting step 100 of the figure shows the initial situation of the method, in which the generation of the noise signal is started for the channel simulator.
• the noise signal being formed can be distributed in a desired manner.
• the noise can for instance be AWGN- distributed (Additive White Gaussian Noise) or alternatively, the noise can be coloured.
• method step 102 substantially uniformly distributed random num- bers are formed, on the basis of which the noise signal is generated.
• the random numbers are formed by the M sequence technique, for instance, that is later described in more detail in connection with the other figures.
• a random number formed during one clock cycle of a processor is a 16-bit number, for instance, that is divided into two 8- bit random numbers.
• the length of the random number is naturally not restricted to the above-mentioned embodiment.
• the amplitude value of the signal component is read from an amplitude table on the basis of the random number formed in step 102.
• the reading is preferably done in such a manner that the random number is used as a pointer on the basis of which the value is read from the amplitude table. For instance, 8 bits can point to 256 separate values in the table.
• the amplitude values formed in the table are approximate values formed with calculation formulas presented by Box-Muller, for instance.
• the accuracy of the read values can be selected as desired for instance in such a manner that even though the random number read in to the amplitude table is 8 bits in length, the approximate value of the amplitude value is presented in 16 bits to obtain a greater accuracy.
• the amplitude values in the amplitude table can be arranged in order of magnitude for instance in such a manner that the value 0 of the random number corresponds to the lowest possible amplitude value and the value 255 of the random number corresponds to the highest amplitude value.
• the amplitude values can also be in a random order in the amplitude table. Both implementation methods produce random amplitude values, because a uniformly distributed random number is used as the pointer.
• the numerical value corresponding to the second random number formed in step 102 is read from the phase table.
• the use of an 8-bit random number produces as output two 16-bit words, one of which is the sine value of the phase angle and the other the cosine value of the phase angle.
• the signal component to be transmitted to the radio channel is formed in step 108 that can be presented by equation (3), wherein r is read in method step 104 and the sine and cosine values of the phase angle in step 106.
• method step 110 several signal components formed in the above manner are summed. By summing several signal components, a distribution is obtained, the CF of which, i.e. the ratio of the maximum value of the distribution and the standard deviation of the distribution, is considerably better than in the case that the entire noise signal is formed by using a signal generated in the manner described above.
• Method step 112 shows how the trans- mission power of the actual information signal transmitted in the channel is measured.
• the input of step 114 is the known power level of the noise signal and the power level of the information signal.
• the ratio of the power levels can be maintained as desired, for instance as constant.
• the signal-to-noise ratio can also be changed with time in a desired manner.
• the ratio of the power levels can be adjusted to be exactly as desired in comparison with analogue noise by means digitally generated noise.
• Figure 1 describes the method of one preferred embodiment in the form of method steps, it is clear that the method is not bound to the order of performance of the steps. It is also clear that some method steps, such as steps 104 and 106, can be performed simultaneously. Even though it has been stated above that the desired values are read from the memory means on the basis of the random number, it is clear that the function to be executed is not only limited to reading, but it is also possible to per- form a few desired calculations in connection with it. The value of the component of the component number is then not necessarily exactly the value read from the table, but it can be multiplied by a constant, for instance. Figure 1 describes the method with respect to one formed random number. The described steps are preferably performed during one clock cycle so the described steps are performed several times during one channel simulation session.
• Figure 2 is a block diagram illustrating the structure of the channel simulator.
• a radio-frequency input signal U is taken in to the channel simulator simulating phenomena occurring in the radio channel, and a radio-frequency signal S is obtained as output from the simulator.
• the inputs and outputs of the signals are connected to radio devices being tested, such as mobile phones or base stations.
• the ideally modulated carrier is converted to an intermediate frequency.
• the intermediate-frequency signal is converted to complex baseband I and Q signals that are converted to digital in an analogue-to-digital converter 204.
• the digital signal samples are forwarded to FIR (finite impulse response) filter elements, in which several components delayed and weighted in different ways are generated from the signal received by the simulator.
• FIR finite impulse response
• the delays caused by the radio channel are added to the signal in delay elements 208 to 214.
• the delayed signals are taken to multipliers 216 to 222, to which the amplitude values A1 to A4 of the signal component are also brought.
• the signal components delayed and weighted in different ways are summed in an adder 224.
• Noise generated in a noise generator 206 is added to the digital signal in an adder 205.
• the noise signal also passes through the same channel as the information signal. This provides the advantage that it makes an input connector and external noise generator unnecessary when generating fading noise.
• Noise can also pass through its own channel, whereby fading interference can also be simulated.
• the combination signal formed in the adder 205 is converted to an analogue signal in a digital-to-analogue converter 226.
• the analogue I and Q signals are converted to an intermediate frequency in device unit 228 and further to radio frequency in device unit 230.
• Figure 2 also shows a control unit of the channel simulator, which in addition to the general functions of the channel simulator also controls for instance the delays of delay elements, the amplitude values fed into the multipliers 216 to 222, and noise generation.
• Figure 3 illustrates the weighting of a noise signal according to a preferred embodiment.
• a noise generator 206 supplies a noise signal whose power is measured and can be adjusted by a control unit 300.
• the power of an input signal U to the channel simulator is measured in a measuring unit 304.
• the power level measured by the measuring unit is transmitted to a control unit 302 that forwards the measuring result on to the control unit 300.
• the control unit can adjust the power level of the noise signal to obtain a desired ratio between the input signal and the noise signal.
• the control unit can adjust the noise signal by multiplying it by a weighting coefficient, for instance.
• Means for measuring an information signal 304 and means for weighting a noise signal 300 are implemented for instance by program or as FPGA circuits or by combining known methods.
• the noise signal and the information signal are summed in an adder 105 and forwarded to a digital-to-analogue converter.
• FIG. 4 describes the operation of the noise generator shown in Figure 2 in more detail.
• the noise generator has a device unit 400 that feeds a seed number to an M sequencer 402 generating a random number and having one or more lines.
• the seed number is read into the M sequencer 402 prior to starting the generation of the random number.
• the M sequencer Prior to starting the generation of the random numbers, the M sequencer can be filled entirely or only partially with seed numbers. Seed numbers can be zeros or ones in such a manner, however, that depending on the feedbacks, the seed number is preferably not only zero or only one.
• the clock of the device controls the operation of a converter 406.
• the converter receives 2*8 bits of data from one M sequencer during one clock cycle.
• the converter 406 uses the thus received 2*8-bit random number to read random numbers from tables T1 408 and T2 410. Even though the figure shows two tables, the information can also be in one table or it can be distributed to more than two tables.
• table T1 stores the amplitude information of the signal and table T2 the phase information of the signal.
• the converter receives the random numbers from tables T1 and T2 and generates a noise signal from them. It is clear that the information read from amplitude table T1 can also partly be used in determin- ing the imaginary component of the noise signal and correspondingly, the information read from table T2 can partly be used in controlling the real component of the noise signal.
• Figure 5 illustrates how only a part of the noise is generated by means of one converter.
• the partial noises generated by several converters 406A to 406N are combined in an adder 500, and the noise signal can be weighted as desired by a weighting coefficient in a multiplier 502.
• FIG. 6 shows a preferred embodiment of an M sequencer that produces the random number entering the converter 406 in Figure 4, for instance.
• the M sequencer contains flip-flops 600A to 600N that are shown as D flip-flops in the figure, but can also be other known flip-flops.
• the flip-flops serve as a shift register that can be stepped for instance 16 or 8 bits at a time, or 4 bits at a time, as shown in Figure 6.
• the number of lines in the sequencer depends on length of the random number read into the converter, for instance an 8-bit random number requires 8 lines.
• the lines are connected to opera- tional units 602A to 602D, and a logical operation, such as XOR or XNOR, is executed on bits that are fed back to the units. By means of the feedbacks, it is possible to produce uniformly distributed random numbers.
• Figure 7 illustrates the operation of the M sequencer 402 in terms of data transmission.
• the sequencer 402 has four parallel lines 700A to 700D underneath each other and a freely selected seed number is entered into each line during initialisation.
• Figure 7 shows vertically the distribution of the sequencer into shift registers 702A to 702D.
• the length of one line is for instance larger than or equal to 40 symbols, which means that an equal number of shift registers is needed.
• the figure shows how during the next clock cycle, a random number DCBA, formed on the basis of the contents of the last memory locations in each line, is obtained from the sequencer 402.
• the content AB- CDEF of the register gives as output ABCDEF, but in a parallel M sequencer group, the content ABCDEFGHIJKL of the register floats in a four-bit output in such a manner that the output is DCBA, HGFE and LKJI by clock cycle.
• the data hops in such a manner that the symbols in register 702B are shifted during the next clock cycle directly to register 702C.
• the output can also be directed to float in some other manner than four bits at a time depending on the length of the required random number.
• the figure shows one example of the feedback of symbols.

Landscapes

• Physics & Mathematics (AREA)
• Electromagnetism (AREA)
• Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Signal Processing (AREA)
• Monitoring And Testing Of Transmission In General (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=8562159

Family Applications (1)

Country Status (2)

Cited By (4)

Citations (2)

• 2001
  • 2001-10-31 FI FI20012108A patent/FI20012108A/fi not_active Application Discontinuation
• 2002
  • 2002-10-30 WO PCT/FI2002/000843 patent/WO2003039046A1/en not_active Application Discontinuation

Patent Citations (2)

Non-Patent Citations (2)

Also Published As

Similar Documents

Legal Events