WO2006026799A2 - Verfahren zur simulation eines mimo kanals - Google Patents
Verfahren zur simulation eines mimo kanals Download PDFInfo
- Publication number
- WO2006026799A2 WO2006026799A2 PCT/AT2005/000356 AT2005000356W WO2006026799A2 WO 2006026799 A2 WO2006026799 A2 WO 2006026799A2 AT 2005000356 W AT2005000356 W AT 2005000356W WO 2006026799 A2 WO2006026799 A2 WO 2006026799A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- channel
- development platform
- platform according
- simulation
- channel simulation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
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
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/0413—MIMO systems
Definitions
- the invention relates to a MIMO (Multiple Input Multiple Output) channel simulation and MIMO development platform for simulating at least one transmission channel between transmitters and receivers of a digital or analog radio transmission system emitting transmission signals, and further for use as a high performance development platform for software development and testing
- MIMO Multiple Input Multiple Output
- the invention is based in part on the patent application A 904/2003 "Channel Simulator", which has been significantly improved in the present invention to On the one hand, it can be used as a complete development platform for multi-channel (MIMO) radio transmission systems and, on the other hand, beyond the channel simulation.
- MIMO radio transmission systems are systems which, in addition to time and frequency, additionally use spatial multiplexing to increase the channel capacity.
- the transmit signal is nested into several packets, all of which are transmitted separately from the individual antennas of an array at the same carrier frequency (Fig. 1). Due to the slightly different emission location, different spatial propagation conditions result for the individual signals.
- the individual MIMO channels of the received signal can be distinguished from one another and integrated in the receiver to form a signal.
- FIG. 1 shows the principle of this real-time channel simulator: One channel each is formed by the transmission from a transmitting antenna to a receiving antenna. Each of these channels is preferably simulated by a unit of the channel simulator.
- Multi-channel simulations require a MIMO-compatible channel model.
- the channel model of patent application A 904/2003 "Channel Simulator” has therefore been expanded in many respects, including the consideration of multiple scattering (double scattering), which allows a more accurate adjustment of real propagation phenomena Role, because it significantly improves the realistic simulation of the channel capacity of these transmissions.
- the channel model was extended by a near cluster in the area of the base station. This so-called “Near Basestation Cluster” is especially necessary for transmission in micro cells.
- the subject invention is not only suitable for the simulation of MIMO radio transmissions, but it can be used as a powerful development tool for the development and testing of software-based flexible radio transmission systems (SDR).
- the subject invention in addition to the channel simulation system also includes a receiver for software defined radio.
- a platform is created that is suitable for the implementation of different software defined radio structures as well as different algorithms for each structural element and thus also for the testing of these structures.
- FIG. 3 A block diagram of an SDR receiver is shown in Figure 3.
- the left unmarked area represents the RF front end which mixes the baseband digital signals into the RF area and the RF signals into the baseband area.
- This area also includes the A / D converter or the D / A converter.
- the red marked part handles the processing of the digitized data. These include demodulation and decoding of the data.
- This part is entirely implemented by the RTS DSP Board.
- the RF and IF sections are implemented by a separate analog I / O board. Its tasks include the IF mixture and the A / D conversion.
- the overall system now consists of an extensible architecture 8 parallel simulation boards, each of which preferably takes over the simulation of a channel, as well as from one to two receiver boards, each of which can process 4 antenna signals.
- Own analogue I / O boards for all arithmetic boards with separate inputs for digital, analogue, RF and IF signals make the platform extremely flexible and allow the simulation of various MIMO configurations.
- LVDS lines enable the fastest possible data transfer of digital baseband signals between the boards and to the I / O.
- Figure 4 the structure of the development platform is shown schematically, whereby here for the sake of simplicity only the case of baseband transmission was used and the interfaces to other data formats are not shown. 2 channel model implementation
- a near cluster of the base station was taken into account.
- This so-called “Near Basestation Cluster” is especially necessary for transmissions in micro cells.
- the base station (BS) is usually mounted at the level of the roof edge, not high above the rooftops (as in macro cells) the base station lying building Scatterer, which are described in the simulation by the "Near Basestation Cluster".
- the structure of the channel model also changes from Patent Application A 904/2003 "Channel Simulator” in that some functions are shifted from the SS (Small Scale) Update to the LS (Large Scale) Update, in the SS Update mostly only linear or quadratic This saves a lot of computation time, especially the calculation of the path length, which was carried out in patent application A 904/2003 "Channel Simulator” for every single path in the SS update, is now replaced by a quadratic interpolation. The path length calculation accounted for most of the computational power because it required the evaluation of 64-bit precision roots.
- NBS Near Basestation Scatterer
- the NBS are scattered similar to the FS (Far Scatterer) at the beginning of the simulation and maintain their position throughout the simulation. This means that the path lengths between the NBS and the base station antennas, as well as the path lengths between NBS and FS, can either be precalculated by the GUI or calculated in the simulator in the initialization phase. The same applies to the angles between the BS antennas and these scatterers.
- the extension of the channel model proposes to add an independent vertical component to the coordinates of each scatterer. This is expressed by the elevation angle and allows the use of three-dimensional antenna models. Elevation is also important for MIMO capacity simulations.
- the scatterers are now scattered in a cylinder, not in a circle, as in the patent application A 904/2003 "Channel Simulator” (see Figure 7) .
- the height of the cylinders can be adjusted by the user via the GUI or upon initialization of the simulation The random process is used for the "angular spread" and the mean and standard deviation are used to determine the height.
- the polarization values of the signals are taken into account for the simulation, whereby two variants are provided for this purpose:
- random but constant polarization values are determined for all scatterers during the initialization
- the polarization parameters are updated each time a scatterer is generated or placed.
- the polarization values become the respective scatterers (such as the area values, which are the "reflection factors" of the scatterers, which represent the ratio of the reflected wave to the incident wave). assigned.
- this second variant has the disadvantage that the program code can lead to different runtimes, because not all polarization factors are newly determined in the LS update, but only those of the added scatterers.
- the attenuation factors such as shadow fading, scatterer area ..., describe the attenuation in propagation processes, which are limited to the polarization plane W (vertical - vertical).
- additional factors for the other polarization levels in the LS Update are determined, which reflect the change in the polarization directions due to the scattering at the scatterer.
- XPD _VH ( ⁇ PD_VH_mean + XPD_VH_spread * randn ⁇ ) * e j * 2 * ⁇ * randO
- XPD _HV (XPD_HV_mean + XPD_HV_spread * randn ⁇ ) * e J * 2 * ⁇ * randO
- POL_HH (POL_HH_mean + POL _HH _spread * randnQ)) * e * * * * rmi ⁇ i
- XPD_VH Cross Polarization Factor for the Polarization Plane Vertical - Horizontal XPDJHV . Cross Polarization Factor for the Polarization Plane Horizontal - Vertical POL HH ... Polarization Factor for the Polarization Plane Horizontal - Horizontal
- the following simplification has been chosen for the implementation.
- the polarization factors are combined in the LS Update together with the antenna gains to form a factor. This is true for all paths, except for those that pass through Near Mobilestation Scatterer.
- This simplification results from the fact that the antenna direction does not change much by the movement of the MS within a LS update and thus the antenna gains remain nearly constant. For paths via Near Mobilestation Scatterem this simplification can not be chosen because the antenna gains are not constant due to the small distance between antennas and scatterers and are calculated in the SS update.
- the attenuation values for the W polarization are calculated without consideration of the antenna factors (see also section 2.4). These attenuation values are then combined with the polarization factors and the polarized antenna gains to form the resulting attenuation factors, which are passed to the FPGA (Field Programmable Gate Array). This combination takes place either entirely in the LS update, as just described, or partly in the SS update (for paths via Near Mobilestation Scatterer).
- FPGA Field Programmable Gate Array
- the calculation of the antenna gain should be done as far as possible in the LS Update. However, this can not be done for the NMS (Near Mobile Station Scatterer).
- the antenna gains for the NMS are determined from tables in the SS update, whereby the angle calculations are interpolated partly linear or quadratic. This reduces the required computing power compared to the computationally intensive arctangent calculation. Since the base station is fixed, the antenna gains of the base station antennas remain constant until scatterers are replaced with new ones. Due to the consideration of the scatterer polarization, the polarization of the antennas plays an important role in the calculation of the antenna gains.
- the antenna arrays can only rotate around the Z axis, but not around the X and Y axes.
- the rotation of the array around the X and Y axes would make the calculation of the antenna gains much more complex, as it would also change the polarization planes of the antennas.
- the channel model also includes the elevation, 3D antenna patterns must be used to calculate the antenna gains.
- the Near Scatterer Area calculation was carried out in the channel model of the patent application A 904/2003 "Channel Simulator" according to COST259
- the scatterers were equally distributed in the Near Scatterer Cluster and were weighted by the Jacobian Transform, the Power Delay Profile and Angular Power Spectrum given by COST259 to obtain.
- One approach to simplifying the Jacobian transform is to reduce the accuracy of the length resolution.
- 32-bit lengths are resolved.
- the LS update interval at maximum speed is 10 ⁇
- the determination of the area values of a circular ring takes place via the Jacobian transformation.
- a few scatterers (as many as desired circular rings) are distributed over the radius and the area value calculation is carried out at these positions.
- the scatterers do not have to be transferred between the coordinate systems.
- significantly fewer scatterers must be considered than in the entire cluster.
- a possible additional division of the circular rings into sectors could still be carried out. This could allow for a better approximation to the Jacobian transformation.
- a linear interpolation of the area values is performed.
- the FS area and the LOS area correction values are also linearly interpolated.
- Figure 10 shows a single scattering path in red, characterized by a certain direction of incidence and a delay defined over the length. Furthermore, some double scattering paths are shown in blue. Since all DS paths arrive via the same second scatterer as the single path, but have different first "Double Scatterers", they all come from one direction but have different delays and attenuation values.
- NMS _ area (NMSno) NMS area value (weighting of the scatterers)
- NS _area _correcture NMS area value (weighting of the scatterers)
- the variable NMS_DS_weight should allow weighting of the DS over NMS compared to single scattering. This variable is initialized by the GUI (Graphical User Interface) and remains constant for a realization. If the variable is modified, a new MC simulation must be performed. Since both Double Scatterer paths have the same correction values, the root calculation via NS_area_correcture and NC_SF can be omitted. As you can see in the following example, it has to be done with the Far Scatterers.
- NMS _ area ⁇ NMSno * ⁇ NMS _ area _ correcture * NMC _ SF *
- the two roots are combined in the calculation in the DSP to one in order to save computing power. Furthermore, the root only has to be computed once for all DS paths between NMS and FS, since it does not depend on the scatterers used.
- NBS SS (NBSN) NBS areaJNBSnr) * ⁇ S_area __ ⁇ rectrect sp
- the LS update is performed at maximum speed every 10 ⁇ . Therefore, the table step size is also increased to 10 ⁇ . Since all path lengths are expressed in multiples of lambda ( ⁇ / 1024), table access is facilitated. The update should reach 10 ⁇ at maximum speed.
- the Far Cluster Visibility Regions are circular. The following simplifications are used to calculate if the MS is in a Visibility Region:
- the cosinusoidal range extends from 0.9 to 1.1 times the radius of the Far clusters Visibility region. Within this range, the cosine function is read from a table. If the distance is greater, the visibility is set equal to zero, which makes the components of the FC equal to zero. If the distance is less than 0.9 times the radius, the visibility is set to 1.
- Interference signals vary depending on the system being simulated.
- CDMA Code Division Multiple Access
- a number of mobile stations operating on the same frequency and using different spreading and scrambing codes will act as interference if that number in the area of Cell utilization increases.
- the base stations in neighboring cells represent the interference signals.
- UMTS uses uplink a power control for each mobile subscriber, so that the same receive power arrives at the base station from each mobile terminal.
- various mobile stations are assumed to be interferers and normalized by means of the power control to a common level.
- the signals of these mobile stations can be summed and stored as an interference signal.
- the "sturgeon" mobile stations move in circular or elliptical orbits in order to simulate the Doppler effects of the interferers ( Figure 14), whereby a limit does not represent the number of disturbers but the size of the orbit and the speed of the movement the samples of the entire trajectory must fit into the memory of the DSP board.
- the number of disturbers is selected based on the selected spreading codes and thus the utilization of the cell or the sector of the cell.
- a difficulty that arises with a special feature of the simulator is that syncing aids are provided for various receiver tests. These are designed to subtract the time lag of the Line of Sight component.
- the system can tell the mobile stations how many icons they should adjust their "internal clock” and start earlier to send, so they do not timings hurt their timeslots. If one wishes to simulate a system in which all MS have the timing to be examined and the interference MS have a timing advance, then each MS must subtract its respective time delay from the LOS component (shortest path between the antenna arrays of the base station and the mobile station) become. (Each MS should arrive at the BS synchronously.)
- Downlink transmissions assign more base stations to a cell (Figure 15).
- Each of these "sturgeon” base stations receives via their cell the middle attenuation course to Cost Hata or Walfish - Ikegami, similar to the same applies to the mobile station in their "home” cell.
- no spreaders are assumed, the signal comes only from the direction of the base station, the signal strength being determined from the attenuation curve of Cost Hata and Walfish - Ikegami, and this attenuation value is folded with previously stored modulated values ,
- the base stations do not need to be synchronized in the system 1 ). This means for the synchronization help of the simulator that also the withdrawn LOS has no influence. So it does not have to be taken into account whether the LOS component was deducted between BS and MS and the "Stör" BS considers this or not.
- the disturber (Fig. 16) will be placed somewhere in the coordinate system of the cell (of course also outside the cell itself).
- the signal strength of the disturber is determined by the Cost Hata and Walfish - Ikegami Pathloss models as well as by the send level of the disturber.
- the conditions of the Pathloss model of the jammer are the same as those of the base station.
- the phase of the attenuation signal is determined by the distances of the interferers to the MS antenna.
- the interference signal is defined in the baseband over an amount and a phase.
- the phase change per sample value is also transferred and represents the frequency of the baseband signal. This results in baseband in a sinewave signal.
- the calculation takes place within the FPGA and can be performed parallel to the arithmetic unit of the channel model paths. Therefore, no additional board is needed as in the aforementioned chapters on the CDMA signal jammers.
- the delays of all signal components must be taken into account in the simulation, so that the receiver has a correlation between transmission and transmission. Can produce received signals. As a result, no synchronization in the receiver is necessary, which would be very complex to implement there. In the simplest case, this is done by using the transit time of the LOS component as a measure and shifting the received signal by this transit time. For a more accurate calculation, the term of the shortest path that occurs between the antennas is used instead of the LOS.
- Figure 17 illustrates how the shortest path delay is subtracted between the antennas.
- Each board receives the positions of all antennas, but only calculates the LOS connections for all TX to RX paths ( Figure 18).
- the indirect connections via scatterers are only calculated for the TX to RX path active in the respective DSP board.
- Each board receives the maximum distance from the individual TX antennas to the TX center and from the RX antennas to the RX center.
- the DSP then calculates the distance between TX and RX center and then subtracts the maximum distances. This results in the shortest possible delay.
- this calculated delay does not always correspond to the actual shortest path and therefore, under certain circumstances, no perfect synchronicity can be achieved.
- the shortest delay can also be subtracted from each "interferer" MS and the sum of these interferers then in RAM be stored.
- the signals of all MS would arrive synchronously at the BS. This would correspond to a system with timing advance and is shown in Figure 19c).
- the computing power required for a simulation of multichannel transmissions is provided in the subject matter of the invention by the interaction of several (or a plurality of) computing boards.
- the division of the computing power preferably takes place in such a way that each board calculates exactly one MIMO channel (a connection from a transmitting antenna to a receiving antenna).
- MIMO channel a connection from a transmitting antenna to a receiving antenna.
- partial results must be passed between the boards and input signals distributed.
- a very powerful data communication between the boards is required, which is performed in the subject invention via a backplane with fixed wired LVDS connections.
- the backplane design hardware consists of a board with two or more plugs connected to the signal processing boards through a connector or a cable (see Figure 20).
- a connector or a cable see Figure 20.
- a preferred embodiment of this connector by means of a novel plug-less technology is described in Section 3.1.
- Each signal processing board leads to this connector in addition to data signals also control signals, such as. serial control signals (RS232 7), HC bus signals and synchronization signals. Furthermore, the power supply with constant voltages and variable currents can be obtained via this plug.
- data signals are converted to the LVDS channel link Proposal for 32 bit signals used.
- the data signals are serialized into 6 signals and transmitted via LVDS level (Low Voltage Differential Signaling) to 6 pairs of lines.
- LVDS level Low Voltage Differential Signaling
- 2 pairs of LVDS clock signals This data connection described in this way is referred to below as the LVDS channel (see Figure 21).
- the backplane also has the task of addressing each board via a fixed, unique slot address and an adjustable housing address (for example with a DIP switch) for the data direction Backplane -> Board.
- this addressing can be set up with 3 address lines, one data line, one read line and one multiplexer (see Figure 22).
- signals from the signal processing board can also be routed across the backplane (see Figure 23).
- the backplane also has the task of transmitting synchronization signals from a signal processing board (master board, for example in slot 0) to the other boards.
- a signal processing board master board, for example in slot 0
- GTLP drivers and GTLP receivers are used. These sync signals can also be forwarded to the back of the device via a corresponding backplane extension to enable synchronization with other systems (Fig. 24).
- Over temperature monitoring can also be provided on the backplane.
- Simple bimetallic switches are provided for both a warning threshold and for switching off the device.
- each signal processing board On each signal processing board, 4 LVDS channels are routed to the backplane, which can operate as both input and output. 2 of these 4 channels are connected via the backplane to a daisy chain in which each board gets its input from the previous board and its output leads to the next board. In this way, the individual results of the respective MIMO channels can be added up and the result forwarded to the next board (see Figure 25).
- the 1st input of the 1st signal processing board and the last output of the last signal processing board are routed via the backplane extension system to the back of the housing in order to extend this daisy chain to several systems.
- a simple backplane extension system routes the 2 remaining LVDS channels (input and output) of each board to the back of the enclosure. With the help of cables can thus on the back wall of the input signal distributed to obtain the desired MIMO configurations. This wiring can either be made by the customer or fixed by the manufacturer (see Section 4, Fix Wiring of the Boards for Different MIMO Systems).
- This backplane extension system also provides for the use of a large programmable switch matrix for LVDS channels. Thus, no manual plugging is required for distributing the input signal.
- the switching matrix for LVDS channels can be realized by a large FPGA (see Figure 27).
- Circuit board 1 has extensions at the end that are similar in shape to connectors in a PCI connector system, with traces on each side of each extension (see Figure 29).
- the printed circuit board 2 has longitudinally milled recesses (see FIG. 30) which receive the extensions of the printed circuit board 1.
- the design and manufacture of the printed circuit board 1 is standard.
- the tracks should be routed as far as possible to the edge.
- the track width may be e.g. 0.4 mm and the track spacing 0.6 mm. This results in a density of 1 trace / mm for each side, so 2 traces / mm for the entire connector.
- the slot-shaped recesses of the printed circuit board 2 should have a width which corresponds to the thickness of the printed circuit board 1 plus the tolerance dimensions (with standard printed circuit boards the thickness is 1.6 mm, the milling should therefore be 1.8 to 2.0 mm).
- the recesses are milled after two vias have been placed on each trace, which are half milled away, and make the electrical contact to the circuit board 1.
- the conductor tracks of the printed circuit board 2 must be performed with the same grid as on PCB 1 up to the milling edge on the top and bottom (the tracks on the top and bottom must have the same electrical potential). Exactly on the two milling edges a through-connection must be set (connecting conductor of the top and bottom side).
- the drill diameter for the vias must be smaller than the track width and the Pa s ur c ontaktmaschineen must not be much larger than the trace width. In any case, the insulation distance between the pads must be respected. For a trace distance of 1.0mm and a trace width of 0.4mm a via hole of 0.3mm and a pad diameter of 0.5 - 0.6mm could be provided.
- the slots are milled. As described, all through-holes are half milled away (see Figure 31). After mechanical mating, the solder joint can then be achieved by wave soldering (see Figure 28 for the direction of the surge).
- the most varied MIMO / MISO / SIMO configurations can be realized with the subject invention.
- the individual boards are z.T. connected via a fixed cabling (backplane).
- the individual inputs and outputs can be switched by software, in order to be able to use them for different purposes, which results in a large number of possible uses.
- the following solutions are possible:
- each individual LVDS channel fixed to one input or one output.
- one input and one output are conveniently accessible on the front panel, the remaining 2 inputs and 2 outputs are routed to the backplane via a 96-pin plug.
- One output on the backplane is connected to one input of the neighboring board (upper link bar), the second output on the backplane is connected to an input of the next but one board (lower link bar, see sketch).
- the individual channels are used for different purposes depending on the size / configuration of the MIMO system, e.g. for looping the input data or for passing on the intermediate results. Due to the fact that one does not commit to fixed usage modes of each channel, one achieves (with the same, fixed wiring) considerably more possibilities of use.
- MIMO n ⁇ 2 works very much like MISO n ⁇ 1. While in MISO n- »1 the base unit is a single board, a set of 2 boards is always considered here ( Figure 33). The distribution of the input signal within a set takes place via the upper link bar. The lower link bars provide convenient passing on of the intermediate results. (Logically equivalent to two nested MISOs!
- each board needs to know its relative position in the chain in order to properly size the input or output buffers of the channel engine.
- the board is aware of whether it belongs to the first or second receiver string and either obtains the input data from the front panel and hands it over to the backplane or via the backplane. Attention: The handover via the backplane results in a further (minimum: approx. VA to Vz LVDS clock) delay of the input signal, which may be perceived as disturbing.
- the main advantage of the solution with only one fixed wiring is that only (intermediate) results from the board to the next are passed through the daisy chain, but never modulator signals are distributed.
- the modulator signals are then flexibly distributed over user-attachable cables. This is in marked contrast to the solution with two wired daisy chains, where, depending on the MIMO degree, one or the other daisy chain was used for each of these tasks. Due to the more flexible wiring, it is already known when programming the channel model at which input the board receives intermediate results and where modulator signals are received without the MIMO degree having to be known. This makes the development of FPGA logic much easier ( Figures 36 and 37).
- MIMO grades which are not a power of two, can now be filled across multiple racks with no gaps between the racks' boards. By using multiple racks, it is of course again necessary to loop the daisy chain through a cable.
- the central task of a digital receiver is to decode the (digital) base-band signal and recover the (useful) data contained in it.
- the base band data of each receive antenna are supplied as I / Q (amplitude / phase) pairs with a value resolution of 32 bits each.
- the sampling times are selected such that between each (no oversampling) and eight (8x oversampling) data values are measured per bit time.
- GSM uses TDMA (Time Division Multiple Access) in frequency bands of 20OkHz width.
- These 156.25 bits are composed of 3 head bits + 57 data bits + 1 kill bit (spacer, always 0) + 26 bits midamble (fixed pattern for synchronization and (approximate) measurement of the channel impulse response) + 1 killbit (spacers, always 0) + 57 data bits + 3 tail bits + 8.25 bit safety margin against jitter, or in short, 114 (2x57) bit user data, 34 (3 + 1 + 26 + 1 + 3) bit management and 8.25 bit safety margin ,
- UMTS uses WCDMA (Wideband Code Division Multiple Access).
- WCDMA Wideband Code Division Multiple Access
- each subscriber simultaneously transmits on the same wavelength in a 5 MHz wide frequency band.
- UM organ transmits its transmission at 10 ms, each frame having 15 timeslots (at 0.667ms) divided, and each frame containing 2560 chips (at 0.26 ⁇ s * 3.84 MHz).
- An RTS board receives digital baseband signals from the front panel via 2 LVDS connections, which are fed to the FPGA. These two LVDS connections are connected to the DSP board via an additional print.
- the RTS board has a 96-pin connector with additional LVDS connectors, with which up to 4 further digital baseband signals can be transmitted.
- An auxiliary board accepts 2 such digital baseband signals on its front panel and forwards them (without further processing) to the backplane. From there, the data is then transferred to the RTS board.
- the receiver board has 4 digital baseband inputs (2 own + 2 via the auxiliary board).
- receiver configurations are shown.
- RTS boards wider by piggyback board for forwarding the two LVDS Channels from the front panel to the connector in the middle of the board
- receiver expansion boards (2 LVDS channels are passed through from the front panel to the backplane terminals without any active logic).
- the task of the backplane (with fixed cabling) is now (apart from power supply and clock forwarding) to ensure that all LVDS signals can be received by an FPGA.
- the simplest way to set up a MIMO x ⁇ 6 receiving system is with 2 RTS boards (see Figure 41).
- One of the RTS boards is extended by two additional LVDS inputs (in total now 4) via the expansion board.
- This RTS board must now receive 4 LVDS inputs and preprocess the data stream contained in it (eg compressing by redundancy elimination etc. - (see below for data compression possibilities) .
- the other RTS board will receive data via its two LVDS inputs. Receive channels and also preprocessing. Now (in each case a selection of essential) data is played on one of the two boards, where those computation steps are processed, for which a total overview of all data is required, and then distributed the knowledge gained back to both boards.
- the data streams can be safely compressed so that they can be transmitted over a single LVDS line ( Figure 43).
- the Coldfire ⁇ Controller has a 100 MBit / sec Fast Ethernet interface, via which (measured real) a data transmission via TCP / IP of about 2.3 MByte / sec (or 3.3 MByte / sec with UDP) is possible (limited by the CPU performance of the Coldfire).
- the few 100 Kbytes / sec that result in the reception algorithms can therefore be easily transferred via Ethernet to the (Windows) PC running the controlling program and the user interface.
- Rapid prototyping is based on the "golden code” paradigm, ie there is a central code that different teams can work on and that is written in an abstract language that can be mapped to different platforms so that changes can be made centrally and there are not several implementations of an algorithm in different languages necessary.
- the header of the Golden Code specifies the number of inputs / outputs and their data rates.
- the platforms may be either a simulation tool, the actual system, or the signal processing hardware related to the invention. Using the mapping tool, the golden code can now be mapped to the various platforms (see Figure 44).
- the golden code can also be managed in multiple blocks of code that can be mapped to either DSPs or FPGAs, using the mapping tool to Automatically perform this partitioning (which block is mapped to which block).
- GenC program automatically maps algorithms ("Golden Code") implemented in Generic C into either Simulink blocks or the signal processing hardware related to the invention, and also automatically blocks blocks on the DSP or FPGA ,
- a library has been developed which realizes the communication between the signal processing hardware related to the invention and the simulation tool Simulink.
- GenCaddOn the files generated by GenC can be extended by initialization, send and receive routines, so that from the Simulink model individual blocks can be outsourced to the simulation board and executed from the PC on the board. For the simulation, this means that a block controls the data exchange between hardware and PC.
- the present invention does not produce the "loop" over expensive JTAG interfaces, but over a Fast Ethernet
- the signal processing hardware relating to the invention has the possibility of being flashed via Ethernet or by special Simulink I / O blocks, which establishes the connection from Simulink to the signal processing hardware via Fast Ethernet , 7 Implementation features of the subject invention
- the channel engine calculates the convolution of the individual multipath signals and is implemented in essential points in accordance with the patent application A 904/2003 "Channel simulator.”
- the decisive improvements over the aforementioned application are given in the following section.
- the GSCM channel model takes into account many propagation paths of different lengths. Some paths have a much higher length than other paths, causing intersymbol interference. This means that the time delay of these paths is higher than the symbol duration and therefore not the same symbol as over the shorter paths, but the preceding symbol is received.
- a HistoryRAM is used, which records the time course of the modulation signal over a certain period. Since in this special case very many different read accesses are made to this HistoryRAM (the addresses correspond to the delays of the paths), but write accesses are made only with continuously increasing address, the HistoryRAM was implemented with the Double Dual-Ported RAM shown in Section 7.2.
- the Channel Engine receives either an internal (DSP calculated) or an external modulation signal at the input and resamples this signal using a multirate interpolation filter to increase the temporal resolution, in addition to forming the difference value of each sample from the previous sample.
- This upsampled signal is stored together with the delta values in the HistoryRAM.
- the signal value is approximated when read from the History RAM by means of deviation values predetermined by the DSP (successive approximation).
- a temporal resolution in the range of 1 ns can be achieved.
- the HistoryRAM is realized with a novel Double Dual-Ported RAM.
- This invention describes a data buffer which makes the data supplied by a data source (producer) constantly available at high frequency to several data receivers (consumers). In this case, with each access of a consumer, within an interval defined by the buffer size, it is possible to freely choose how many data packets should be accessed from the current time back to the past.
- This invention describes one way of using a buffer made from ordinary single-ported RAM simultaneously for reading and writing, or both ports, by skillfully utilizing the internal RAM organization of the buffer, as well as the fact that the write accesses are strictly ordered Dual-ported RAM for reading and still being able to write data.
- a larger RAM area such as implemented in an FPGA, is composed of several small block RAMs of fixed size, each with its own address and data lines. Only a logic above, which is automatically created by the VHDL compiler, builds a single, homogeneous memory. Instead of this standard logic, the improved logic underlying this invention can now be used as exemplified in Figure 45 below.
- the BlockRAMs used in Figure 45 are Dual Ported.
- the port shown in Figure 45 above is for reading only, with several, exemplarily two, consumers sharing this port in time division, while the port shown below is for simultaneous reading by a plurality, in the figure exemplarily two, consumers and the Writing by the producer serves. If you only have single-ported block RAMs, the upper part of the map will fall away and the single port will be used for simultaneous reading and writing.
- the individual block RAMs are arranged according to the principle of the cyclic chain.
- the foremost link in this chain represents the current write block RAM, while behind it, sorted by age, are the read block RAMs.
- the producer has written his 'blockRAM', it is placed in the front of the read block RAM block and, according to the cyclic chain principle, the 'oldest' read block RAM is assigned as the new write block RAM.
- LVDS connections were selected based on the 21-bit Channellink standard.
- the data sets consist of 64 bits, dividing each into 32 bits for real part (I) and imaginary part (Q).
- Each data packet consists of 21 bits. Of these, 16 bits are payload and 5 bits overhead. The overhead is regenerated for each data packet to be transmitted.
- User data area user data to be transferred
- Flag is used on the receiver side to receive the received
- Timeslotflag This flag is used for the first one
- Validflag identifies a valid data packet In each receiver board or in the FPGA of this board 4 LVDS interfaces (receivers) are implemented. An LVDS receiver now receives 64-bit data sets (symbols) from a transmit board, which represent the receive data of a receive antenna. Due to different LVDS lines and asynchronous data transmission, the individual LVDS signals are offset in time by up to 1 symbol length.
- the data of the individual LVDS receivers must be synchronized for each timeslot. This is done via the timeslot flag and the valid flag of the LVDS data stream.
- each LVDS receiver are buffered in each case in a register for synchronization, then they are written by the controller according to the invention in a DPRAM and are available synchronously for processing.
- the register Path select is available, in which is entered for each timeslot, which senders are active. As soon as a timeslot flag of a send- er board is detected, the respective bit is set in the 4-bit path select register.
- a counter is set to the start value. With the first (and each subsequent) incoming valid flag the counter value is reduced by 1 and from this point on the data is written synchronously by the 4 registers into the DPRAM.
- the used uClinux has to be reconfigured very often and recompiled.
- the uClinux image is saved in flash memory in compressed form and is decompressed into the SDRAM memory by pressing the reset button of the RTS-DSP board from the bootloader, which is also stored in the flash memory, and then started.
- Compiling the image includes compiling the kernel, individual user programs, and merging each binary file
- the make xconfig command calls a very user-friendly tool for configuring the kernel and user programs of ROMFS.
- the processor and the architecture is selected and partly configured.
- Kernel modules such as file systems, interfaces, and drivers are also configured here. Disadvantages prove to be configurations and changes of the source code of a certain program, these must be done in the respective special directory. Here too, there is no need to make any changes unless there is security in terms of compilability and correctness.
- the source directory is used by each module selected in the configuration and the respective make command is executed in it.
- the executable files are created in this, with user programs in the ROMFS being copied into the bin directory and kernel components being present as a whole binary kernel image.
- All sources are of course compiled with the compiler of the target system (eg m68k) on the development system (eg Linux PC / i386).
- ROMFS (usually a directory with all files, scripts and directories required on the target system) is created from the ROMFS on the development system. an original first t. This mage is linked to the kernel image and the resulting binary file is converted to zip format.
- the d Bug Monitor program should be used as a bootloader.
- the dn image via TFTP to SDRAM download
- go image from SDRAM address 0x20000 start
- the image is then available on the target system and it can be used with this, but after a reset no longer exists.
- several changes of the scripts or carried out manual starts of programs must be made after a new download.
- the COLILO Bootloader can be installed. After a reset, this decompresses the uClinux image programmed in the flash memory into the SDRAM memory and then starts it automatically.
- the uClinux image must first be programmed into the flash memory using BDM (Background Debug Mode) cable. Although this method does not require the re-download of the image from the development system (because the image exists locally on the target system), the problem of volatile memory and the lost changes to the scripts and programs is not resolved.
- the method with COLILO is rather impractical for uClinux image development, since every new image must also be programmed into the flash memory as mentioned above using BDM cable and local connection to the target system.
- the development system is used for each change and the respective changes made on it, and then the uCLinux image is created. After that it has to be transferred to the target system to be able to test it.
- the entire flash memory In order to configure or secure certain parts of the system, these must also be changeable separately from the system. If the entire flash memory is considered, then it consists of only a single partition. In this partition is the boot loader at the address OxO, this is connected to the kernel image from address 0x40000 and this in turn to the ROMFS (address 0x40000 + kernel image length). To program a new image into flash memory, all components must first be linked together on the development system and secondly programmed as a whole. To counter this problem, the flash memory is divided into several partitions (Figure 47).
- the bootloader partition also contains the COLILO boot loader. This decompresses after a reset of the kernel and ROMFS image from the second partition into the SDRAM memory and starts this afterwards.
- the kernel image and ROMFS partition contains a minimal uClinux image created on the development system as well as initialization and boot routines that the system uses minimal performance starts. At the end of these scripts, the last partition is mounted (user flash partition) and called in this user-changeable scripts.
- the user flash partition is the actual renewal on the whole system.
- RAMFS 1 ext2 file system, RW SDRAM memory
- RW file system
- the idea behind it was the genromfs tool (generated from a directory and the files in it a ROMFS image), which is used for ROMFS generation on the development system to port to the target system. This makes it possible to generate a ROMFS from directories on the target system so that it can also be programmed into the flash memory and read-connected.
- the files modified and created on the target system and in the RAMFS directory (which can also be a copy of the previous user flash content) are generated with the genromfs tool to a ROMFS image. It also required a minimal text editor for the target system to be ported. This ROMFS image is then also programmed on the target system with the flash memory programming program (flashw) into the user flash partition (Fig. 48).
- the channel simulator For the use in the channel simulator actually only the network is used, since over telnet a connection can be established to the RTS DSP board, which needed Change (usually parameters of the communication programs) of the mirrored flash memory can be made and the boat script for automatic call after the reset is adjusted. Thereafter, the ROMFS image is again generated via telnet and programmed into the flash memory. Basically, the programming of each partition of the flash memory is possible, whereby an update of the uClinux image can be done independently of the development system. For this, only the PC of the channel simulator GUI establishes a telnet connection to the target system, a network directory (eg NFS server on the PC), with the new uClinux image, integrated and this image, into the corresponding flash memory partition, from the target system programmed.
- a network directory eg NFS server on the PC
- the subject matter underlying the patent application can also be used as a channel sounder. It is a measuring device for determining the transmission characteristic of a real radio channel.
- a transmitting antenna sends out a short signal of known shape (amplitude variation, phase), which is received by a measuring antenna at another location.
- the received signal is digitized and can be used for further analysis.
- the subject invention is in an excellent manner.
- the generation of the transmission signal, the reception, the A / D conversion and the storage of the reception data can be realized with the invention.
- the FPGA reads in the data and forwards it to the memory of the DSP.
- the data can be processed or forwarded to a PC for processing.
- a particular advantage is obtained when the signals of several mutually offset measuring antennas with several boards recorded in parallel. This gives a spatial and temporal image of the signal distribution under exactly identical boundary conditions.
- the signal processing boards of the subject invention can be used to evaluate detector signals of a quantum cryptographic system to store and pass on.
- the detectors which detect quanta with different polarizations, have the disadvantage of a high error rate. This means that as soon as one detector trips, the others will also trip after a short while.
- a detection event If a detection event occurs, it is evaluated using a temporal window function.
- the timing of the window is communicated to the system via a separate conventional optical channel (synchronization channel). If the rising edge of the signal lies within the window, the signal is stored as activation. If it is outside the time window, it is detected as a faulty trip.
- the width and the delay of this window can be adjusted in increments of 1.25 ns. This exact resolution is achieved by using the positive and negative edges of the clock for window generation. In addition, the simultaneous occurrence of several detection events within a window is also checked, and the result is stored.
- Another important parameter of the detection logic is the number of individual detection signals within a defined interval. For this purpose, one uses a counter which detects the signal edges of the individual detector signals, counts and is reset to zero after a settable time interval. As soon as an event occurs on a detector, the signal is fed to the FPGA and subsequently evaluated and stored. For this purpose, it is forwarded to the DSP who performs these tasks. Using the interface between the DSP and the CPU, the information can be transferred to a GUI via an Ethernet link (Fig. 49).
- a further possible use beyond the actual purpose for the subject of the invention underlying the patent application is to provide a plurality of arbitrarily shapeable test signals to the LVDS sockets.
- This provides a universal, scalable, modular test system capable of testing systems that require a large number of test signals with different protocols.
- test signals could be generated which rarely occur in real operation and so e.g.
- a synchronized test signal can be made available by means of the graphical user interface and via the device-internal synchronization (CLK-START, LOCK and TRIG signals).
- Each signal processing board has a total of 78 usable test pins, which are directly connected to the existing on the boards FPGA.
- this FPGA it is now possible to generate any desired signal sequences or protocols, on the other hand, because of the parallel processing in the FPGA, the test signals can also be output in parallel.
- the modularity of the test system according to the invention can be extended to larger units than 1 board by interconnecting several racks to an even larger synchronous test system.
- the scalable test system shown here is able to replace the signals of the transducers. Thus, it is easily possible to simulate a variety of scenarios and operating conditions that occur during a driving operation.
- Loading the I / O data and controlling the test system happens through the Fast Ethernet interfaces of each board.
- the GUI controls the test system via these connections.
- a rack can be expanded up to a maximum of 776 test signals, whereby 624 work synchronously, 152 asynchronously (Fig. 50).
- the RTS-DSP board can also be used for image processing. This feature is provided by the two 26-pin sockets on the front panel. These can be configured as a LVDS connection to the board as a Camera Link. This allows either camera modules with a Camera Link output, or the RTS-DSP board as a video generator can be used. In the first case, the RTS-DSP board functions as a frame grabber and may also include other algorithms for image processing in the FPGA and the DSP. Relevant information can then be passed via Ethernet to a computer or other RTS-DSP boards.
- generator configuration data can be transmitted to the RTS-DSP board via Ethernet from a PC, then the signals to be output in the DSP or FPGA can be calculated on the board and output to further boards or external frame grabbers via the Camera Link (Fig. 51, 52).
- FIG. 54 shows a transmission frame in GSM mobile radio
- FIG. FIG. 55 shows a transmission channel between GSM transmitter and GSM receiver
- FIG. Fig. 56 shows an example of a radio transmission channel in the urban area
- Fig. 57 is a block diagram of the channel simulator according to the invention
- Fig. 58 is a block diagram of N channel simulators coupled together
- Fig. 59 is a block diagram of a channel simulator constructed of integrated circuits
- Fig. 60 shows the structure of a COST structure
- FIG. 61 is a symbolic representation of macro, micro and picocells in a GSM network
- FIG. 62 shows the principle of a geometrically stochastic channel model GSCM
- FIG. 63 shows the geometrically stochastic channel model in a Cartesian coordinate system
- Fig. 65 shows an Azimuth Delay Power Spectrum ADPS
- Fig. 66 shows a power delay profile and an azimuth power spectrum
- Fig. 67 is a diagram for determining scatterer positions
- FIG. 68 is a diagram of the simulation accuracy as a function of the scatterer number
- FIG. Fig. 69 is a Jacobian "bathtub" diagram of a Jacobian transform
- Fig. 70 is a plot of the amplitude probability density distribution in a Rayleigh process
- 71 is a diagram of the phase probability density distribution in a Rayleigh process
- FIG. 72 is a sum vector diagram of multipath components
- FIG. 73 shows a representation of the phase shift of a transmission channel caused by the high-frequency carrier
- FIG. 73 shows a representation of the phase shift of a transmission channel caused by the high-frequency carrier
- 74 shows a diagram of the average attenuation between transmitting station and receiving station as a function of the distance from each other, according to the COST-Hata model and the COST-Walfish-Ikegami model, respectively;
- Fig. 75 is a flow chart of the basic structure of the simulation algorithm
- Fig. 76 is a flow chart of the flow of the channel simulation
- Fig. 77 is a diagram showing example values of CAS, CDS and shadowing;
- the invention is explained below with reference to a mobile radio system, such as GSM, but is not limited thereto.
- Other applications of the invention relate to the development and research of intelligent antenna systems, in particular the testing of algorithms for changing the directional characteristic of such intelligent antenna systems.
- Other applications include wireless microphone systems and wireless networks.
- the invention offers the possibility of providing an image of a radio transmission channel between a radio transmitter and a radio receiver, so that expensive and time-consuming field tests can be dispensed with in favor of laboratory tests, wherein the simulated radio transmission channel can be reproduced at any time.
- the channel simulator according to the invention in the laboratory any interference and interference of the radio transmission channel can be adjusted by environmental conditions or imperfect radio equipment.
- the channel simulator according to the invention operates in real time, wherein also the parameters of the simulated radio transmission channel can be changed in real time.
- the channel simulator can be used for long-term measurements and individual data segments can be selected and analyzed during the simulation.
- the channel simulator may be equipped with a noise source to emulate channel noise, where the noise source may be based on a pseudo-random generator.
- GSM Global System for Mobile communications
- TDMA Time Division Multiplex Access
- FDMA Frequency Division Multiplex Access
- the chronological sequence is structured in so-called time frames, whereby each individual frame is assigned to a specific discussion participant. Each conversation is thus assigned a frequency and a time frame.
- GSM uses "gaussian minimum shift keying" (GMSK), which is a phase modulation technique with phase-in information, with channel bandwidth being 20OkHz due to modulation with Gauss pulses GSM can be considered as a narrowband system
- GMSK Gaussian minimum shift keying
- the simulation of transmission over a narrowband channel is much simpler than that of a wideband channel, simply because different frequencies have different transmission behavior. Obstacles like houses or mountains reflect different frequencies differently. Therefore, it can be assumed that all spectral components of a signal with a narrow bandwidth of 20OkHz are reflected equally.
- the carrier frequencies are arranged in a frequency range of 20OkHz, for example, in GSM900 in the downlink direction (ie when transmitting from a base station to a mobile station) between 935MHz and 960MHz 124 carrier frequencies are accommodated.
- Each of these channels is in turn divided into eight timeslots (TS), which is referred to as TDMA (Time Division Multiple Access), with a timeslot of a transmission channel being provided for each GSM subscriber.
- TS timeslots
- a timeslot again consists of 156.25 bits, which have different functions.
- Fig. 1 shows a GSM transmission frame consisting of eight time slots, wherein the Timeslot No. 2 shown in detail has a burst as it occurs in an information (voice) transmission
- Fig. 55 shows a block diagram of a GSM radio communication system comprising a GSM transmitter emitting broadcast signals, a transmission channel in which the broadcast signals are transmitted, and a GSM receiver receiving the transmitted signals.
- the GSM transmitter has a channel coder, which adds further redundant bits for error detection and error correction to the data stream to be transmitted in order to be able to correct disturbances in the transmission channel which lead to incorrectly transmitted bits at the receiver. Cyclic codes with error detection and correction are used.
- the bit stream at the output of the channel coder is modulated with the aid of the GMSK modulator, thereby generating signals whose information content lies in the phase transitions.
- the GSM receiver has a GMSK demodulator to demodulate the received signals.
- SDMA Space Division Multiplex Access
- GSM Global System for Mobile communications
- the transmission channel for radio signal propagation The quality of each radio transmission is primarily influenced by the transmission channel between transmitter and receiver, in which the propagation of the radio waves takes place.
- An example of a transmission channel in the urban area will be explained with reference to FIG. 56.
- radio waves In free space, radio waves propagate in a straight line regardless of their frequency. In the atmosphere, however, various factors influence the propagation characteristic of radio waves. The attenuation of the radio waves is dependent on their frequency. While low frequencies are hardly attenuated by atmospheric disturbances, at higher frequencies above about 12 GHz a strong attenuation due to scattering and absorption already occurs in fog or rain.
- the transmitted signal can be divided into several components, so-called multipath components (MPC), which reach the receiver on different paths.
- MPC multipath propagation
- MPP multipath propagation
- the transmission pulse is thereby widened or smeared in time.
- individual signals are superimposed, which is referred to as intersymbol interference ISI.
- ISI intersymbol interference
- GSM must work up to runtime differences of up to 16 ⁇ s.
- the other effect of multipath propagation is fading.
- a moving mobile station as a receiver generally receives a multipath signal that is time and location dependent.
- the various signal components arriving at the receiver may therefore have traveled different distances.
- These signal components may also be one which has traveled a direct path between transmitter and receiver (line of sight).
- other signal components have a time delay and thus a phase shift.
- These are superimposed on the antenna of the mobile station and can give a higher reception level than that without multipath components (constructive interference).
- these reusable components can cancel each other out by destructive interference. In the latter case one speaks of fading (fading).
- this effect is noticeable by periodic level dips in the received signal, at intervals of the order of half the wavelength. Because these break-ins are fairly rapid (the time to traverse this line at medium speed is approximately in the millisecond range), this fading is also called fast fading (fast fading). Another name for this fading is Small Scale Fading.
- the fading breaks become smaller the stronger one of the components of the multipath signal is over the other components. In direct line of sight between the transmitter and receiver, there is usually such a dominant signal component.
- a transmission channel which has a dominant signal component, a Rice channel and accordingly one speaks of Rice'schem - Fading.
- slow fading Another type of level drop of the received signal is the slow fading (slow fading). This is caused by large-scale obstacles in the line of sight between transmitter and receiver. Due to these obstacles, transmission paths disappear and new ones are added. The absence of this direct signal component (line of sight) is noticeable in the received signal as attenuation. Such obstructions causing shadowing can remain effective for a mobile station in distances of up to 100 meters. The time to traverse this route at medium speed is approximately in the seconds range. This is very long compared to the period of level fades in fast fading, and accordingly this fading is called slow fading. Another name for this fading is Large Scale Fading.
- the present invention has become Channel simulator for simulating at least one transmission channel between transmitters emitting emitters and receivers of a digital or analog radio transmission system developed, which is shown in an embodiment in Fig. 57 in block diagram.
- the channel simulator includes a plurality of signal inputs to which input signals representing transmit signals may be applied.
- the channel simulator includes an input for analogue low frequency signals (analog NF in) which are converted from an input analog to digital converter (A / D) into digital signals; an input for analog high-frequency signals (analog RF in), which are converted in a mixer into analog baseband signals and then converted into digital signals in an analog-to-digital converter (A / D); an input for digital baseband signals (dig. BB in) which are fed to a Low Voltage Differential Signaling (LVDS) circuit.
- the digital or digitized input signals are fed to a signal processing unit, which consists of a Field Programmable Gate Array (FPGA) and a digital signal processor DSP.
- FPGA Field Programmable Gate Array
- DSP digital signal processor
- a control means controls the field programmable gate array FPGA and the digital signal processor DSP and forms an interface to a control unit.
- SYNCJn and SYNC_out lines are used to couple N channel simulators to multiply the number of channels that can be computed, as shown in the block diagram of Figure 58, where all N channel simulators are controlled by a GUI operator.
- the control means ⁇ C provide the signal processing unit FPGA, DSP with one or more channel parameters defined channel model (s) defining the signal shaping characteristics of at least one transmission channel to be simulated.
- the signal processing unit FPGA, DSP uses the channel parameters to model the set channel model as a geometric model with a plurality of scattering objects and applies this geometric model to form the input signal, thus producing an output signal in real-time calculation of the propagation curve of multipath components of the input signal. wherein the multipath components are caused by scattering, reflection and diffraction of the input signal at the scattering objects of the geometric model.
- the output of the output signal is either via a digital / analog converter (D / A) as a low-frequency analog signal (anal NF out), or via a digital / analog converter (D / A) and a mixer, the signal in the Highfrequen Scheme transformed, as an analog high frequency signal (anal. HF out), or directly via a differential signal circuit (LVDS) as a digital baseband signal (dig. BB out).
- the signal processing unit maps the geometric model of the channel model by multiplying the multipath components of the input signal and then superimposing the multipath components, additionally weighting the multipath components with weighting factors that correspond to certain physical effects of the transmission channel to be simulated give the channel parameters of the channel model.
- the simulation calculation can be done according to the following formula:
- H H L0S + ⁇ H NS (k) + ⁇ H F s (n, m) k-0 n, m
- N FS Number of scatterers in a "far scatter area"
- the signal processing unit has been realized with a hybrid architecture, the signal processor DSP for the entire geometric-statistical calculations and the determination of the weighting factors, and the ReId Programmable Gate Array (FPGA) to perform the multiplications of the multipath components with the weighting factors and the superimposition of the weighted reusable components.
- FIG. 1 A block diagram of a structure of the integrated circuit channel simulator according to the present invention is shown in FIG.
- the input and output signals and the power supply are routed via a 96-pin plug connector.
- the FPGA is the type Xilinx® Virtex 2, which forms its own channel engine and architecture highly parallel processing and is also reconfigurable.
- This block can be configured using the hardware programming language VHDL.
- VHDL hardware programming language
- Softmacros are parameterizable digital basic structures such as counters, memories, etc., which are integrated into the design in blocks.
- Hardmacros are fully functional blocks with internal placement and wiring.
- the Virtex 2 device used has forty 18bit x 18bit multipliers and as many RAM blocks as hard macros.
- the digital signal processor DSP is a type DSP TMS320C6416 from Texas Instruments®.
- the firmware of both the FPGA and DSP is in a 2 + 2MB flash memory connected to the FPGA and the DSP via an EMIF B data bus (External Memory Interface).
- the DSP is connected via another data bus EMIF A with two 32 MB SDRAM banks.
- the FPGA is connected via a 32-bit memory data bus to a flash memory, an expansion connector, an SDRAM and a Motorola® Coldfire MCF5272 interface block that controls various standardized interfaces, such as a serial RS232 interface, a USB interface and an ethernet interface.
- the entire simulation environment is controlled by a personal computer (PC), not shown.
- PC personal computer
- initialization data are transmitted from the PC to the DSP.
- the transmission signal is first generated and prepared.
- the transmission signal is modulated and the geometric calculations are performed.
- the modulated data is passed on to the FPGA in the next step, where the actual signal interference of the channel is performed.
- a noise generator is implemented in the channel engine to simulate the noise in the radio transmission channel, whereby the noise power can be varied
- the channel engine model enables continuous communication with the DSP.
- the Channel Engine input signal uses the complex transmit signal in the magnitude and phase representation.
- the calculated received signal should appear in the representation with real and imaginary part.
- a geometrically stochastic channel model GSCM is used in the channel simulator according to the invention, which achieves the best simulation results currently available.
- GSCM geometrically stochastic channel model
- transmission channel means the sum of all propagation conditions of the entire route between transmitter and receiver.
- the atmosphere (with snowfall, rain, %), urban planning (streets, buildings, %), landscapes (forests, mountains, %), vehicles, people, etc. contribute to the extent to the entire extent Wave propagation at. It is often necessary in mobile communication to investigate certain effects in isolation. Because of the high number of influencing factors, it is extremely difficult to study the influence of the effects as a whole. In connection with the establishment and standardization of mobile radio systems, the transmission characteristics of the mobile radio channel play an important role.
- the propagation of radio waves is heavily dependent on the often extremely different topographical and physical features of the environment.
- a standardized three-level structure of COST259 SWG2.1 has been defined, which is shown in FIG.
- the top level contains a first distinction of the cell types.
- some REs radio environment
- G for "general” to emphasize that an RE describes a whole class of propagation constellations with similar or typical features.
- level 2 of the model defines different typical environments (Radio Environments). Furthermore, for typical environments, a set of parameters will be defined that will feed into the simulations based on this model. These are divided into External, Local and Global parameters. External parameters are those that remain the same for all simulations, e.g. Frequency, average height of base stations (BS) and mobile stations (MS), average distance between MS and BS, etc. The global parameters primarily affect the probability density of the local parameters. The influence of the local parameters can be seen in level 3 of the COST model. Local parameters include information about the size, placement and electrical properties of objects (mountains, buildings, ...) that are defined in specific simulation scenarios. During simulation, these parameters influence the following simulation variables:
- FIG. 61 shows a GSM network with macro, micro and pico radio cells.
- Macrocells are radio cells capable of covering a large area. They are used only in areas with low call volumes. These cells have the largest radii with an extent between 1km and 10km and are thus designed exclusively for outdoor use.
- the antennas are mounted above the level of the surrounding building roofs.
- Microcells are used in areas where there is a large volume of calls, e.g. in cities.
- the transmission power of these radio cells is very low, they cover an area that roughly corresponds to the range of the base of a DECT in-house telephone. Due to the short range of these antennas within a city, for example, the transmission frequencies can be repeated very often, thus the GSM network can provide sufficient call channels in these areas.
- Pico cells are smaller than microcells. In most cases, they should only ensure supply within a building.
- the geometrically stochastic channel model is based on the exact geometrical determination of a stochastic, thus randomly generated virtual landscape whose transmission properties are described by means of probability density functions and power profiles.
- the propagation conditions of each RE are described by certain specified parameters, probability density function (PDF) and power profiles. Since they characterize the propagation conditions of the perfect RE, they are called global parameters (GP). These serve as key parameters and provide the necessary information for basic system design decisions about modulation technique, burst length, coding scheme, etc. Global parameters are obtained through comprehensive measurement campaigns.
- PDF probability density function
- GP global parameters
- the scatterers can be found.
- Each of these scatterers carries a signal multipath component MPC.
- Measurements have shown that scatterers occur predominantly in groups (clusters) within which the reflection points of several MPCs are so close to each other that they experience approximately the same propagation conditions and reach the receiver approximately at the same time.
- These local dispersions typically occur in the immediate vicinity of the mobile station MS, which are then distributed as Near Scatterer (NS) in the Near Cluster (NC). But also at distant objects reflections can occur.
- These areas are called Far Cluster (FC) and include Far Scatterer (FS) (Fig. 62).
- BS it is assumed in macro cells that the antenna is placed high enough so that no local reflections occur here and thus there is no near cluster for the BS.
- Fig. 62 shows the principle of a geometrically stochastic Kanai model GSCM.
- the base station BS is at the origin of the coordinate system and thus has the coordinates:
- h B s is determined depending on the RE and describes the central position of the BS antenna.
- BS can exist in the cell, but there can be several BS antennas.
- a mobile station can normally move freely in the area between the inner and outer borders. Their position is determined by the Cartesian coordinates:
- the channel simulation according to the invention is carried out by an exact calculation of the propagation curve of the individual multipath components of the signal, which are caused by scattering, reflection and diffraction on the according to the selected channel model and statistical models arranged scattering objects (scatterers, scatterers) (ray tracing).
- the individual physical effects of a channel are calculated, which are taken into account by a statistical distribution of scatterers and randomly distributed Visibility Regions. From this weight factors are calculated which subsume the individual physical effects of the channel.
- the simulation is performed by multiplying the multipath components, which were delayed in time relative to the baseband signal according to the propagation paths, with these weighting factors and by superimposing these weighted multipath components. This corresponds to the superposition of multipath components and the emergence of interference in reality.
- MPC multipath component
- the parameters NCI.min and m are global parameters.
- PPP (m) is used to calculate the number of additional clusters (Far Cluster).
- the Poisson Point Process generates a Poisson distribution from a uniform distribution
- the probability of clusters occurring is the same at any point within a certain range around the BS. In the model, they are therefore spread with an equal distribution in this area.
- the position of a cluster becomes its center determined, which may not exceed the maximum distance dmax to BS.
- the location of Far Clusters can not change because they represent fixed objects in the landscape.
- a scatterer may be at most RCI + dCI, i away from the BS. Since the position of a scatterer can now be determined exactly, the signal transit time ⁇ Si, j can be calculated exactly via this path, c is the speed of light.
- the index of scatterer-related parameters is only the letter "s" (eg, ⁇ s ).
- Grouping of MPCs could be determined by measuring Azimuth Delay Power Spectrum (ADPS). This shows which power and time delay the individual MPCs arrive from which azimuth direction, as shown in FIG. 65.
- ADPS Azimuth Delay Power Spectrum
- the ADPS Power Delay Profile
- APS Azimuth Power Spectrum
- the PDP shows with what power and time delay the MPCs reach the receiver.
- the APS shows from which azimuth direction the individual power spectra arrive (FIG. 66).
- the angle 0 ° corresponds here to the LOS direction to the transmitter or to the cluster.
- EPS Elevation Power Spectrum
- the geographical position of a scatterer can also be determined. If the signal delay of an MPC is known, then the Scatterer lie only on an ellipse around BS and MS, as shown in Fig. 67. The actual position is where a straight line at an azimuth angle ⁇ intersects the ellipse.
- the GSCM provides a weighting of the scatterers that should correspond to their "areas.” This weighting is dependent on the MS position and thus can be temporally altered by the MS moving around Part of this weighting is determined by the ADPS function Ps (FIG. ⁇ , ⁇ , ⁇ ), which determines the attenuation of the scatterers depending on their position in the cluster. This relationship characterizes a cluster and its properties depend on its position, the spreads (CDS, CAS, CES), the power and The ADPS formula is applied in macrocells for each individual cluster, averaging the exponential decaying shape in the PDP typical of the receiver, and averaging the Laplace shape of the APS and EPS real measurements.
- Ps FIG. ⁇ , ⁇ , ⁇
- the function causes MPCs with a longer runtime and paths that deviate further from the LOS direction to be attenuated more strongly. It was developed from the results of many measurement campaigns.
- ⁇ indicates the azimuth angle between the LOS component and the MPC from the receiver's point of view.
- ⁇ T the delay difference between the LOS path and the complete scatterer path (transmitter-scatter receiver) must be used - at least this was the result of the simulator's analysis.
- the LOS path in the case of a Far Cluster is the route BS Far Cluster Center MS. This setting may cause some MPCs to have a negative T. Since the formula is only designed for positive delays, the amplitude of the corresponding scatterers must be set to zero.
- a cluster is represented as a cylinder, but since there is no elevation in the GSCM, a cluster can be treated two-dimensionally as a circle in the base area.
- the APS shows the power of the MPCs as a function of their azimuth angles. It can be observed that most of the power of a cluster arrives at a relatively small angle. If, from the receiver's point of view, one considers all MPCs in the angle range between - ⁇ and + ⁇ , the cluster would have to be infinitely large. Similarly, the exponential nature of the PDP can cause a very weak power component to arrive at the receiver infinitely long, which is also caused by an infinitely large cluster radius.
- the cluster radius is made dependent on the delay and azimuth spread, which are a measure of the spread of the delays and the azimuth and are mainly used as input values for the calculation of PDP and APS.
- the radius depending on the distance of the cluster to the BS is indicated by the CAS
- cluster radius R ⁇ the larger of the two radii is used as cluster radius R ⁇ .
- the distribution of the scatterers in the cluster is determined stochastically and has a great influence on the appearance of the ADPS.
- Formulas (9) and (10) originate from COST259 and were designed for equal distribution of scatterers. Only in this way the PDP and the APS get their characteristic forms. If the probability of occurrence PDF and thus the density of the scatterers near the cluster center were different to the density at the cluster edge, the sum of all MPCs would distort the resulting ADPS.
- the equal distribution of the scatterers is done with
- r s is the distance of the scatterer from the cluster center and U (0,1) is an equal distribution between 0 and 1.
- the associated azimuth angles are subject to an equal distribution between 0 ° and 360 °.
- J (x s , y s ) The relationship between the scatterer position in the radio cell and its attenuation is described by the Jacobian transformation J (x s , y s ).
- This stochastic model in the GSCM tries to transform the probability of occurrence (PDF) of a scatterer depending on its "location” and to derive an attenuation value for it.
- PDF probability of occurrence
- J (x s , y s ) is actually not a damping but a gain factor , which assumes values below 0 dB for all scatterer positions.
- the distribution density of a VR type p c is derived from the statistical probability of occurrence and is with
- m is the average number of additional clusters and r c is the radius of the corresponding VR type.
- the cluster VRs are equally distributed on the area within the cut-off distance d co .
- Their number N V R, CI, I is determined so that the ratio of the area of all VRs to the total area within the cut-off distance equals the probability of cluster visibility. The result must be rounded up or down to the nearest integer value. The presence of visibility is given by V ⁇ .
- the number of LOS VRs is determined by
- large-scale fading describes changes in the average power of the received signal when a time- or space-varying multipath channel is used as the transmission medium. Normally these mean values remain nearly constant for short time or space distances.
- the main cause of the occurrence of large-scale fading is the change in channel characteristics that occurs when the MS moves into or out of a VR. This causal relationship has led to the term shadow fading, which is often used instead of large-scale fading.
- small-scale fading describes the rapid amplitude changes of a signal received from a time- or space-varying multipath channel. These rapid changes can already occur with MS movements on the order of a wavelength ⁇ . They are caused by constructive or destructive superimposition of the time-delayed and with different phase angles incoming MPCs. Small scale fading is often referred to as fast fading or short term fading.
- the investigations of small-scale fading are mostly based on measurement campaigns, and many studies in recent years have also obtained their results through simulations for deterministic propagation prediction.
- the most widely used model for small-scale fading is the assumption of a Rayleigh fading amplitude.
- the Rayleigh distribution arises in an amplitude distribution of a sum a of a large number of uncorrelated rotating vectors with the same amplitude and uniformly distributed phase.
- Figures 70 and 70 show the Probability Density Function PDF for amplitude and phase for the Rayleigh process. This shows the probability with which a certain amplitude a (or phase ⁇ ) can occur.
- the received signal a (t) consists of the superposition of a plurality of uncorrelated MPCs (with approximately the same amplitudes)
- the probability of occurrence of the time-dependent amplitude ⁇ a (t) ⁇ can be described as a Rayleigh distribution.
- the phase of the resulting vector ⁇ fi (a (t)) is uniformly distributed in the interval [-T ⁇ , ⁇ ].
- the variance of the fading process 2 ⁇ 2 corresponds to the signal power and can therefore be determined from the result of the large-scale fading model.
- the Rayleigh distribution can be decomposed into uncorrelated Gaussian distributions (mean 0) for the real and imaginary parts of the vector a, which is advantageous for simulation purposes (see Fig. 72).
- the fading amplitude is modeled by the Rice distribution
- This distribution describes the amplitude distribution of the resulting vector a, which consists of the sum of one dominant vector and a plurality of much smaller uncorrelated vectors of equal amplitude and equally distributed phase.
- the Rice fading can therefore be interpreted as a superposition of a dominant component (BS-MS-LOS) with a Rayleigh fading signal.
- the parameter / 0 (x) in formula (22) is the modified Oth order Bessel function
- an MPC Due to the relative movement of the MS to the transmitter, an MPC experiences a frequency shift.
- a wave spreads in air at the speed of light c. If (in the downlink) the MS moves towards the BS at a speed v res , it "sees" the propagation velocity as the sum v res + c and a wavelength ⁇ is correspondingly faster, so the wave does not appear on the moving MS the transmission frequency f c , but shifted by the Doppler frequency f D.
- Pathloss The power losses of a radio transmission path in free space are described by the Pathloss (Pathloss). This is calculated using the COST-Hata model (for GRA and GHT) and the COST-Walfish-Ikegami model (for GTU and GBU). Both methods yield the average attenuation between transmit and receive signal as a function of the BS-MS distance, as shown in the graph of FIG. 74. They were first described in COST231 and marginally for use in COST259 customized. It should be noted that these models are only valid for narrowband channel simulators.
- L MC becomes three correction values over the ratio between the detected attenuation of the clusters (Uos, Lws. Lps) and COST path loss L
- the Rice factor K 0 is also included in the calculation.
- the APS distinguishes between uplink and downlink, so for these two cases different path weight calculations are used in the channel model.
- the path weights represent the respective proportion of an MPC in the total attenuation of the channel. They are multiplied by the input signal delayed by the signal propagation delay, and the sum of all products added to the two noise levels yields the received signal.
- FIG. 75 outlines the basic structure of the simulation algorithm for an antenna.
- the calculation of the line-of-sight component (LOS), the near-scatterer components (NS) and the far-scatterer components (FS) must be done separately for each antenna to be simulated because each antenna is separated by the spatial distance has a different channel transfer function.
- LOS line-of-sight component
- NS near-scatterer components
- FS far-scatterer components
- the program starts depending on the user request with the initialization of the selected radio cell and the modulator.
- the geographical conditions of the channel are diced according to specified criteria and all start values of the variable parameters are calculated.
- the program enters a loop which has the task of calculating the path weights and the output signal of the channel for the period of one time slot.
- the modulator is called and creates the input signal.
- the small-scale update whose task is to update the fast-variable parameters responsible for the smail-scale fading and to determine the path weights for each time in the Ah of the symbol duration T.
- the output signal is last determined in the GSCM channel engine.
- the loop Whenever the loop is run anew, it must be checked if the MS has moved further than the distance of 10 wavelengths. In this case, after the modulator call, the large-scale update is performed to update the slowly changing channel parameters.
- the number of loop passes is determined by the user by specifying the number of TDMA frames to simulate, with the calculation of one frame being eight
- Initialization starts with a few queries to the user. He is asked for the input of the desired RE and the modulator, he also has the number of simulated TDMA frames, which of the eight time slots to be simulated and specify the sampling rate.
- This area of the simulation is realized as a loop, whereby a pass can simulate the output signal of the channel for the duration of one time slot.
- the already calculated path weights for the current time (beginning of the time slot) are used and the path weights for the time at the end of the time slot are determined.
- the size of the GSM / EDGE symbol duration T is interpolated between the associated path weights in a time frame.
- the output is generated for each of these times. Since the path weights at the end of the time slot are the same as those at the beginning of the following time slot, only the path weights at the end of the time slot must be determined in each subsequent loop cycle by Compute Channel. In detail, the following steps are performed in one pass.
- the Large Scale Update updates values that cause large-scale fading and therefore need to be re-determined only after a ten-wavelength MS path has been traveled. Thus, this part of the program is not in every run of
- the MS will be positioned where it will be at the end of the current timeslot.
- interpolation is performed between the individual path weights at the beginning and at the end of the current time slot.
- the task of the Channel Engine is to determine the runtimes of the MPCs and, in combination with the path weights, to calculate the output signal. This program part is only called if the user has requested the calculation of the output signal for the current time slot.
- the signal propagation time of an MPC can be calculated on the basis of the distances.
- Each path weight is multiplied by the thus delayed component of the channel input signal (modulator baseband signal). Due to the in-phase addition of these products produces the output signal in a temporal grid of the symbol duration T.
- the MS position is either user-defined or random in the cell unless it is already given because of the Monte Carlo simulation. It sets the number of Far Clusters (FC) and distributes their positions within the cell area equally with a random function.
- FC Far Clusters
- the radii of the FC are only set in the setup and are not updated later. Thus, their diameter remains constant throughout the simulation and depends purely on the starting position of the MS and the associated spreads (CAS, CDS). Also, as described above, the number of FCs is not determined with a random number from a Poisson distribution, but only a single possible value from the distribution is used for the determination of the number of FCs, and thus the number of FCs remains during one whole channel simulation constant.
- tables are created with the correlating values of Shadowing, CAS and CDS, which are created based on the procedure described above.
- the distance between two table entries corresponds to a further movement of the MS by one meter, no matter in which direction. Crucial here is only the distance traveled since the start of the simulation. Because the three parameters are different for each cluster, the tables are created separately for each cluster.
- An example of the 3 tables of a cluster, shown as a diagram, can be seen in FIG. 77.
- the shadowing values of the individual clusters are interpolated linearly from the tables, irrespective of the actual MS position, but depending on their distance traveled, so that intermediate values from the 1-meter grid can be better approximated.
- the MS table must be started again when the MS is exceeded.
- the correction values L MC for calculating the path weights of LOS, NS and FS were determined in the Monte Carlo simulation in 128 equal steps between cell center and cell edge and are now taken from these tables depending on the BS-MS distance by linear interpolation.
- the cluster VRs and the LOS VRs are spread in the radio cell. It does not pay attention to whether the VRs overlap in certain places. At an overlap location, the visibility V is equal to the larger of the two values. It is also not considered whether a cluster VR is positioned in the area of the associated FC, ie the MS could be in the model in the cluster and still not see it. However, in the model analyzed, the shape of the VRs is generated quadratically, in contrast to the COST proposal, because of the savings in computational effort.
- the radii ⁇ CI L OS and r C ⁇ SF from the external parameters correspond in each case to the halved side length of a VR and the number is changed by the factor ⁇ / 4 with
- the size of the clusters is determined using CAS and CDS, and the near and far scatterers can be scattered in the clusters (see chapter 0). Their number per cluster, however, must be kept constant throughout a simulation in order to minimize the computational effort. However, more than twice as many scatterers are generated in each cluster than are needed. Because many MPCs are omitted from the calculation by a negative signal delay ⁇ r ⁇ 0, a portion of the path weights is calculated for the remainder and only the number of the strongest paths needed for the simulation is used.
- the antenna gain is interpolated for the LOS and each individual NS depending on the orientation of the antenna from the given antenna pattern.
- the antenna factor is only determined via the angle to the cluster center and therefore assumed to be identical, in order to minimize the computational effort.
- the influence of antennas is usually eliminated by setting the antenna gain in each direction to 1.
- the new scatterers will be spread in the entire cluster with an equal distribution. If the new cluster radius becomes larger than the old one, the new scatterers will be placed in the area of the new cluster that does not overlap with the old cluster.
- a scatterer that falls out of the NC is deleted from memory and thus does not reappear when the MS moves back into that area.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Mobile Radio Communication Systems (AREA)
- Radio Transmission System (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
Abstract
Description
Claims
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE112005002146T DE112005002146A5 (de) | 2004-09-06 | 2005-09-06 | Kanalsimulations- sowie Entwicklungsplattform und Verwendung derselben |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| ATA1491/2004 | 2004-09-06 | ||
| AT14912004A AT501645B1 (de) | 2004-09-06 | 2004-09-06 | Kanalsimulations- sowie entwicklungsplattform und verwendung derselben |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| WO2006026799A2 true WO2006026799A2 (de) | 2006-03-16 |
| WO2006026799A3 WO2006026799A3 (de) | 2006-06-15 |
Family
ID=35788193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/AT2005/000356 Ceased WO2006026799A2 (de) | 2004-09-06 | 2005-09-06 | Verfahren zur simulation eines mimo kanals |
Country Status (3)
| Country | Link |
|---|---|
| AT (1) | AT501645B1 (de) |
| DE (1) | DE112005002146A5 (de) |
| WO (1) | WO2006026799A2 (de) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008131814A1 (de) * | 2007-04-25 | 2008-11-06 | Rohde & Schwarz Gmbh & Co. Kg | Messgerät mit serieller digitaler schnittstelle |
| DE102008013011A1 (de) * | 2008-03-07 | 2009-09-10 | Rohde & Schwarz Gmbh & Co. Kg | Verfahren zur Generierung von Mehrantennensignalen |
| CN102122996A (zh) * | 2011-03-11 | 2011-07-13 | 电信科学技术研究院 | 模拟射频信号生成方法及系统 |
| CN102223192A (zh) * | 2011-06-15 | 2011-10-19 | 北京交通大学 | 一种高速铁路复合小尺度无线信道模型构建的方法和装置 |
| CN102571171A (zh) * | 2012-01-10 | 2012-07-11 | 西安交通大学 | 一种多输入多输出无线通信系统信道模型的硬件实现方法 |
| CN103888196A (zh) * | 2013-04-25 | 2014-06-25 | 中国科学院上海微系统与信息技术研究所 | 双向移动通信环境模拟系统 |
| CN103984836A (zh) * | 2014-05-30 | 2014-08-13 | 电子科技大学 | 一种3d-mimo天线极化增益的联合存储方法 |
| CN105743590A (zh) * | 2016-03-04 | 2016-07-06 | 清华大学 | 多天线系统的Nakagami信道m参数估测系统 |
| AT519270A1 (de) * | 2016-11-11 | 2018-05-15 | Ait Austrian Inst Tech Gmbh | Verfahren zur Emulation eines Funkkanals |
| CN108418611A (zh) * | 2018-03-08 | 2018-08-17 | 成都坤恒顺维科技股份有限公司 | 一种大规模多输入输出无线信道仿真仪 |
| US10397811B2 (en) | 2016-10-14 | 2019-08-27 | At&T Intellectual Property I, L.P. | Wireless channel sounder with fast measurement speed and wide dynamic range |
| CN113010106A (zh) * | 2021-02-25 | 2021-06-22 | 北京遥测技术研究所 | 一种基于fpga的总线复用型flash读写系统 |
| CN113595665A (zh) * | 2020-04-30 | 2021-11-02 | 华为技术有限公司 | 获取模拟信道的方法和装置 |
| CN114124262A (zh) * | 2021-11-25 | 2022-03-01 | 江苏科技大学 | 基于智能反射面的宽带高空平台信道模型建立方法 |
| CN114268982A (zh) * | 2021-11-30 | 2022-04-01 | 重庆长安汽车股份有限公司 | 一种车载移动通信终端测试系统 |
| CN116048514A (zh) * | 2023-02-06 | 2023-05-02 | 昆易电子科技(上海)有限公司 | 模型处理方法、测试方法、装置、终端设备及存储介质 |
| EP4198750A1 (de) * | 2021-12-17 | 2023-06-21 | dSPACE GmbH | Verfahren zur datenkommunikation zwischen teilbereichen eines fpgas |
| CN117707654A (zh) * | 2024-02-06 | 2024-03-15 | 芯瑞微(上海)电子科技有限公司 | 一种多物理场核心工业仿真处理软件信号通道继承方法 |
| US12158849B2 (en) | 2021-12-17 | 2024-12-03 | Dspace Gmbh | Method for data communication between subregions of an FPGA |
| CN119995742A (zh) * | 2025-04-15 | 2025-05-13 | 集美大学 | 信号多径效应模拟系统 |
| CN120196005A (zh) * | 2025-03-07 | 2025-06-24 | 成都立思方信息技术有限公司 | 一种可灵活配置的实时高精度雷达回波仿真系统 |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7154959B2 (en) * | 2001-08-29 | 2006-12-26 | Intel Corporation | System and method for emulating a multiple input, multiple output transmission channel |
| FI114596B (fi) * | 2001-11-09 | 2004-11-15 | Elektrobit Oy | Menetelmä ja laite radiokanavan simuloimiseksi |
| EP1447927A1 (de) * | 2003-02-17 | 2004-08-18 | France Telecom | Vorrichtung und Verfahren zur Signalverarbeitung |
-
2004
- 2004-09-06 AT AT14912004A patent/AT501645B1/de not_active IP Right Cessation
-
2005
- 2005-09-06 DE DE112005002146T patent/DE112005002146A5/de not_active Withdrawn
- 2005-09-06 WO PCT/AT2005/000356 patent/WO2006026799A2/de not_active Ceased
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008131814A1 (de) * | 2007-04-25 | 2008-11-06 | Rohde & Schwarz Gmbh & Co. Kg | Messgerät mit serieller digitaler schnittstelle |
| US8548767B2 (en) | 2007-04-25 | 2013-10-01 | Rohde & Schwarz Gmbh & Co. Kg | Measuring device with serial digital interface |
| US8711962B2 (en) | 2008-03-07 | 2014-04-29 | Rohde & Schwarz Gmbh & Co. Kg | Method for generating multi-antenna signals |
| DE102008013011A1 (de) * | 2008-03-07 | 2009-09-10 | Rohde & Schwarz Gmbh & Co. Kg | Verfahren zur Generierung von Mehrantennensignalen |
| CN102122996A (zh) * | 2011-03-11 | 2011-07-13 | 电信科学技术研究院 | 模拟射频信号生成方法及系统 |
| CN102122996B (zh) * | 2011-03-11 | 2013-10-16 | 电信科学技术研究院 | 模拟射频信号生成方法及系统 |
| CN102223192A (zh) * | 2011-06-15 | 2011-10-19 | 北京交通大学 | 一种高速铁路复合小尺度无线信道模型构建的方法和装置 |
| CN102571171A (zh) * | 2012-01-10 | 2012-07-11 | 西安交通大学 | 一种多输入多输出无线通信系统信道模型的硬件实现方法 |
| CN102571171B (zh) * | 2012-01-10 | 2014-04-23 | 西安交通大学 | 一种多输入多输出无线通信系统信道模型的硬件实现方法 |
| CN103888196A (zh) * | 2013-04-25 | 2014-06-25 | 中国科学院上海微系统与信息技术研究所 | 双向移动通信环境模拟系统 |
| CN103888196B (zh) * | 2013-04-25 | 2015-09-23 | 中国科学院上海微系统与信息技术研究所 | 双向移动通信环境模拟系统 |
| CN103984836A (zh) * | 2014-05-30 | 2014-08-13 | 电子科技大学 | 一种3d-mimo天线极化增益的联合存储方法 |
| CN103984836B (zh) * | 2014-05-30 | 2017-06-23 | 电子科技大学 | 一种3d‑mimo天线极化增益的联合存储方法 |
| CN105743590A (zh) * | 2016-03-04 | 2016-07-06 | 清华大学 | 多天线系统的Nakagami信道m参数估测系统 |
| US10945142B2 (en) | 2016-10-14 | 2021-03-09 | At&T Intellectual Property I, L.P. | Wireless channel sounder with fast measurement speed and wide dynamic range |
| US10397811B2 (en) | 2016-10-14 | 2019-08-27 | At&T Intellectual Property I, L.P. | Wireless channel sounder with fast measurement speed and wide dynamic range |
| AT519270A1 (de) * | 2016-11-11 | 2018-05-15 | Ait Austrian Inst Tech Gmbh | Verfahren zur Emulation eines Funkkanals |
| AT519270B1 (de) * | 2016-11-11 | 2018-07-15 | Ait Austrian Inst Tech Gmbh | Verfahren zur Emulation eines Funkkanals |
| CN108418611A (zh) * | 2018-03-08 | 2018-08-17 | 成都坤恒顺维科技股份有限公司 | 一种大规模多输入输出无线信道仿真仪 |
| CN113595665A (zh) * | 2020-04-30 | 2021-11-02 | 华为技术有限公司 | 获取模拟信道的方法和装置 |
| CN113595665B (zh) * | 2020-04-30 | 2022-09-23 | 华为技术有限公司 | 获取模拟信道的方法和装置 |
| CN113010106B (zh) * | 2021-02-25 | 2024-05-14 | 北京遥测技术研究所 | 一种基于fpga的总线复用型flash读写系统 |
| CN113010106A (zh) * | 2021-02-25 | 2021-06-22 | 北京遥测技术研究所 | 一种基于fpga的总线复用型flash读写系统 |
| CN114124262B (zh) * | 2021-11-25 | 2023-10-24 | 江苏科技大学 | 基于智能反射面的宽带高空平台信道模型建立方法 |
| CN114124262A (zh) * | 2021-11-25 | 2022-03-01 | 江苏科技大学 | 基于智能反射面的宽带高空平台信道模型建立方法 |
| CN114268982A (zh) * | 2021-11-30 | 2022-04-01 | 重庆长安汽车股份有限公司 | 一种车载移动通信终端测试系统 |
| CN114268982B (zh) * | 2021-11-30 | 2023-06-27 | 重庆长安汽车股份有限公司 | 一种车载移动通信终端测试系统 |
| EP4198750A1 (de) * | 2021-12-17 | 2023-06-21 | dSPACE GmbH | Verfahren zur datenkommunikation zwischen teilbereichen eines fpgas |
| US12158849B2 (en) | 2021-12-17 | 2024-12-03 | Dspace Gmbh | Method for data communication between subregions of an FPGA |
| CN116048514A (zh) * | 2023-02-06 | 2023-05-02 | 昆易电子科技(上海)有限公司 | 模型处理方法、测试方法、装置、终端设备及存储介质 |
| CN117707654A (zh) * | 2024-02-06 | 2024-03-15 | 芯瑞微(上海)电子科技有限公司 | 一种多物理场核心工业仿真处理软件信号通道继承方法 |
| CN117707654B (zh) * | 2024-02-06 | 2024-05-03 | 芯瑞微(上海)电子科技有限公司 | 一种多物理场核心工业仿真处理软件信号通道继承方法 |
| CN120196005A (zh) * | 2025-03-07 | 2025-06-24 | 成都立思方信息技术有限公司 | 一种可灵活配置的实时高精度雷达回波仿真系统 |
| CN119995742A (zh) * | 2025-04-15 | 2025-05-13 | 集美大学 | 信号多径效应模拟系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2006026799A3 (de) | 2006-06-15 |
| DE112005002146A5 (de) | 2007-09-06 |
| AT501645B1 (de) | 2007-11-15 |
| AT501645A1 (de) | 2006-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2006026799A2 (de) | Verfahren zur simulation eines mimo kanals | |
| DE69522918T2 (de) | Verfahren und vorrichtung zur simulation der von teilnehmern empfangenen interferenz in einem spreizspektrum-kommunikationssystem | |
| DE60313741T2 (de) | Mobilfunknetzwerksimulator | |
| DE69535615T2 (de) | Verteilsystem für drahtloses telefon mit raum- und zeitdiversity | |
| DE60119121T2 (de) | Emulation von übertragungsverzerrung | |
| DE69718665T2 (de) | Positionsbestimmung mittels satelliten auf niedriger umlaufbahn | |
| DE69535410T2 (de) | Verfahren und anordnung zum testen von digitalen übertragungskanälen mit variablen oder festen datenraten | |
| De et al. | Design considerations for a multihop wireless network testbed | |
| DE102014102406B4 (de) | Fading-Simulator und Fading-Simulationsverfahren | |
| DE69811399T2 (de) | Simulationsverfahren für funkfrequenzszenario in mobilfunksystem und testsytem zur durchführung des verfahrens | |
| KR20170020491A (ko) | 무선 디바이스를 테스트하기 위한 자연적 무선 환경의 가상화 | |
| DE102004002145B4 (de) | Verfahren und Vorrichtung zur Anpassung eines Funknetzmodells an die Bedingungen eines realen Funknetzes | |
| AT500923B1 (de) | Kanalsimulationsverfahren | |
| DE3145992A1 (de) | Digitale mobilfunkanlage hoher kapazitaet | |
| WO2003094538A3 (en) | Methods, apparatus, and systems for simulation of mixed traffic in a wireless network | |
| DE112009004412T5 (de) | Dynamischer hf-matrix-emulator | |
| DE2921780B2 (de) | Verfahren und Schaltungsanordnung zur Übertragung von Daten über eine Zweidrahtleitung | |
| Eyceoz et al. | Prediction of fast fading parameters by resolving the interference pattern | |
| DE69833187T2 (de) | Sehr bandbreiteneffiziente kommunikation | |
| DE102013217899A1 (de) | Audiowiedergabesystem und -verfahren, die in handgehaltenem elektronischem Gerät verwendet werden | |
| DE4030825C2 (de) | Verfahren zum Simulieren der Auswirkung alternativer Antennenrichtdiagramme auf die Abdeckungs- und Interferenzbilder in einem Mobilfunksystem | |
| DE112006000207T5 (de) | Verfahren und Vorrichtung zur Durchführung von Kanalsimulation | |
| EP3094019A1 (de) | Hf-testgerät und verfahren zum testen von d2d-fähigen mobilfunkendgeräten | |
| Burkhardt et al. | MIMOSA–a dual approach to detailed land mobile satellite channel modeling | |
| DE69835002T2 (de) | Funksystem und verbindungsaufbauverfahren |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU LV MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| WWE | Wipo information: entry into national phase |
Ref document number: 1120050021467 Country of ref document: DE |
|
| REF | Corresponds to |
Ref document number: 112005002146 Country of ref document: DE Date of ref document: 20070906 Kind code of ref document: P |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 05777706 Country of ref document: EP Kind code of ref document: A2 |












