BRC Series. BRK Series. LL Series. LS Series. C-Band LNA. LC Series. LRC Series. LXA Series. LRX Series. BRK Series. LL Series. LS Series. C-Band LNA. LC Series. LRC Series. LXA Series. LRX Series. LK Series. Furthermore, the individual antennas within a group may be differentiated by space, polarization, beam pattern, or orientation to achieve spatial, polarization, or angular decorrelation.
The embodiments described herein are applicable to all time-division duplex TDD systems, and are particularly suited to those that incorporate orthogonal frequency division multiplexing OFDM. Assuming a relatively static propagation environment, TDD systems exhibit channel reciprocity since both uplink and downlink transmissions take place on the same frequency.
The algorithms described in this invention calculate and apply spatial weights to a signal received from a particular transmission source hereinafter called uplink transmissions and then, if a set of rules testing channel reciprocity permit, apply the complex conjugate of the same weights on one or more downlink transmissions to the same transmission source, which is now expecting to receive downlink signals.
For downlink transmissions, each spatial processing channel is also multiplied by complex calibration coefficients that compensate for the relative amplitude and phase errors between spatial channels due to receiver and transmitter implementation imperfections.
The invention includes database algorithms that parse the spatial weights produced from multiple uplink sources operating from different locations within the bounds of a TDD framework. Each antenna element receives one or more delayed version s of the signal in addition to interference.
The correlation of signals between antenna elements is a function of antenna geometry and multipath scattering in the vicinity of receiving antennas. The invention applies spatial weights in the frequency domain so as to reduce the bandwidth of signals that are processed by an adaptive algorithm, thereby preserving degrees of freedom for interference rejection and diversity.
The invention describes unique frequency domain algorithms that exhibit fast convergence and low processing overhead and provides a specific embodiment of the invention for an IEEE Std The embodiment includes automatic gain control, synchronization, and diversity switch selection algorithms for this application. In particular, FIG. A high-speed backhaul link is used to carry information between the base station and a core network The transmission media for backhaul link is not specified as it may be fiber, coaxial cable, wireless, or some other future transmission media.
The invention is primarily embodied in base station and the system in FIG. Information is exchanged between the base station and remote devices through over wireless media through The invention imposes no restrictions on the number of remote devices, or their geographic orientation relative to the base station.
Diagram illustrates the sharing of a single frequency between a plurality of remote devices using the TDD technique. Transmissions from the base station to remote device s are hereinafter referred to as downlink transmissions. Transmissions from a remote device to the base station are hereinafter referred to as uplink transmissions. The interface to the core network in FIG. The upper layer protocol processor interfaces to the spatially aware media access control shown as media access c'ntrl and hereinafter abbreviated as SP-MAC.
The individual antennas are located on the right side of FIG. The description focuses on the following blocks in transmit mode: , , , through , through , and through In the receive mode, the description focuses on blocks: , , , through , through , and through Prior to operation, a calibration procedure is performed to measure the differential amplitude and phase of the receive and transmit hardware implementations for spatial channels 1 through M.
A set of frequency domain complex correction values is calculated and stored for eventual application in transmit spatial processing shown as tr'nsmt spatial proc'sng. During initialization, the radio frequency switches RF SW through shown as rf sw are initialized in a known state and the RF transceivers through shown as rf tr'nscvr are set to the receive mode. The RF transceivers through amplify and convert the signals from a radio frequency RF to an intermediate frequency IF or baseband frequency.
The analog to digital converters ADC through sample, quantize, and code the IF or baseband signals into digital format. To effectively process the signal using spatial algorithms, the receive automatic gain control AGC and synchronization function sets the gain of transceivers through so that their outputs are kept within the dynamic range of ADCs through , respectively.
This functional block also corrects for timing and frequency offsets. The lines connected to represent a multiplicity of input and output signals used for timing, gain, and frequency control. The digital downconverters DDC through mix the respective signals to baseband 0 Hz and, optionally, digitally filter and reduce the sample rate of each spatial channel. The digital mix frequency is generated by Alternately, the interface from RF transceivers through may be at baseband 0 Hz , thereby increasing the number of ADCs by a factor of two and changing the functionality of DDCs through to only offset the receive frequency around 0 Hz, as opposed to around a receive IF , if required.
The choice of using IF or baseband sampling may be determined by cost, performance, and availability of implementation options.
The AGC algorithm calculates a gain value for RF transceivers through based on ADC outputs through , DDC outputs through , and, optionally, received signal strength indications from RF transceivers through The cyclic extension is removed from the time domain signal by through shown as r'mv cy'lc ex'tsn under the control of The signals are converted to parallel format by through shown as serial to p'rll and converted into the frequency domain by N-point discrete Fourier transforms DFT through After the spatial channels have been transformed into the frequency domain, receive spatial processing calculates and applies one or more complex spatial weights that have been calculated to minimum mean-squared error MMSE or another criteria.
The weight vector s and a time-stamp are passed to the spatially aware media access control SA-MAC SA-MAC correlates and stores the spatial information, or some transform thereof, as spatial weights with the decoded source address of the device that transmitted the uplink signal. The output of is then optionally processed by an N-point inverse discrete Fourier transform IDFT based on whether the demodulating process is applied to time domain or frequency domain signals.
The resultant signal is then processed by the remaining blocks in the system , , , , , respectively shown as p'rll to serial, de-mod'lt, de-mux, ch'nl decode, source decode and format that perform parallel to serial conversion, demodulate, demultiplex, channel decode, source decode, and format functions, respectively, before being output to The output of demodulate is provided to receive spatial processing for decision-aided adaptation e.
De-interleaving, if included, is encapsulated in the channel decode function De-scrambling, if included, is encapsulated in the source decode function The SA-MAC parses the output of and determines if the information should be passed to upper layers for further processing. The receive processing sequence is now complete. In transmit mode, the information is received from upper network layers and parsed in the SA-MAC before being passed to format block The SA-MAC examines the destination address of the transmit information and determines if a set of spatial weights are available from a prior reception within a pre-determined timeout period.
If these, and potentially other tests, are true, then the SA-MAC transfers the stored spatial weights to Such other tests may include determination of whether the message is unicast, multicast or broadcast.
Such calculation may be based on a mathematical combination of a plurality of weights stored as with a multicast message and reference weights, or an omni-directional weight as with broadcast messages or messages where receive spatial information is either unavailable or unreliable. After formatting, the signal is processed by and shown as source encode and ch'nl encode.
Scrambling, if included, is encapsulated in the source encode function Interleaving, if included, is encapsulated in the channel encode function The multiplex function shown as mux accepts inputs from other encoding stages and also provides an insertion point for zero tones and training tones in OFDM applications. Multiplex also serves as the insertion point for test signals that are activated during system calibration. In OFDM systems, block shown as mod'lt maps data bits to symbols for data-carrying subcarriers and sets the training tones to the appropriate amplitude and phase.
For other modulation formats, the modulator uses appropriate modulating techniques. Block shown as serial to p'rll converts the signal to parallel format for processing by DFT where it is converted to the frequency domain DFT is bypassed for an OFDM application. Transmit spatial processing receives the frequency domain signal representation and applies complex calibration coefficients to each spatial channel.
The calibration coefficients are applied in the frequency domain to compensate for frequency-selective amplitude and phase errors.
Next, the spatial weights, or some transform thereof, passed to the block from SA-MAC are applied to the corrected frequency domain representation of the transmit signal. Similarly, the order of application of calibration coefficients and spatial weights to the frequency domain representation of the transmit signal may be juxtaposed. Furthermore, the receive weights may be transformed to transmit weights and calibration coefficients may be applied prior to storage and retrieval and application to the frequency domain transmitted signal.
The order of these processes is determined by SA-MAC response time requirements, architecture, and implementation. The transmit timing, frequency and gain control generates timing signals that may be used to delay the output of IDFTs through if alignment with other sources is required.
The lines connected to represent a multiplicity of input and output signals used for timing, frequency, and gain control. The output of transmit spatial processing is processed by IDFTs through that convert the signal to the time domain. The time domain signal is converted to serial format by through shown as p'rll to serial under the control of A cyclic extension is added in through shown as add cy'lc ex'tsn prior to mixing to a transmit IF by the digital upconverter DUC through The digital to analog conversions are performed by DACs through The converted signal is passed to RF transceivers through along with the control signals generated by Alternately, the interface to RF transceivers through may be at baseband 0 Hz , thereby increasing the number of DACs by a factor of two and changing the functionality of DUCs through to only offset the transmit frequency around 0 Hz, as opposed to around a transmit IF under the control of , if required.
The RF transceivers through are set to the transmit mode by signals originating from These transceivers amplify and convert the signals from baseband or an intermediate frequency IF to a radio frequency RF. The gain of these transceivers may be set based on implementation requirements related to linearity of the transceiver and modulation used, by regulatory requirements related to spectral emissions or maximum power output requirements, or for power control based on a closed loop power control using information sent by the receiver, channel reciprocity based on the strength of signals recently received from the intended destination, or other algorithms.
The radio frequency switches through are set to a state determined from a prior reception and passed from SA-MAC The wireless signal is simultaneously transmitted over M selected antennas from the set of antennas through The transmit processing sequence is now complete.
Referring now to FIG. The time index for frequency domain signals, nT DFT , is not shown for the sake of clarity.
0コメント