Abstract:
A vector processor is provided having an instruction set with a vector convolution function. The disclosed vector processor performs a convolution function between an input signal and a filter impulse response by obtaining a vector comprised of at least N1+N2-1 input samples; obtaining N2 time shifted versions of the vector (including a zero shifted version), wherein each time shifted version comprises Ni samples; and performing a weighted sum of the time shifted versions of the vector by a vector of Ni coefficients; and producing an output vector comprising one output value for each of the weighted sums. The vector processor performs the method, for example, in response to one or more vector convolution software instructions having a vector input. The vector can comprise a plurality of real or complex input samples and the filter impulse response can be expressed using a plurality of coefficients that are real or complex.
Abstract:
An low power vector summation apparatus is provided for using 2's complement arithmetic without the high switching activity of the prior art. In particular, the invention operates to exploit the sign-extension property of a 2's complement number. A reduced representation for 2's complement numbers is provided to avoid sign-extension and the switching of sign-extension bits. The maximum magnitude of a 2's complement number is detected and its reduced representation is dynamically generated to represent the signal. A constant error introduced by the reduced representation is also dynamically compensated.
Abstract:
A data structure, method and protocol wherein synchronization data indicative of a data frame delineation point is inserted within an inter-packet gap (IPG) proximate a data frame during transmission. Optionally, a cyclical redundancy check (CRC) length indicative data, pointer data, and other data is inserted within the IPG to further insure appropriate delineation of data frames within a data stream.
Abstract:
An arrangement is provided for using 2's complement arithmetic without the high switching activity of the prior art. In particular, the invention operates to exploit the sign-extension property of a 2's complement number. A reduced representation for 2's complement numbers is provided to avoid sign-extension and the switching of sign-extension bits. The maximum magnitude of a 2's complement number is detected and its reduced representation is dynamically generated to represent the signal. A constant error introduced by the reduced representation is also dynamically compensated.
Abstract:
A method and apparatus are disclosed for increasing the effective processing speed of a parallel decision-feedback equalizer (DFE) by combining block processing and look-ahead techniques in the selection (multiplexing) stage. The present invention extends a parallel DFE by using look-ahead techniques in the selection stage to precompute the effect of previous blocks on each subsequent block, and to thereby remove the serial output dependency. The parallel DFE includes a multiplexor tree structure that selects an appropriate output value for each block and precomputes the effect of previous blocks on each subsequent block. A multiplexing delay algorithm on the order of logN is employed to resolve the output dependency and thus speeds up parallel block processing DFEs. The disclosed DFE architecture can be combined with pipelining to completely eliminate the critical path problem. Pipelining reduces the required critical path timing to one multiplexing time. The disclosed multiplexor tree circuitry for the parallel DFE groups multiplexor blocks into groups of two, referred to as block pairs, and provides at least one multiplexor for each block, i, to select an output value, yi, from among the possible precomputed values. The output of each parallel block depends on the possible precomputed values generated by the look-ahead processors for the block, as well as the actual values that are ultimately selected for each previous block. In order to reduce the delay in obtaining each actual output value, the present invention assumes that each block contains each possible value, and carries the assumption through to all subsequent blocks. Thus, the number of multiplexors required to select from among the possible values grows according to N·logN, where N is the block number.
Abstract:
Convergence of blind fractionally spaced equalizers is improved, and misconvergence is corrected by training the equalizers to detect convergence of one adaptive filter, copying the tap weights of the converged adaptive filter to the other adaptive filters and shifting the tap weights of the other adaptive filters according to the expected phase difference between the respective filters. In a two-dimensional orthogonal modulation scheme the converged weights of a first filter are copied to a second filter and shifted .pi./2. For the two dimensional orthogonal modulation scheme, the probability of a proper convergence can be increased by choosing initial tap weights for the two adaptive filters with a 3.pi./4 phase difference.
Abstract:
A method and system for canonical channel estimation in the Long Term Evolution uplink where a multi-frequency signal is generated and then converted to frequency spectrum which is then convolved in the frequency domain with a truncated window function to obtain a time domain channel impulse response. The time domain channel impulse response can be then transformed to a frequency domain to produce a down sampled user channel response, which can be then linearly interpolated to provide a channel estimate for a plurality of subcarriers. Such an approach achieves channel estimation within Long Term Evolution at only canonical locations to reduce complexity without loss in channel entropy.
Abstract:
A method of applying an order N fast Hadamard transform (FHT) of a vector U using a mixed radix FHT in a receiver of a communication system, the N a positive integer, when receiving signals from a transmitter over a channel and generating the vector U. The method includes, in an FHT module of a decoder in the receiver, planning n stages of the mixed radix FHT, where the n is a positive integer, each stage defined by corresponding logic, decomposing the order N FHT into n low order FHTs, such that N=KnKn−1 . . . K1 and U=UKnKn−1 . . . K1, where the K is a positive integer, calculating, via the corresponding logic, each low order FHT at each stage, wherein input vectors of a subsequent stage are calculated in a proceeding stage, and reconstructing, by the decoder, calculated results of the each low order FHT to form an output vector output the decoder.
Abstract:
In one embodiment, the present invention is a method for performing incremental preamble detection in a wireless communication network. The method processes non-overlapping chunks of incoming antenna data, where each chunk is smaller than the preamble length, to detect the signature of the transmitted preamble. For each chunk processed, chips of the chunk are correlated with possible signatures employed by the wireless network to update a set of correlation profiles, each profile comprising a plurality of profile values. Further, an intermediate detection is performed by comparing the updated profile values to an intermediate threshold that is also updated for each chunk. Upon receiving the final chunk, the correlation profiles are updated, and a final preamble detection is made by comparing the updated profile values to a final threshold. Detections are performed on an incremental basis to meet latency requirements of the wireless network.
Abstract:
A method and system for parallel computation of a linear sequential circuit (LSC) based on a state transition matrix is disclosed herein. A multistep state transition matrix and a multistep output generation matrix can be pre-computed and stored in association with the linear sequential circuit. The multiple state transitions and the multiple output bits can be computed by multiplying the current input-state vector with a multistep next state transition matrix and a multistep output generation matrix, respectively. Multiple state transitions and multiple output bits can be generated in parallel in a single clock cycle based on the pre-computed state transition matrix and the output generation matrix utilizing a dot product in order to improve computational speed. Such a simple augmentation provides a flexible and inexpensive solution for high speedup linear sequential circuit computation with respect to a processor.