St-ericsson Sa

Ahmet Bastug

Pierre Demaj

Fabrizio Tomatis

Andrea Ancora

Giuseppe Montalbano

Eric Alliot

A symbol-level adaptation method to adapt at least one coefficient of an equalizer, the method comprising the steps of:

Symbol-level adaptation method for equalizer coefficients, memory, equalizer and receiver for implementing the method

The invention is directed towards a method and apparatus for estimating a noise variance in a receiver of a code multiplex telecommunications system with orthogonal spreading codes, wherein respective real and imaginary parts of a sum and a difference of two estimated symbols having substantially same amplitudes are determined, and then respective minimum energy components of the determined real and imaginary parts of the sum and difference are determined. The respective minimum energy components are finally combined to obtain the noise variance. Thereby, noise variance can be estimated simply based on addition and comparison operations, which reduces complexity and processing load.

Method and apparatus for estimating noise variance

The present invention relates to a receiver apparatus and method of channel estimation in a telecommunication system which provides at least two pilot sequences, and to a computer program product. Channel estimation is achieved by estimating channel taps separately for each of the at least two pilot sequences in every transmission block, and for applying estimated channel taps obtained from the estimation to at least one of a temporal and spatial filtering or combining operation to refine the channel estimate. Accordingly, temporal correlations and cross-correlations of the at least two pilot sequences are exploited without requiring knowledge of path delays and beamforming parameters.

Channel estimation using common and dedicated pilots

A symbol-level adaptation method to adapt at least one coefficient of an equalizer, wherein the method comprises the steps of: a) determining () if there exists a shortened pilot channelization code which has a spreading factor shorter than the spreading factor of a full pilot channelization code and which is, at the same time, still orthogonal to any other simultaneously active channelization codes, and, b) if there exists a shortened pilot channelization code, despreading () the pilot channel using the shortened pilot channelization code to obtain a pilot symbol estimation, and c) adapting () the value of the equalizer coefficient according to the error between the pilot symbol estimation obtained from step b) and a corresponding expected pilot symbol.

Symbol-level adaptation method, memory, equalizer and receiver for implementing this method

The present invention relates to a method and a device for detecting active spreading codes for a signal in a communication system and estimating power of the active spreading codes. Further, the present invention ensures efficient interference cancellation. Multiple correlations are performed at a single level of a spreading code tree, the produced output samples are then observed at this level during a defined period, and a decision for the activity and power is met for all existing codes via extracting and processing information from these samples. At the correlation step, unitary fast Walsh Hadamard transformation is performed at practically the highest spreading factor in the system.

Determination of active spreading codes and their powers

A filter weight estimation device (D), for an equaliser of a communication receiver, comprises i) a tap delay line connected to N branches each comprising a descrambler (DS-DSN) and a despreader (DE-DEN), for despreading received signals, corresponding to one or more available multiplexed channels associated with different channelization codes, with a reference code equal to the sum of all the channelization codes associated to the available multiplexed channels, &#xcb;) an adaptive filter (AF) comprising N input filter taps respectively connected to the N branches and a regression input and arranged to estimate the sum of the symbols outputted by the despreaders (DE-DEN) and associated to each available channel, said symbol sum constituting a pseudo-symbol estimate defining a reference symbol estimate (rs), and to implement a mechanism of a LMS type to deliver a chosen number of filter weights (fi), i&#xcb;) a quantiser (SM) arranged to quantise the reference symbol estimate (rs) to estimate a desired reference symbol (d), and iv) an error computation module (AM) for subtracting the reference symbol estimate (rs) from the desired reference symbol (d) to output an error signal (e) feeding the regression input.

Growing community of inventors

Istanbul, Turkey

Ahmet Bastug