Mithat C Dogan

US Patent:

6535666, Mar 18, 2003

Filed:

Jan 18, 2000

Appl. No.:

09/484567

Inventors:

Mithat Can Dogan - Sunnyvale CA
Stephen Deane Stearns - Los Altos CA

Assignee:

TRW Inc. - Redondo Beach CA

International Classification:

G02B 626

US Classification:

385 31, 385 12, 359127

Abstract:

Method and apparatus for separating signals transmitted over a waveguide using different propagation modes in overlapping frequency bands. The signals are received at an array of probes installed in an operative relationship to the waveguide, and are separated using a recovery system based on cumulants of the fourth or higher even order.

US Patent:

6650714, Nov 18, 2003

Filed:

Apr 24, 2001

Appl. No.:

09/841456

Inventors:

Mithat C. Dogan - Sunnyvale CA
Paul Petrus - Santa Clara CA
Mitchell D. Trott - Mountain View CA

Assignee:

ArrayComm, Inc. - San Jose CA

International Classification:

H04L 2706

US Classification:

375343, 375267, 375347, 375354

Abstract:

A method and apparatus are provided that performs spatial processing, timing estimation and frequency offset using a training sequence of a received burst. According to one aspect of the present invention, the invention includes receiving a burst having a known training sequence at a set of diversity antennas, sampling the received burst at each antenna, determining a coarse timing estimate for samples from at least one antenna, and determining a spatial weighting vector using the coarse timing estimate. The embodiment further includes applying the spatial weighting vector to the received burst samples for each antenna to form a single channel signal, determining a fine timing estimate for the single channel signal, determining a second spatial weighting vector using the fine timing estimate, applying the second spatial weighting vector to the received burst samples for each antenna to form a second single channel signal, and demodulating the second single channel signal.

US Patent:

6650881, Nov 18, 2003

Filed:

Nov 30, 2000

Appl. No.:

09/727261

Inventors:

Mithat C. Dogan - Sunnyvale CA

Assignee:

ArrayComm, Inc. - San Jose CA

International Classification:

H04B 106

US Classification:

4552761, 455272, 455561, 4555621, 370252, 370334, 375267

Abstract:

A method and apparatus are provided that allows initial beamforming weights to be determined before timing or delay and frequency offset are known. According to one aspect of the present invention, the invention includes receiving a burst from a terminal at a set of spatial diversity antennas, the burst including a known training sequence and storing measurements of the received burst from each spatial diversity antenna corresponding to receive times within a selected analysis window. The invention further includes determining weights for each diversity antenna for at least a portion of the training sequence of the burst before timing for the burst has been determined, applying the determined weights to the stored measurements and determining timing for the burst based on the stored measurements to which the determined weights have been applied.

US Patent:

6697633, Feb 24, 2004

Filed:

Jan 18, 2000

Appl. No.:

09/484724

Inventors:

Mithat Can Dogan - Sunnyvale CA
Stephen Deane Stearns - Los Altos CA

Assignee:

Northrop Grummar Corporation - Los Angeles CA

International Classification:

H04Q 730

US Classification:

455509, 455450, 455562

Abstract:

A method for use in a communication network having multiple users and limited bandwidth, wherein up to N users are allocated to the same frequency band at the same time, and the method provides for processing signals received at an M-element antenna array, using a cumulant-based signal recovery system to separate and recover the signals transmitted by the respective users.

US Patent:

6731689, May 4, 2004

Filed:

Feb 19, 2002

Appl. No.:

10/079620

Inventors:

Mithat C. Dogan - Sunnyvale CA

Assignee:

ArrayComm, Inc. - San Jose CA

International Classification:

H04L 2700

US Classification:

375259

Abstract:

A training sequence for a radio communications system is provided. According to one aspect of the present invention, the invention includes a core sequence of symbols, a successive repetition of the core sequence, and a marker sequence having a sequence of symbols different from the core sequence.

US Patent:

6768747, Jul 27, 2004

Filed:

Nov 30, 2000

Appl. No.:

09/726799

Inventors:

Mithat C. Dogan - Sunnyvale CA

Assignee:

ArrayComm, Inc. - San Jose CA

International Classification:

H04J 306

US Classification:

370509, 370503, 370350

Abstract:

A method and apparatus are provided that allows timing to be determined in two steps. According to one aspect of the present invention, the invention includes searching a received digital signal burst for a known repeating core sequence, the core sequence having a first symbol sequence, and, upon finding one of the core sequences, determining sequence timing of the found core sequence using the found core sequence. The invention further includes searching the received burst for a marker sequence following the repetition of the core sequence, the marker sequence having a second symbol sequence different from the first symbol sequence and upon finding the marker sequence, determining timing of the received digital signal burst using the found marker sequence.

US Patent:

6801589, Oct 5, 2004

Filed:

Nov 30, 2000

Appl. No.:

09/727177

Inventors:

Mithat C. Dogan - Sunnyvale CA

Assignee:

ArrayComm, Inc. - San Jose CA

International Classification:

H03D 100

US Classification:

375343, 375347, 375356, 375324, 375326, 375267, 375354, 370334, 370310

Abstract:

A method and apparatus are provided that allows a coarse timing approximation to be determined from analyzing only a portion of a received burst. The coarse timing can be refined by focusing on the coarse timing approximation. According to one aspect of the present invention, the invention includes receiving a burst having a known repeating core training sequence, selecting an analysis window to be substantially the same size as a multiple of a single repetition of the core training sequence, and over sampling the received burst for the portion overlapping the analysis window. The invention further includes calculating a cross correlation vector for a portion of the samples with respect to a selected part of the training sequence, each cross correlation vector corresponding to a relative timing hypothesis and each cross correlation vector combining samples that occur at intervals within the analysis window, the intervals corresponding to the period of the repeating core training sequence, calculating a least squares fit for each hypothesis using the calculated cross correlation vectors; and selecting the combination of samples corresponding to the minimal least squares fit as the relative timing of the selected part of the repeating core training sequence.

US Patent:

6931030, Aug 16, 2005

Filed:

Nov 30, 2000

Appl. No.:

09/727262

Inventors:

Mithat C. Dogan - Sunnyvale CA, US

Assignee:

ArrayComm, Inc. - San Jose CA

International Classification:

H04J003/06

US Classification:

370509, 370330, 370278

Abstract:

A method and apparatus are provided that allows a delay to be chosen to be applied to a user terminal. According to one aspect of the present invention, the invention includes receiving a downlink broadcast burst from a base station, determining timing for an uplink burst from the received broadcast burst, selecting an amount of delay for the uplink burst, and transmitting an uplink burst to the base station using the determined timing and the selected amount of delay.