Non-integer oversampled timing recovery for higher order quadrature modulation communication systems using quadrature-phase samples
AssigneeThe United States of America as represented by the Secretary of the Air Force
InventorGerard Wohlrab
Life of this patent
Estimating the term…
Abstract
Apparatus and method for performing entirely digital timing recovery for high bandwidth radio frequency communications. The received digital data source can be sampled from any (minimum 2×) non-integer oversampled transmitted data. This method re-samples the data through interpolation and phase adjustment. The output phase error adjusts the receiver's Analog-to-digital Convertor sampling clock to improve synchronization with the transmitter's Digital-to-analog Convertor clock phase, thus improving transmitted symbol recovery.
Claims6
An apparatus for timing recovery in digital communications, comprising:
- a timing recovery subsystem into which quadrature-phase components of a digitized communications signal are input and a phase adjustment signal is output;
- and
- an interpolation subsystem into which said phase adjustment signal and said quadrature-phase components are input and recovered in-phase and quadrature-phase symbols are output;
- wherein said timing recovery subsystem further comprises:
- a buffer into which said quadrature-phase components of said digitized communications signal are input;
- a gain block outputting an update gain signal;
- timing recovery logic, output of which is a phase offset signal, and into which is input:
- an array of buffered quadrature-phase components;
- said update gain signal;
- and
- a feedback sample of said phase offset signal;
- and
- a lowpass filter for removing phase variability extremes in said phase offset signal;
- and
- wherein said timing recovery logic further comprises:
- a signum function calculator into which said array of buffered quadrature-phase components is input;
- a first sampler into which an output of said signum function calculator is input;
- an index calculator having as an input a previous iteration phase offset signal and having as an output sampling indices being input into said sampler;
- a phase error calculator having as an input a real-valued array output from said sampler and having as an output an amount of and direction of current iteration phase offset within a particular sampling iteration;
- and
- a phase adjustment block having as inputs said update gain signal, said current iteration phase offset, and said previous iteration phase offset signal and having as an output an adjusted phase offset signal;
- and
- wherein said buffer samples said quadrature-phase components a minimum of N samples, wherein
- N
- =
- ⌈
- F
- samp
- F
- sym
- ⌉
- ;
- where
- Fsym is the symbol rate;
- and
- Fsamp is the sample rate;
- and
- wherein said index calculator computes said sampling indices according to:
- t
- c
- =
- 1
- +
- k
- ·
- N
- +
- N
- 2
- +
- ϕ
- 0
- t
- s
- =
- t
- c
- -
- N
- 2
- t
- e
- =
- t
- c
- +
- N
- 2
- wherein
- ts is the start of the incoming symbol within said array;
- tc is the center of the incoming symbol within said array;
- te is the end of the incoming symbol within said array;
- k is a counter of which symbol is being sampled within said buffer;
- and
- Φ0 is the previous symbol phase offset.
A method for timing recovery in a digital communications signal having in-phase and quadrature components, comprising the steps of:
- buffering the quadrature-phase components of the digitized communications signal;
- recovering timing of said quadrature-phase components so as to generate a phase adjustment signal;
- lowpass filtering said phase adjustment signal so as to remove phase variability extremes therefrom;
- and
- interpolating between said filtered phase adjustment signal and said in-phase and quadrature components so as to recover quadrature-phase symbols;
- wherein said step of recovering the timing of said quadrature-phase components further comprises the steps of:
- performing a signum function on said buffered quadrature-phase components;
- sampling said signed and buffered quadrature-phase components to form a complex array comprising start (xs), center (xc) and end (xe) symbols;
- determining sampling indices from a previous phase adjustment signal;
- performing a phase offset calculation on said complex array so as to determine an amount and direction of the current phase offset within a sampling iteration;
- and
- generating an adjusted phase offset signal as a function of a gain signal, said current phase offset, and said previous phase adjustment signal;
- wherein said step of sampling further comprises sampling a minimum of N samples, wherein
- N
- =
- ⌈
- F
- samp
- F
- sym
- ⌉
- ;
- where
- Fsym is the symbol rate;
- and
- Fsamp is the sample rate;
- and
- wherein said step of determining said sampling indices further comprises the steps of computing said sampling indices according to:
- t
- c
- =
- 1
- +
- k
- ·
- N
- +
- N
- 2
- +
- ϕ
- 0
- t
- s
- =
- t
- c
- -
- N
- 2
- t
- e
- =
- t
- c
- +
- N
- 2
- wherein
- ts is the start of the incoming symbol within said array;
- tc is the center of the incoming symbol within said array;
- te is the end of the incoming symbol within said array;
- k is a counter of which symbol is being sampled within said buffer;
- and Φ0 is the previous symbol phase offset.
Dependencies are inferred from the claim text.