Adaptive equalizer system for short burst modems and link hopping radio networks

• US 6,628,707 B2
• Filed: 05/04/2001
• Issued: 09/30/2003
• Est. Priority Date: 05/04/2001
• Status: Expired due to Fees

• Current Assignee: Intel Corporation
• Original Assignee: Radiant Networks PLC
• Inventors: Lu, Jun, Shah, Tushar, Fu, Dengwei, Rafie, Manouchehr S.
• Primary Examiner(s): Chin, Stephen
• Assistant Examiner(s): Ha, Dac V.
• Application Number:
  US09/849,687
• Publication Number:
  US 20020196844A1
• Time in Patent Office: 879 Days
• Field of Search: 375/229-234, 375/350, 708/305, 708/304, 708/322, 708/323
• US Class Current: 375/233
• CPC Class Codes: H04L 2025/03477 not time-recursive
• CPC Class Codes: H04L 2025/03668 to the value at the end of a previous adaptation period
• CPC Class Codes: H04L 25/03038 with a non-recursive structure H04L25/03031 takes precedence

Abstract

A method for an adaptive equalization apparatus in a multiple-link hopping radio system includes hopping among a plurality of radio links to receive variable-length bursts of radio signals on the plurality of radio links and equalizing amplitude and phase variations of a slow channel for each radio link from a received burst on the radio link. Further, the method includes storing the estimated tap coefficients pertinent to each radio link and using the tap weights of the current burst of the radio link to reliably pre-compensate the channel amplitude and phase distortion of a next received burst on the radio link.

Claims

1. A method for receiving and decoding burst radio signals from multiple sources, the method comprising:

receiving bursts of radio signals on a plurality of radio links in a radio system;

compensating amplitude and delay distortions for each radio link from a received burst on the plurality of radio links using a non-adaptive equalizer, a first adaptive equalizer I, a second adaptive equalizer II, the equalizers being K/M fractionally-spaced, K and M being integers and K≦

M;

pre-equalizing a received burst of a radio link using the non-adaptive equalizer, wherein tap coefficients of the non-adaptive equalizer are obtained from the adaptive equalizer II operated on a previous received burst of the radio link;

estimating and removing carrier phase offset for the received burst;

estimating initial tap coefficient values of the adaptive equalizer II using the first adaptive equalizer I operating on the current burst of a present radio link;

adaptively equalizing a delayed version of the current burst using the adaptive equalizer II with the estimated tap coefficient values;

storing the estimated tap coefficient values of the adaptive equalizer II into a memory unit at the end of the current burst of the operating radio link; and

retrieving stored tap coefficients of an adaptive equalizer II from the memory unit into the non-adaptive equalizer to pre-compensate any residual amplitude and delay distortion of a next received burst on each radio link.

2. The method of claim 1 further comprising:

pre-compensating amplitude and delay distortions of a received burst in a non-adaptive, K/M fractionally-spaced equalizer for each radio link using stored tap coefficients of a previous burst of the respective radio link.

3. The method of claim 1 further comprising:

compensating for the carrier phase offset of the received bursts of radio signals;

estimating carrier phase of each link using available pilot and data symbols in a received burst.

4. The method of claim 1 further comprising:

providing the initial tap coefficient values for the second adaptive, fractionally-spaced equalizer II from an adaptive, K/M fractionally-spaced equalizer I.

5. The method of claim 1 further comprising:

in the second adaptive equalizer II, using the initial tap coefficient values of the first adaptive equalizer I as initial coefficient weights to equalize a delayed and carrier phase compensated received burst on a radio link of the plurality of radio links.

6. The method of claim 5 further comprising:

using resulting equalizer tap coefficient values at the end of a current burst of a respective radio link of the second adaptive equalizer II as the initial tap coefficient values for a next data burst in the non-adaptive equalizer on the radio link.

7. The method of claim 5 further comprising:

using resulting equalizer tap coefficient values at the end of a current burst of a respective radio link of the second adaptive equalizer II as initial tap coefficient values for a next data burst in the first adaptive equalizer I on the radio link.

8. The method of claim 1 wherein receiving bursts of radio signals comprises:

receiving one of a data burst in the steady-state operation and multiple invitation packets on radio links associated with neighbors in the radio system for each joining new neighbor in a join operation.

9. The method of claim 8 further comprising:

receiving a programmable equalizer training sequence (ETS) multiplexed with a data burst every S programmable frames per radio link, wherein each respective invitation packet comprises known training symbols at a known timing position in the respective invitation packet.

10. The method of claim 9 wherein receiving the programmable ETS comprises:

periodically multiplexing a maximum-length shift-register (MLSR) code sequence with good correlation properties positioned at the beginning of every S frames with a known modulation format.

11. The method of claim 9 further comprising

adapting tap coefficients using a plurality of known code sequences per each received ETS burst at the beginning of each Sth frame to allow the equalizer to follow changing channel characteristics and to minimize the lag error in the LMS (least-mean square) algorithm per respective link.

12. The method of claim 8 further comprising:

after receiving S frames, receiving an ETS burst multiplexed with the transmitted frame structure for each link among the plurality of radio links; and

training the tap coefficients of the first adaptive equalizer I and the second adaptive equalizer II using a received data burst.

13. The method of claim 8 further comprising:

compensating for amplitude and delay distortion of radio link; and

demodulating the received burst using a known training sequence.

14. The method of claim 13 wherein demodulating comprises:

estimating channel phase and amplitude variations from the known training sequence in the invitation packets during the join operation;

estimating channel phase and amplitude variations from ETS sequences in the steady-state operation;

adapting tap coefficients in a decision-directed mode using data bursts in the steady-state operation; and

detecting processed symbols based on a selected modulation scheme M-QAM.

15. The method of claim 1 further comprising:

receiving an invitation burst for a radio link with a new joining node joining the radio system;

retrieving stored equalizer weights obtained using the invitation packet in the first adaptive equalizer I and obtaining an estimate of tap coefficients as the initial tap coefficient values for steady-state operation of the second adaptive K/M-spaced equalizer II; and

retrieving stored equalizer weights obtained using the invitation burst into a non-adaptive, K/M-fractionally-spaced equalizer for pre-processing a data burst before the carrier phase recovery unit.

16. The method of claim 1 further comprising:

establishing link parameters for a radio link with a new joining node joining the radio system; and

storing the link parameters as initial stored channel information for steady-state operation.

17. The method of claim 1 further comprising:

compensating carrier phase for a data burst by forming an estimate of carrier phase based on the pilot and data symbols in the data burst;

loading the tap coefficient values of the first adaptive equalizer I after L symbols into a second adaptive equalizer II as initial tap coefficient values of the second adaptive equalizer II;

updating and adapting the tap coefficient values for the second adaptive equalizer II using a (K+L) symbol delayed version of the current data burst in a decision-directed mode operation;

freezing the tap coefficient values of the first adaptive equalizer I after the Lth symbol in the burst and providing them to the second adaptive equalizer II;

resuming adaptation of the tap coefficient values of the first adaptive equalizer as soon as a new data burst is received;

storing the tap coefficient values of the second adaptive equalizer II into a memory block after all symbols of the current burst are processed; and

using the stored equalizer weights in the non-adaptive equalizer for pre-processing the data burst for carrier-phase recovery preceding the first adaptive equalizer I and the second adaptive equalizer II.

18. The method of claim 1 wherein estimating carrier phase offset comprises:

estimating carrier phase of the bursts of radio signals using a coarse-compensated amplitude and phase distorted radio signal burst.

19. The method of claim 1 wherein estimating and removing carrier phase offset comprises:

removing impact of carrier phase due to channel impairments to improve convergence of the tap coefficients of the first adaptive equalizer I to an acceptable mean-square error value.

20. The method of claim 19 further comprising:

varying a gain parameter of a least-mean square process in the first adaptive equalizer I and the second adaptive equalizer II based on one or more measured channel quality parameters.

21. The method of claim 20 wherein the one or more measured channel quality parameters comprises received signal strength (RSSI) and signal to noise ratio (SNR).

22. The method of claim 1 wherein estimating and removing carrier phase offset comprises:

receiving at a second adaptive equalizer II a delayed signal from a carrier phase recovery unit, the delayed signal representing the bursts of radio signals from which the impact of carrier phase due to channel impairments is removed; and

using the delayed signal at the second adaptive equalizer II to converge the tap coefficients of the second adaptive equalizer II to an acceptable mean-square error value.

23. A method for receiving radio signals in a link-hopping, burst mode radio receiver, the method comprising:

receiving a first burst of a first radio signal on a first radio link;

determining equalizer weights for a first adaptive equalizer I about the first radio link using the first burst;

subsequently, receiving a next burst of the first radio signal on the first radio link using the equalizer weights obtained in the first burst as initial tap coefficient values;

storing equalizer weights of a second adaptive equalizer II at the end of the current burst to be used in a fixed equalizer to pre-compensate for amplitude and phase variations of received signals;

receiving a first burst of a second radio signal on a second radio link; and

determining equalizer weights for the second radio link using equalizer weights obtained in the first burst of the second radio signal as initial tap coefficient values of the first burst of the second radio signal;

wherein receiving the first burst of the second radio signal comprises;

determining channel information for the second radio link and storing equalizer weights for the second radio link for subsequent use; and

using the stored information as initial equalizer weights of the first equalizer for a second burst of the second radio signal.

24. A method for receiving radio signals in a link-hopping, burst mode radio receiver, the method comprising:

receiving a first burst of a first radio signal on a first radio link;

determining equalizer weights for a first adaptive equalizer I about the first radio link using the first burst;

subsequently, receiving a next burst of the first radio signal on the first radio link using the equalizer weights obtained in the first burst as initial tap coefficient values; and

storing in a memory circuit coupled to the second adaptive equalizer II next burst equalizer weights for the fixed equalizer and the adaptive equalizer I on a present radio link;

wherein determining the equalizer weights comprises downloading to a fixed equalizer complex coefficients from a second adaptive equalizer II for the first radio signal.

25. An adaptive, fractionally-spaced equalizer method for use in a multiple-link hopping, burst adaptive modem, the method comprising:

receiving an amplitude and phase distorted complex signal as a series of bursts, each burst including one or more pilot symbols and data symbols;

pre-compensating for amplitude and phase variations of the received complex signal using a fixed, fractionally-spaced equalizer;

forming an estimation of carrier phase using a coarse estimation of carrier phase to estimate carrier phase for a burst using pilot and data symbols of the burst;

estimating initial tap values of the second adaptive equalizer using a first adaptive equalizer I after an integer number of symbols; and

equalizing the delayed version of the input received signal of the first adaptive equalizer in a second adaptive equalizer II.

26. An adaptive equalizer system for use in a link hopping burst adaptive modem in steady state operation, the system comprising:

a random access digital memory with a plurality of storage locations that are predefined for multiple radio links and storing the tap coefficient values for the respective radio links;

a fixed equalizer for pre-compensating amplitude and phase variations of a present data burst of a present link through the fixed equalizer for which the tap coefficient values pertinent to a previous burst of the present link are loaded from the memory, the fixed equalizer producing a pre-compensated amplitude and phase signal;

a carrier phase recovery unit extracting distorted carrier phase for a plurality of communication links, the carrier phase unit including;

a first stage which acquires a coarse estimate of the distorted carrier phase through a pilot-aided technique using multiple known pilot symbols contained in the present burst of data;

a fixed phase bias removal stage coupled to the first stage;

a complex conjugate phasor generating stage coupled to the fixed phase bias removal stage;

a de-rotating means for de-rotating the equalized signal by a phasor signal generated by the phasor generating stage to remove estimated coarse carrier phase from the equalized signal to produce a coarse phase compensated signal;

a data-aided phase estimator stage for removing modulated signal from the coarse phase compensated signal using a data-aided, decision-directed technique;

an averaging means for averaging M symbols from the data-aided phase estimator stage to reduce noise variance;

a first adaptive equalizer I unit for generating initial tap coefficient values for a second adaptive equalizer II which produces an equalized signal and stores tap coefficient values of the second adaptive equalizer II at the end of the current burst in the memory to be used as the tap coefficient values of a fixed equalizer for the next burst of a respective radio link.