PARALLEL SOFT SPHERICAL MIMO RECEIVER AND DECODING METHOD

• US 20070286313A1
• Filed: 03/30/2007
• Published: 12/13/2007
• Est. Priority Date: 03/31/2006
• Status: Active Grant

• Current Assignee: BCE Incorporated, University of Waikato
• Original Assignee: BCE Incorporated, University of Waikato
• Inventors: Nikopour-Deilami, Hosein, Saleh, Aladdin, Khandani, Amir
• Granted Patent: US 7,986,752 B2
• Time in Patent Office: Days
• US Class Current: 375/341
• CPC Class Codes: H04L 1/005 Iterative decoding, including iteration between signal detection and decoding operation
• CPC Class Codes: H04L 2025/0342 QAM
• CPC Class Codes: H04L 2025/03426 transmission using multiple-input and multiple-output channels
• CPC Class Codes: H04L 25/0204 of multiple channels
• CPC Class Codes: H04L 25/0248 Eigen-space methods
• CPC Class Codes: H04L 25/03191 in which the receiver makes a selection between different metrics
• CPC Class Codes: H04L 25/03216 using the M-algorithm
• CPC Class Codes: H04L 25/03242 Methods involving sphere decoding
• CPC Class Codes: H04L 25/03286 with channel-decoding circuitry
• CPC Class Codes: H04L 25/067 providing soft decisions, i.e. decisions together with an estimate of reliability H04L25/068 and H04L25/069 take precedence; sequence estimation techniques H04L25/03178
• CPC Class Codes: H04L 27/38 Demodulator circuits; Receiver circuits

Abstract

A method and system for detecting and decoding multiple signals. A low-complexity MIMO detector that combines sphere decoding and m-algorithm approaches, while accounting for the effect of channel condition on the decoding operation, is provided. Taking into account the channel condition effectively controls the size of the search tree, and consequently the search complexity, in an adaptive manner. The channel condition is exploited in the construction of the tree to manage the number of branches in the tree and to avoid undesirable growth.

Claims

1. A method of detecting and decoding a multiple-input multiple-output (MIMO) signal vector, comprising:

determining, in parallel, soft decision outputs for each symbol in the signal vector in view of a priori soft decoded symbol probabilities;

receiving soft decoded symbol probabilities determined in accordance with the soft decision outputs; and

iteratively repeating the steps of determining and receiving to decode the signal vector.

2. The method of claim 1, wherein determining the soft decision outputs comprises determining the soft decision outputs in accordance with a channel condition.

3. The method of claim 1, wherein determining the soft decision outputs comprises generating a tree of lattice points by generating all nodes in a level in parallel, and retaining a predetermined number of nodes as nodes for a next level of the tree.

4. The method of claim 3, wherein generating the tree comprises applying sphere decoding techniques.

5. The method of claim 4, wherein applying sphere decoding techniques comprises determining a radius of a hyper-sphere in accordance with the channel condition.

6. The method of claim 5, wherein applying sphere decoding techniques comprises searching for all lattice points contained within the hyper-sphere to determine the generated nodes;

and determining a branching factor for each generated node.

7. The method of claim 6, further determining a branching metric for a node by adding branch factors of previous nodes leading to a root node.

8. The method of claim 7, wherein retaining only the predetermined number of nodes comprises retaining m nodes having a highest branch metric.

9. The method of claim 8, wherein m is determined adaptively in response to the channel condition.

10. The method of claim 5, wherein the radius is adaptively determined at each level.

11. A MIMO receiver for detecting and decoding a MIMO signal vector, comprising:

an a posteriori probability (APP) detector operable to iteratively determine, in parallel, soft decision outputs for each symbol in the signal vector in view of a priori soft decoded symbol probabilities;

an error correction code (ECC) decoder operable to iteratively provide soft decoded symbol outputs to the APP detector; and

a hard decision processor operable to determine a decoded signal set from the soft decoded symbol outputs after a predetermined number of iterations.

12. The receiver of claim 11, wherein the APP detector is further operable to determine the soft decision outputs in accordance with a channel condition, and the APP detector comprises:

a channel estimator operable to estimate the channel condition;

a parallel sphere decoder operable to determine, in parallel, a set of soft symbol values, and a corresponding branch metric for each soft symbol value, of the signal vector in view of the channel condition and a priori soft decoded symbol probabilities; and

a log-likelihood ratio processor operable to provide the soft decision outputs in accordance with of the set of soft symbol values and branch metrics.

13. The receiver of claim 12, wherein the parallel sphere decoder is operable to generate a tree of lattice points by generating all nodes in a level in parallel, and comprises a pruning module to prune nodes having the lowest branch metrics at each successive level in the tree.

14. The receiver of claim 13, wherein the parallel sphere decoder is operable to determine a radius of a hyper-sphere in accordance with the channel condition.

15. The receiver of claim 14, wherein the parallel sphere decoder is operable to search for all lattice points contained within the hyper-sphere to generate nodes of the tree, and is further operable to determine a branching factor for each determined node.

16. The receiver of claim 15, wherein the parallel sphere decoder is operable to determine a branching metric for a node by adding branching factors of previous nodes leading to a root node.

17. The receiver of claim 13, wherein the pruning module is operable to retain m nodes having a highest branch metric.

18. The receiver of claim 17, wherein the pruning module is further operable to adaptively modify m in response to the channel condition.

19. The receiver of claim 11, wherein the ECC decoder is a turbo decoder.

20. The receiver of claim 11, wherein the ECC decoder is a low-density parity check (LDPC) decoder.

21. The receiver of claim 11, wherein the receiver is compliant with IEEE 802.16 (WiMAX) standards.

22. The system of claim 21, wherein the receiver is compliant with the IEEE 802.16e (Mobile WiMAX) standard.

23. A soft MIMO APP detector, comprising:

a channel estimator operable to estimate a channel condition;

a parallel sphere decoder operable to receive a signal vector, and to determine, in parallel, a set of soft symbol values, and a corresponding branch metric for each soft symbol value, of the signal vector in view of the channel condition and a priori soft decoded symbol probabilities; and

a log-likelihood ratio processor operable to provide soft decision outputs in accordance with the set of soft symbol values and branch metrics.

24. The APP detector of claim 23, wherein the parallel sphere decoder is operable to generate a tree of lattice points by generating all nodes in a level in parallel, and comprises a pruning module to prune nodes having the lowest branch metrics at each successive level in the tree.

25. The APP detector of claim 24, wherein the parallel sphere decoder is operable to determine a radius of a hyper-sphere in accordance with the channel condition.

26. The APP detector of claim 25, wherein the parallel sphere decoder is operable to search for all lattice points contained within the hyper-sphere to generate nodes of the tree, and to determine a branching factor for each determined node.

27. The APP detector of claim 26, wherein the parallel sphere decoder is operable to determine a branching metric for a node by adding branching factors of previous nodes leading to a root node.

28. The APP detector of claim 24, wherein the pruning module is operable to retain m nodes having a highest branch metric.

29. The APP detector of claim 28, wherein the pruning module is further operable to adaptively modify m in response to the channel condition.

30. A parallel sphere decoder for a soft MIMO APP detector, comprising:

a signal conditioning block operable to condition a MIMO signal vector in accordance with a channel estimation;

a parallel search block operable to receive the conditioned signal vector, and to determine, in parallel, a set of soft symbol values, and a corresponding branch metric for each soft symbol value, of the signal vector in view of the channel condition and a priori soft decoded symbol probabilities; and

a pruning block operable to retain a set of m soft symbol values having highest branch metrics at each level of a decision tree.

31. The parallel sphere decoder of claim 30, wherein the parallel search block is operable to generate a the decision tree by generating all nodes in a level in parallel, and the pruning block is operable to prune nodes having the lowest branch metrics at each successive level in the tree.

32. The parallel sphere decoder of claim 31, wherein the parallel search block is operable to determine a radius of a hyper-sphere in accordance with the channel estimation.

33. The parallel sphere decoder of claim 32, wherein the parallel search block is operable to search for all lattice points contained within the hyper-sphere to generate nodes of the tree, and to determine a branching factor for each determined node.

34. The parallel sphere decoder of claim 33, wherein the parallel search block is operable to determine the branch metric for a node by adding branching factors of previous nodes leading to a root node.

35. The parallel sphere decoder of claim 30, wherein the pruning block is further operable to adaptively modify m in response to the channel estimation.

36. The parallel sphere decoder of claim 30, wherein the signal conditioning block is further operable to condition the channel estimation.

37. The parallel sphere decoder of claim 36, wherein the signal conditioning block is operable to decompose and triangularize a channel estimation matrix.

38. The parallel sphere decoder of claim 30, further comprising a log-likelihood ratio processor operable to provide soft decision outputs in accordance with of the set of soft symbol values and branch metrics.

39. A method of parallel sphere decoding, comprising:

conditioning a MIMO signal vector in accordance with a channel estimation;

determining, in parallel, a set of soft symbol values, and a corresponding branch metric for each soft symbol value, of the signal vector in view of the channel condition and a priori soft decoded symbol probabilities; and

pruning the set of soft symbol values to retain those having the highest branch metrics.

40. The method of claim 39, wherein determining the set of soft symbol values comprises generating a the decision tree by generating all nodes in a level in parallel, and pruning away nodes having the lowest branch metrics at each successive level in the tree.

41. The method of claim 40, wherein determining the set of soft symbol values comprises determining a radius of a hyper-sphere in accordance with the channel estimation.

42. The method of claim 41, wherein determining the set of soft symbol values comprises searching for all lattice points contained within the hyper-sphere to generate nodes of the tree, and determining a branching factor for each determined node.

43. The method of claim 42, wherein determining the set of soft symbol values comprises determining the branch metric for a node by adding branching factors of previous nodes leading to a root node.

44. The method of claim 39, wherein pruning the set of soft symbol values further comprises adaptively pruning in response to the channel estimation.

45. The method of claim 39, wherein conditioning the MIMO signal vector further comprises conditioning the channel estimation.

46. The method of claim 45, wherein conditioning the MIMO signal vector comprises decomposing and triangularizing a channel estimation matrix.

47. The method of claim 39, further comprising determining a log-likelihood ratio to provide soft decision outputs in accordance with of the set of soft symbol values and branch metrics.

48. A MIMO communication system, comprising:

a transmitter operable to provide an ECC encoded signal block of multiple signals; and

a receiver having an APP detector operable to iteratively determine, in parallel, soft decision outputs for each symbol in the signal vector in view of a priori soft decoded symbol probabilities;

an ECC decoder operable to iteratively provide soft decoded symbol outputs to the detector; and

a hard decision processor operable to determine a decoded signal set from the soft decoded symbol outputs after a predetermined number of iterations.

49. The system of claim 48, wherein the APP detector comprises:

a channel estimator operable to estimate a channel condition;

a parallel sphere decoder operable to receive a signal vector, and to determine, in parallel, a set of soft symbol values, and a corresponding branch metric for each soft symbol value, of the signal vector in view of the channel condition and the a priori soft decoded symbol probabilities; and

a log-likelihood ratio processor operable to provide the soft decision outputs in accordance with the set of soft symbol values and branch metrics.

50. The system of claim 48, wherein the ECC encoded signal block is turbo encoded and the ECC decoder is a turbo decoder

51. The system of claim 48, wherein the ECC encoded signal block is LDPC encoded, and the ECC decoder is a LDPC decoder.

52. The system of claim 48 employing code division multiple access (CDMA) modulation.

53. The system of claim 48, wherein the receiver and transmitter are compliant with IEEE 802.16 (WiMAX) standards.

54. The system of claim 53, wherein the receiver and transmitter are compliant with the IEEE 802.16e (Mobile WiMAX) standard.

55. A method for performing APP signal detection, comprising:

estimating a channel condition;

decoding symbols of a signal vector, in parallel, to determine a set of soft symbol values, and a corresponding branch metric for each soft symbol value, of the signal vector in view of the channel condition and a priori soft decoded symbol probabilities; and

providing soft decision outputs in accordance with of the set of soft symbol values and branch metrics.

56. The method of claim 55, wherein providing the soft decision outputs comprises determining a log-likelihood ratio.

57. The method of claim 55, wherein decoding symbols of the signal vector comprises generating a tree of lattice points by generating all nodes in a level in parallel, and pruning nodes having the lowest branch metrics at each successive level in the tree.

58. The method of claim 57, wherein decoding symbols of the signal vector further comprises determining a radius of a hyper-sphere in accordance with the channel condition.

59. The method of claim 58, wherein decoding symbols of the signal vector further comprises searching for all lattice points contained within the hyper-sphere to generate nodes of the tree, and determining a branching factor for each determined node.

60. The method of claim 59, wherein decoding symbols of the signal vector further comprises determining a branch metric for a node by adding branching factors of previous nodes leading to a root node.

61. The method of claim 57, wherein pruning the nodes comprises retaining m nodes having a highest branch metric.

62. The method of claim 61, wherein pruning the nodes further comprises adaptively modifying m in response to the channel condition.