Calibration data
Apparatus is provided for: storing at least one three-dimensional matrix C[M][N][K] of calibration data; performing singular value decomposition of each at least one three-dimensional matrix C[M][N][K] of calibration data to produce at least one first unitary rotation matrix U, at least one diagonal scaling matrix S and at least one second unitary rotation matrix V; and resizing each of the at least one first unitary rotation matrix U, the at least one diagonal scaling matrix S and the at least one second unitary rotation matrix V by removing dimensions therefrom, thereby producing resized matrices U, S and V. Corresponding decompression is also provided.
1. Apparatus, comprising at least one processor, at least one memory, and computer-readable code stored on the at least one memory, wherein the computer-readable code when executed controls the at least one processor to perform a method comprising:
storing by the apparatus, at least one three-dimensional matrix C[M][N][K] of calibration data, wherein M is an azimuth angle range 0 to 360 degree, N is an elevation angle range 0 to 90 degree, and K is channels to transmit or receive signal, wherein the calibration data is antenna array calibration data;
compressing by the apparatus, the calibration data to produce compressed calibration data, at least a portion of which is wirelessly transmitted to a mobile device using a wireless carrier, the compressed calibration data being capable of being decompressed by the mobile device for use in determining a bearing, the compressing the calibration data comprising the steps of:
performing by the apparatus, singular value decomposition of each at least one three-dimensional matrix C[M] [N] [K] of calibration data to produce at least one first unitary rotation matrix U, at least one diagonal scaling matrix S and at least one second unitary rotation matrix V; and
resizing by the apparatus, each of the at least one first unitary rotation matrix U, the at least one diagonal scaling matrix S and the at least one second unitary rotation matrix V by removing dimensions therefrom, thereby producing resized matrices U, S and V.
2. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform: quantising float type data in the resized matrices U, S and V into signed fixed point data.
3. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform: causing transmission of a first layer of each of the resized matrices U, S and V.
4. Apparatus as claimed in claim 3 , wherein the computer-readable code when executed controls the at least one processor to perform: causing transmission of a second layer of each of the resized matrices U, S and V subsequent to and separately from transmitting the first layers of the resized matrices U, S and V.
5. Apparatus as claimed in claim 4 , wherein the computer-readable code when executed controls the at least one processor to perform: causing transmission of a third layer of each of the resized matrices U, S and V subsequent to and separately from transmitting the second layers of the resized matrices U, S and V.
6. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the singular value decomposition of each at least one three-dimensional matrix C[M] [N] [K] of calibration data to produce the first unitary rotation matrix U, the diagonal scaling matrix S and the second unitary rotation matrix V by decomposing each of K 2-D matrices X[M] [N] into the matrices U, S, and V by the singular value decomposition to provide resized first unitary rotation matrices U of size M-by-M, resized diagonal scaling matrices S of size M-by-N, and resized second unitary rotation matrices V of size N-by-N, wherein K matrices X[M][N] are provided and wherein each matrix X[M] [N] is composed by the first and second dimensions of the three-dimensional matrix C[M] [N] [K] of calibration data.
7. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the singular value decomposition of each at least one three-dimensional matrix C[M] [N] [K] of calibration data to produce the first unitary rotation matrix U, the diagonal scaling matrix S and the second unitary rotation matrix V by:
swapping over second and third dimensions N and K of the three-dimensional matrix C[M] [N] [K] of calibration data to provide a rearranged three-dimensional matrix C[M][K][N]; and
decomposing each of N 2-D matrix X[M] [K] into the matrices U, S, and V by the singular value decomposition to provide resized first unitary rotation matrices U of size M-by-M, resized diagonal scaling matrices S of size M-by-K, and resized second unitary rotation matrices V of size K-by-K, wherein N matrices X[M][N] are provided and wherein each matrix X[M][K] is composed by the first and second dimensions of the rearranged three-dimensional matrix C[M][N][K].
8. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the singular value decomposition of each at least one three-dimensional matrix C[M] [N] [K] of calibration data to produce the first unitary rotation matrix U, the diagonal scaling matrix S and the second unitary rotation matrix V by:
rearranging each three-dimensional matrix C[M] [N] [K] of calibration data by merging second and third dimensions N and K as a first dimension of a single three-dimensional matrix X[N*K][M]; and
decomposing the three-dimensional matrix X[N*K] [M] by the singular value decomposition to produce a resized first unitary rotation matrix U of size N*K-by-N*K, a resized diagonal scaling matrix S of size N*K-by-M, and a resized second unitary rotation matrix V of size M-by-M.
9. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of the first unitary rotation matrix U by discarding one or more columns of the first unitary rotation matrix U.
10. Apparatus as claimed in claim 9 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of at least some of the first unitary rotation matrix U by keeping the first P columns and discarding other columns of the first unitary rotation matrix U.
11. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of the diagonal scaling matrix S by discarding one or more columns and one or more rows of the diagonal scaling matrix S.
12. Apparatus as claimed in claim 11 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of at least some of the diagonal scaling matrix S by keeping the first Q rows and the first Q columns and discarding other columns and rows of the diagonal scaling matrix S.
13. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of the second unitary rotation matrix V by discarding one or more rows of the second unitary rotation matrix V.
14. Apparatus as claimed in claim 13 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of at least some of the second unitary rotation matrix V by keeping the first Q rows and discarding other rows of the second unitary rotation matrix V.
15. Apparatus as claimed in claim 1 , wherein the computer-readable code when executed controls the at least one processor to perform the resizing of at least some of each of the first unitary rotation matrix U, the diagonal scaling matrix S and the second unitary rotation matrix V by:
discarding one or more columns of the first unitary rotation matrix U;
discarding one or more columns and one or more rows of the diagonal scaling matrix S; and discarding one or more rows of the second unitary rotation matrix V.
16. Apparatus as claimed in claim 1 , wherein the at least one three-dimensional matrix of calibration data comprises first and second three-dimensional matrices Cv[M] [N] [K] and Ch[M] [N] [K] of calibration data, each three-dimensional matrix relating to a different polarisation.
17. Apparatus, comprising at least one processor, at least one memory, and computer-readable code stored on the at least one memory, wherein the computer-readable code when executed controls the at least one processor to perform a method comprising:
compressing uncompressed calibration data to produce plural matrices of compressed calibration data;
storing by the apparatus, data selected from the plural matrices of compressed calibration data, wherein the plural matrices together define at least one three-dimensional matrix C[M][N][K] of the uncompressed calibration data and comprise at least one first unitary rotation matrix U, at least one diagonal scaling matrix S and at least one second unitary rotation matrix V obtained by singular value decomposition, wherein M is an azimuth angle range 0 to 360 degree, N is an elevation angle range 0 to 90 degree, and K is channels to transmit or receive signal, and the calibration data is antenna array calibration data; and
causing transmission by the apparatus, of the data selected from the plural matrices of compressed calibration data, to a mobile device using a wireless carrier, for determining a bearing by the mobile device.
18. Apparatus as claimed in claim 17 , wherein the computer-readable code when executed controls the at least one processor to perform: causing transmission of data comprising a first layer of each of the plural matrices of compressed calibration data.
19. Apparatus, comprising at least one processor, at least one memory, and computer-readable code stored on the at least one memory, wherein the computer-readable code when executed controls the at least one processor to perform a method comprising:
receiving data by a mobile device, using a wireless carrier, the data comprising a first layer of each of plural matrices of compressed calibration data received separately from other layers of the plural matrices, a second layer of each of the plural matrices of compressed calibration data received separately from the other layers of the plural matrices, and a third layer of each of plural matrices of compressed calibration data received separately from the other layers of the plural matrices;
using the received data comprising the first, second, and third layers of each of the plural matrices of compressed calibration data, by the mobile device, to reconstruct the plural matrices of compressed calibration data, by generating respective sets of coarse calibration matrices, one set generated for each of the first, second, and third layers, the plural matrices comprising at least one first unitary rotation matrix U, at least one diagonal scaling matrix S and at least one second unitary rotation matrix V, wherein the compressed calibration data is antenna array calibration data;
performing by the mobile device, matrices multiplication of U, S and V to obtain the reconstructed matrices;
receiving by the mobile device, positioning packets over the wireless carrier; and
computing a bearing by the mobile device by operating on the received positioning packets with the reconstructed matrices.