Unbiased code phase discriminator
A feedback control law steers a reference phase that tracks the phase of a received code sequence. The reference phase clocks a track-reference signal consisting of a series of correlation kernels, over which data is extracted and then summed in various combinations. The correlation kernels are designed in such a manner that errors caused by multipath are eliminated or substantially reduced. Furthermore, the areas of the correlation kernels are balanced across level-transitions of a code and non-transitions to eliminate phase biases when tracking specific satellites. Extra care must be taken to balance the correlation kernels in this manner due to a little known aspect of GPS C/A codes. Specifically, not all C/A codes have the same ratio of level-transitions to non-transitions as has been assumed in prior art.
1. A method of determining the phase of a code signal where the code C consists of a known sequence of positive and negative symbols of average width T, where T>0, which method comprises the following steps: generating a first correlation kernel, by using a correlator system, that corresponds to non-transitions of the code signal and has a first correlation kernel shape; generating a second correlation kernel, using the correlator system, that corresponds to transitions of the code signal and has a second correlation kernel shape independent of the shape of the first correlation kernel; and producing a correlative phase discriminator by multiplying at least one of the said correlations by at least one scale factor, and adding together the first and second correlation kernels after at least one is multiplied by the said scale factor.
2. A method of determining the phase of a code portion of a signal where the code C consists of a known sequence of positive and negative symbols of average width T, where T>0, with the method comprising the following steps:
generating, by using a correlator system, a first track reference signal S 1 consisting of both zero and non-zero segments such that when S 1 and C are aligned in phase, all non-zero segments of S 1 occur within a width T/2 of the symbol boundaries of C that exhibit level transitions, either positive-to-negative or negative-to-positive; and
providing the signal S 1 with the further properties that:
each non-zero segment of the product of C and S 1 , when phase aligned, has a right-hand portion and a left-hand portion with the right-hand portion occurring later in time than the left-hand portion and the division between right and left coinciding with a symbol boundary of C;
the mass-balance sum of all right-hand portions is zero; and
the mass balance sum of all left-hand portions exhibits a non-zero value B 1 ,
generating, using the correlator system, a second track reference signal S 2 consisting of both zero and non-zero segments such that when S 2 and C are aligned in phase, all non-zero segments of S 2 occur within a width T/2 of the symbol boundaries of C that exhibit no level transitions, with the signal S 2 having the further property that:
the mass-balance sum of the product of C and S 2 , when phase aligned, has a non zero value B 2 ;
performing, using the correlator system, a first correlation by accumulating of the product of D and S 1 to provide a first correlation kernel with a first correlation kernel shape;
performing, using the correlator system, a second correlation by accumulating of the product of D and S 2 to provide a second correlation kernel with a second correlation kernel shape independent of the shape of the first correlation kernel;
determining at least one scale factor based on the values of B 1 and B 2 ;
multiplying at least one of the said correlations by at least one said scale factor;
forming an error discriminator by adding together the first and second correlation after at least one is multiplied by the said scale factor; and
using the error signal, control the phase of S 1 and S 2 so that the error discriminator is driven to zero.
3. The method of claim 2 where the scale factor is based on the number of symbol boundaries of C that exhibit level transitions and on the number of symbol boundaries of C that do not exhibit level transitions.
4. The method of claim 2 where the signal S 1 and the signal S 2 are combined into one composite signal S where at least two correlations are formed by accumulating the product of S and D, with at least one correlation corresponding to symbol boundaries of C with level transitions and with at least one correlation corresponding to symbol boundaries of C with no level transitions.
5. The method of claim 2 where S 1 and S 2 are further broken down into sub-signals, that when all sub-signals are added together produce the sum of S 1 and S 1 .
6. A method of providing dead-zone compensation to a first correlation kernel in a correlative phase error discriminator system, which method comprises the steps of: providing a first correlation kernel, where said first correlation kernel has a zero value at a nominal track point, a non-zero value to the left and right of the nominal track point, and a zero value in a dead-zone that is not located at the nominal track point; providing a second correlation kernel with a zero value within a zero-band that includes the nominal track point of said first correlation kernel and a non-zero value on one side of the zero-band; measuring a phase error using said first correlation kernel; and providing dead-zone compensation to said first correlation kernel using said second correlation kernel by adding a compensation term to the measured phase error of said first correlation kernel whenever said second correlation kernel reads a non-zero correlation value outside of its zero-band with said compensation term proportional to the size of the non-zero correlation value.
7. The method of claim 6 , which includes the additional steps of: providing an Early-Minus-Late (EML) correlation kernel; providing a prompt correlation kernel; scaling said prompt correlation kernel; summing said EML correlation kernel with said scaled prompt correlation kernel to shift said EML correlation kernel and thereby provide said second correlation kernel; and providing said dead-zone compensation to the phase error discriminator system by zeroing said second correlation kernels within said zero-band including said nominal track point and employing said combined first and second correlation kernels in a code phase tracking loop.
8. The method of claim 7 where the shifted EML correlation kernel is zero for code phase errors that are near to zero.
9. The method of claim 8 where the shifted EML correlation kernel is zero for positive code phase errors.
10. The method of claim 8 where the shifted EML correlation kernel is zero for negative code phase errors.
11. The method of claim 7 where:
the shifted EML correlation kernel is employed in a code phase tracking loop for the purpose of preventing the track loop from losing lock;
the track loop employs a separate correlator for tracking code phase errors; and
said separate correlator is designed to aid in the tracking of code phase error while not being responsive to at least some forms of multipath.
12. The method of claim 7 where:
the shifted EML correlation kernel is employed in a code phase tracking loop for the purpose of aiding in the initial acquisition of code phase;
the said track loop employs a separate correlator for tracking code phase errors; and
said separate correlator is designed to aid in the tracking of code phase error while not being responsive to at least some forms of multipath.