Adaptive filtering for precision time protocol with physical layer support
Systems and methods for reducing phase delay variation impact are described. A microcontroller can receive a sequence of phase offsets determined by a slave device over time. The microcontroller can determine a weight vector based on a metric associated with the sequence of phase offsets. The microcontroller can adjust a set of filter coefficients based on the weight vector. The set of filter coefficients can be filter coefficients of a filter being implemented by the slave device to filter incoming packet data.
1. An integrated circuit that implements adaptive filtering to reduce packet delay variation impact, the integrated circuit comprising:
a microcontroller configured to:
receive a sequence of phase offsets determined by a slave device over time;
form a metric vector based on a set of elements sampled from a metric associated with a portion of phase offsets among the sequence of phase offsets;
determine a weight vector based on the metric vector; and
adjust a set of filter coefficients based on the weight vector, wherein the set of filter coefficients are filter coefficients of a filter being implemented by the slave device to filter incoming packet data.
2. The integrated circuit of claim 1 , wherein the microcontroller is configured to perform a statistical analysis to determine the metric.
3. The integrated circuit of claim 1 , wherein the microcontroller is configured to:
extract the portion of phase offsets from the sequence of phase offsets; and
sample the set of elements from the metric associated with the portion of phase offsets among the sequence of phase offsets.
4. The integrated circuit of claim 3 , wherein the microcontroller is configured to execute a sliding window algorithm to extract the portion of phase offsets.
5. The integrated circuit of claim 4 , wherein a size of a sliding window of the sliding window algorithm is based on a physical layer of the slave device.
6. The integrated circuit of claim 1 , wherein the microcontroller is configured to:
normalize the metric vector; and
saturate the normalized metric vector to determine the weight vector.
7. The integrated circuit of claim 1 , wherein the microcontroller is configured to:
determine a minimum value of the metric vector;
determine a mean of the metric vector;
determine a difference between the mean and the minimum value of the metric vector;
compare the difference with a threshold;
in response to the difference being less than the threshold, set every element among the weight vector to a value of one; and
in response to the different being greater than the threshold, saturate a normalization of the metric vector to determine the weight vector.
8. The integrated circuit of claim 1 , wherein the microcontroller is configured to multiply the weight vector with the set of filter coefficients to adjust the set of filter coefficients.
9. A device that implements adaptive filtering to reduce packet delay variation impact, the device comprising:
a physical layer configured to receive packet data from a network; and
a microcontroller configured to be in communication with the physical layer, the microcontroller being configured to:
receive a sequence of phase offsets determined by the device over time;
form a metric vector based on a set of elements sampled from a metric associated with a portion of phase offsets among the sequence of phase offsets;
determine a weight vector based on the metric vector; and
adjust a set of filter coefficients based on the weight vector, wherein the set of filter coefficients are filter coefficients of a filter being implemented by the device to filter packet data received by the physical layer.
10. The device of claim 9 , wherein the microcontroller is configured to perform a statistical analysis to determine the metric.
11. The device of claim 9 , wherein the microcontroller is configured to:
extract the portion of phase offsets from the sequence of phase offsets; and
sample the set of elements from the metric associated with the portion of phase offsets among the sequence of phase offsets.
12. The device of claim 11 , wherein the microcontroller is configured to execute a sliding window algorithm to extract the portion of phase offsets.
13. The device of claim 12 , wherein a size of a sliding window of the sliding window algorithm is based on a physical layer of a slave device.
14. The device of claim 9 , wherein the microcontroller is configured to:
normalize the metric vector; and
saturate the normalized metric vector to determine the weight vector.
15. The device of claim 9 , wherein the microcontroller is configured to:
determine a minimum value of the metric vector;
determine a mean of the metric vector;
determine a difference between the mean and the minimum value of the metric vector;
compare the difference with a threshold;
in response to the difference being less than the threshold, set every element among the weight vector to a value of one; and
in response to the different being greater than the threshold, saturate a normalization of the metric vector to determine the weight vector.
16. The device of claim 9 , wherein the microcontroller is configured to multiply the weight vector with the set of filter coefficients to adjust the set of filter coefficients.
17. A method for performing adaptive filtering to reduce phase delay variation impact, the method comprising:
receiving a sequence of phase offsets determined by a slave device over time;
forming a metric vector based on a set of elements sampled from a metric associated with a portion of phase offsets among the sequence of phase offsets;
determining a weight vector based on the metric vector; and
adjusting a set of filter coefficients based on the weight vector, wherein the set of filter coefficients are filter coefficients of a filter being implemented by the slave device to filter incoming packet data.
18. The method of claim 17 , further comprising:
extracting the portion of phase offsets from the sequence of phase offsets; and
sampling the set of elements from the metric associated with the portion of phase offsets among the sequence of phase offsets.
19. The method of claim 17 , further comprising:
determining a minimum value of the metric vector;
determining a mean of the metric vector;
determining a difference between the mean and the minimum value of the metric vector;
comparing the difference with a threshold;
in response to the difference being less than the threshold, setting every element among the weight vector to a value of one; and
in response to the different being greater than the threshold, saturating a normalization of the metric vector to determine the weight vector.
20. The method of claim 17 , further comprising multiplying the weight vector with the set of filter coefficients to adjust the set of filter coefficients.