Optimization of all software modem using flexible configuration parameters for high-performance computing (HPC)
A method to provide flexibility on the configuration and operation of the modulator, demodulator, and modem, where purpose-built (legacy) devices are not traditionally capable of exposing a level of control and flexibly for a user or an autonomous program for optimizing performance. Providing user or programmatic control of algorithms is traditionally not possible for purpose-built modems. Parameters such as the number of decoder iterations that are performed on Forward Error Correction (FEC), Interference Mitigation algorithm, or dynamic adjustment loop bandwidth to combat phase noise can be adjusted autonomously to optimize receiver performance. The all software modem, supported by a High-Performance Computing (HPC) architecture, removes the limitation due to the flexibility of programming resources and available performance. Unlike most purpose-built hardware, the HPC allows processing resources to dynamically be reallocated, so that as additional performance is desired, the resources may be increased and decreased as required.
1. A method for optimization of an all software modem using flexible configuration parameters for high-performance computing (HPC) platform comprising:
receiving a digital IQ data at the all software modem,
wherein the all software modem comprises a demodulator, and wherein the demodulator is configured to adjust demodulator parameters;
storing and queuing the digital IQ data for downstream processes for each demodulator stage;
adjusting the demodulator parameters of a demodulator stage in response to decoding of data fails; and
re-attempting the demodulator stage with the adjusted demodulator parameters,
wherein the HPC platform can detect in real time a type of jamming environment and apply, in real time, a technique to mitigate interference.
2. The method of claim 1 , wherein the digital IQ data is stored and queued for downstream processes for each demodulator stage.
3. The method of claim 1 , wherein a carrier recovery loop (CRL) bandwidth is adjusted in response to decoded errors.
4. The method of claim 2 , wherein the demodulator comprises a digital IF transport receiver, an Interference Mitigation (IM), a Root-Raised Cosine Filter (RRCF), an equalizer, a timing recovery, a carrier recovery, a signal demodulator, a PL Deframer, a FEC Decoder, a BB Deframer.
5. The method of claim 4 , wherein the demodulator is configured to adjust parameters of the IM, the equalizer, the timing recovery, the carrier recovery, the demodulator, and the decoder.
6. The method of claim 4 , the digital IQ data at a demodulator stage is stored and retrieved with full fidelity.
7. The method of claim 4 , wherein the method further comprises altering FEC parameters and decreasing a number of Low-Density Parity Check (LDPC) iterations.
8. The method of claim 1 , wherein the data is reprocessed in a serialized or parallelized manner.
9. The method of claim 4 , wherein a signal is sent back to the RRCF and the equalizer to attempt to reprocess a block of data that failed a decoding of a FEC block using a different equalizer and passed back into a Low Density Parity Check (LDPC) decoder.
10. The method of claim 9 , wherein the equalizer operation is repeated for multiple equalizer configurations until the LDPC decode can decode the incoming data or reaches a threshold of iterations or time limit.
11. The method of claim 4 wherein stored packets are reprocessed with the interference mitigation program and passed back into a Low Density Parity Check (LDPC) decoder.
12. The method of claim 4 wherein stored packets are reprocessed with multiple equalizers in parallel and a voting performed to see if more than one equalizer configuration results in the same data and passed back into a Low Density Parity Check (LDPC) decoder.
13. The method of claim 12 , wherein the stored packets are reprocessed with multiple equalizer configurations until the LDPC decode can decode the incoming data or reaches a threshold of iterations or time limit.
14. The method of claim 4 , wherein stored packets are reprocessed simultaneously in serial and a voting performed to see if more than one equalizer configuration results in the same data and passed back into a Low Density Parity Check (LDPC) decoder.
15. The method of claim 14 , wherein the stored packets are reprocessed repeatedly for multiple equalizer configurations until the LDPC decoder can decode the incoming data or reaches a threshold of iterations or time limit.
16. The method of claim 4 , wherein the received IQ data is stored in a queue sufficiently sized to allow multiple carrier recovery loop bandwidth of increasing size to be tried that minimizes the decoded errors.
17. The method of claim 1 , wherein carrier recovery bandwidth is incrementally changed to minimize decoded errors.
18. The method of claim 17 , wherein the change is done in a streaming manner, without employing any processing queues.
19. A method for sending data via all software modem using flexible configuration parameters for high-performance computing (HPC) platform, the method comprising:
sending a first user network data from a source;
receiving the first user network data at a first modem, wherein the first user network data is modulated to a digital IQ data;
sending the digital IQ data to a destination via a network;
receiving the digital IQ data at a second modem, wherein the second modem is configured to run an interference mitigation program on the incoming data;
demodulating the digital IQ data to a second user network data; and
receiving the second user network data at the destination,
wherein the all software modem comprises a digital IF transport receiver, an Interference Mitigation (IM), a Root-Raised Cosine Filter (RRCF), an equalizer, a timing recovery, a carrier recovery, a demodulator, a PL Deframer, a FEC Decoder, a BB Deframer, and
wherein a signal is sent back to the RRCF and the equalizer to attempt to reprocess a block of data that failed a decoding of a FEC block using a recursive least squares or a zero forcing algorithm in response to the block of data is not processed.