With increasingly dense wireless traffic in 5G and 6G networks, the incidence of message faults due to interference is increasing, leading to wasted time and energy on multiple re-transmissions. Disclosed are procedures for assembling a fault-free copy of a message from two corrupted copies. First, measure the modulation quality of each message element. A faulted message element usually has poor modulation quality. Then, select the best message elements from each of the two corrupted copies, and test the merged version against an embedded error-detection code. If the merged copy still fails the test, select each of the message elements that are different in the two faulted copies since they are all suspicious, and test each version with the error-detection code. By recovering a message despite reception errors, another transmission is avoided, saving time and energy, and avoiding contributing yet further to the background noise. Many additional aspects are disclosed.

Network throughput can be increased and the message failure rate can be reduced in 5G and 6G communications by use of AI-based fault mitigation: that ism detection, localization, and correction of faulted message elements in real-time. A receiver provides the demodulated message, along with amplitude and phase measurements of each message element, directly to a properly trained artificial intelligence model. The model determines the most-likely faulted message elements, and in some cases can indicate the most probable correct value of the faulted message elements. The AI model can also determine the fault probability of each message element. The expected message content (such as value ranges and predetermined format) can also be provided to the AI model, for further corruption sensitivity. By correcting faulted messages in less time than required for a retransmission, the system can save time, reduce backgrounds, and greatly reduce dropped messages.

In current practice, faulted messages are typically discarded and a retransmission is requested. Forward error-correction codes (FEC) in 5G and 6G are bulky, resource-expensive, and often unable to resolve the problem. Disclosed are systems and methods for determining which specific message elements are faulted, so that just the faulted portion can be retransmitted, instead of the entire message. For example, the amplitudes of the I and Q branches, of each message element, can be compared to the calibrated amplitude levels of the modulation scheme. Any message element with a large amplitude deviation is suspect. Other factors, such as the SNR, can also be considered in evaluating the validity of each message element. Usually, all of the faulted message elements occupy just a portion of the message. Compact formats are disclosed specifying which portion of the message is to be retransmitted, thereby saving time, power, and background generation.

An improved way is disclosed for recovering a message by merging two corrupted copies of the message in 5G or 6G. Message faults, from noise or interference, distort the modulation of one or more message elements. In a modulation scheme of amplitude-modulated quadrature (I and Q) signals, noise can change the amplitude values, which results in demodulation faults. Often the reception is so poor that a retransmission of the message is also faulted. Nevertheless, the receiver can recover the correct message by measuring a modulation quality of each message element, and assembling a merged message from the best-quality message elements of the two copies. The modulation quality depends on the message element's amplitude versus the calibrated amplitude levels of the modulation scheme. By selecting the individual message elements from the two faulted copies, based on the modulation quality, users can obtain better reception at longer distances while expending less power.

Prior art includes error detection according to an embedded CRC (cyclic redundancy code) or the like, and error correction using FEC (forward error correction) codes, but achieves only partial success in practice, leading to frequent requests for message retransmission. Disclosed is a method for detecting errors in individual message elements using 5G or 6G technologies, by measuring the modulation quality according to how far the amplitude or phase of the message element deviates from the calibrated modulation levels of the modulation scheme. A large deviation indicates a faulted message element, whereas a close match with the calibrated modulation levels is likely correct. By identifying faulted message elements individually, the receiver can recover the message using a number of strategies, disclosed herein. With improved error detection, and localization to individual message elements, network communications can be substantially upgraded at negligible cost, according to some embodiments.

Wireless receivers in 5G and 6G are generally configured to discard faulted messages and request retransmission of the entire message. Disclosed is a procedure enabling the receiver to determine which specific message elements are faulted, and then to request only the faulted portion be retransmitted. Substantial time and wasted power can thereby be saved. The receiver can identify the faulted message element(s) by calculating a modulation quality of each message element and specifying only that portion of the message containing those message elements. For example, the receiver can determine the modulation quality by comparing a difference between the message element's amplitude or phase modulation and the closest predetermined amplitude or phase level of the modulation scheme. A deviation larger than a threshold value strongly suggests that the message element is wrong. The SNR and other factors can also be included in a formula to identify faulted message elements.

Upon receiving a corrupted message in 5G or 6G, a receiver generally rejects the message or ignores it entirely, because determining which message elements are faulted is difficult and complex. AI-based procedures are provided for localizing faults in specific message elements, and for determining the corrected values when possible. AI inputs may include the amplitude or phase modulation quality of each message element, the measured SNR of each message element, the modulation quality of a preceding demodulation reference, and current backgrounds, among other factors. After training (adjusting according to measured network data), the AI model may then determine the most likely faulted message elements, and may also direct the search for the most likely corrected values. By recovering the original corrected message without an unnecessary retransmission, the system can save time, reduce transmission energy, and avoid generating backgrounds. Many additional aspects are disclosed.

An improved way is disclosed for recovering a message by merging two corrupted copies of the message in 5G or 6G. Message faults, from noise or interference, distort the modulation of one or more message elements. In a modulation scheme of amplitude-modulated quadrature (I and Q) signals, noise can change the amplitude values, which results in demodulation faults. Often the reception is so poor that a retransmission of the message is also faulted. Nevertheless, the receiver can recover the correct message by measuring a modulation quality of each message element, and assembling a merged message from the best-quality message elements of the two copies. The modulation quality depends on the message element's amplitude versus the calibrated amplitude levels of the modulation scheme. By selecting the individual message elements from the two faulted copies, based on the modulation quality, users can obtain better reception at longer distances while expending less power.

Disclosed are systems and methods to determine which specific message elements, of a 5G or 6G message, are faulted. By comparing the amplitude or phase modulation of each message element to a predetermined modulation level, and comparing the difference to a threshold, the faulted message elements can be identified and, potentially, corrected. For example, the modulation scheme may provide two superposed orthogonal signals, thereby providing two amplitude-modulated signals per message element, and a modulation quality can be derived according to the differences between those two amplitudes and the closest predetermined amplitude levels of the modulation scheme. The SNR or SINR of each message element can also be measured and included in the modulation quality determination. Artificial intelligence may enable improved or faster determination of the faulted message elements by including additional input factors. The receiver may then mitigate the message by altering just the faulted message elements, saving re-transmission costs.

This application relates to determining transmission quality of a communication channel, in particular for determining a measure of errors in data transmitted as multi-bit symbols. Described is an error checker with an input for receiving an input signal comprising a series of modulated symbols, wherein each symbol encodes multiple bits of a pseudo-random bit sequence. A demodulator is configured to receive the input signal and only partially demodulate at least some of the symbols to generate a partially demodulated bit sequence. A PRBS module is configured to receive the partially demodulated bit sequence and determine the pseudo-random bit sequence and a comparator compares the output of the demodulator to an expected output based on the pseudo-random bit sequence determined by the PRBS module.