Intelligent retransmission method enhanced with a risk-scoring machine learning layer.

Abstract

The invention relates to an intelligent retransmission method, enhanced with a risk-scoring machine learning layer, used in error correction and hybrid automated data retransmission request processes in wireless communication systems.

Description

1 TARIFF Intelligent re-transfer powered by a risk-scoring machine learning layer. method Technical Area The invention relates to wireless communication systems (specifically 5G / 6G (5th / 6th Generation Wireless Communication 5 System), IoT (Internet of Things), URLLC (Ultra-Reliable Low-Profile Communications) Latency Communication (Ultra-Reliable Low Latency Communication), V2X (Vehicle-to- Everything (Vehicle-to-Everything Communication) and NTN (Non-Terrestrial Networks) (Non-Network) environments require error correction and hybrid automated data retransmission. (HARQ - Hybrid Automatic Repeat Request) processes are smarter and more adaptable. and relates to a method used to make it more efficient. The invention, in particular, concerns time- For sensitive, low-latency, or bandwidth-limited applications, classic HARQ Developed to prevent inefficiencies in mechanisms, network load, channel Learned risk based on quality, data priority, user type, and other contextual parameters. It customizes the retransmission decision with its score. 15 State of the Art HARQ protocols used in current wireless communication systems are for retransmission. They generally base their decisions on fixed threshold values, channel quality indicators (SINR (Signal- Signal-to-Interference-plus-Noise Ratio (CQI), Channel Quality Indicator – Channel Quality Indicator)) or according to one-way feedback mechanisms 20 These systems provide information on network dynamics, user priorities, or system overall load. It does not take this into account. For example, retransmitting some packets can create unnecessary resources on the system. This usage can lead to increased latency or overprioritization of non-critical data. Yes. Furthermore, current techniques can minimize the potential systemic impacts of the retransmission decision. (e.g., spectral efficiency, QoS (Quality of Service) balance, energy 25 (consumption) cannot be predicted. Enriched with external data and contextual learning. Because no structure with the capability is offered, classic HARQ systems cannot be customized and It produces inexplicable decisions. In conclusion, due to the negative aspects described above and the current solutions being the subject of discussion... Due to its shortcomings, an improvement in the relevant technical field is necessary. 30 It has been made. 2 Purpose of the Invention The invention was created by drawing inspiration from existing situations and overcoming the aforementioned drawbacks. It aims to solve the problem. The method described in the invention involves integrating a HARQ mechanism into the upper layer. A Machine Learning (ML) model and a Risk-Scoring engine are presented. The invention concerns 5 In this method, when a retransmission request comes in, it's not just the physical channel quality that matters; the data... priority, packet delay profile, device type, user class, current network congestion, power constraint A risk / reward score is generated based on several attributes, such as application type. The score estimates the “systemic benefit” of the decision to retransmit each packet. Thus, the HARQ retransmission decision is made not by classical threshold mechanisms, but by very 10 The scores are based on multidimensional data, are scenario-sensitive, and are given using learned scores. In the architecture of the system that enables the implementation of the method described in the invention, each A decision point for the package is generated dynamically. Input feature collector module, By taking channel measurements along with network resource status, connection class, and data type, risk assessment can be performed. It feeds the model. This structure allows the machine learning model to work with only the physical layer. contextual integration not limited to the network and application layers It provides. Machine learning algorithms are used for supervised learning or online. It can be trained using one of the reinforcement learning methods. The method that is the subject of the invention... The system that enables its implementation is a learned policy, repeated under specific conditions. You can also choose not to transmit or to redirect the resource to a user with a different priority. 20 can. In addition, the explainability of the decision-making process is ensured through the decision-making explanatory layer. This layer provides information on which characteristics are effective for each decision made. What gains / harms would the system experience if retransmission did not occur or if it did? It produces assessment information that predicts how long it will live. This situation, system 25 It provides critical insights to operators and researchers. It also updates the model. layer, the system's ability to adapt to changing network conditions, user behavior, or data over time. This allows it to adapt to different types. As a result, the invention in question... The method considers not only technical performance but also learnability, explainability, and flexibility. It also offers an advanced HARQ architecture in terms of features. 30 To achieve the objectives described above, errors in wireless communication systems must be avoided. Risk-scoring systems used in correction and hybrid automated data retransmission request processes. 3 An intelligent retransmission method has been developed, enhanced with a machine learning layer. and the method mentioned is: - Real-time environmental information via the input feature collector module. gathering as - This collected information is processed by a risk scoring engine using machine learning. 5 processing in a model based on and assigning a risk / benefit score to each package production, - this score is the retransmission decision block that will make the retransmission decision. interpreted by [the system / organization], retransmission is made if the score is high, score If the value is low, no retransmission will occur, 10 - through the model update layer, the system's relationship with real data by analyzing decision performance of the machine learning model updating, - through the decision explanation layer, which attributes the decision belongs to and its potential effects on the user or upper layers 15 reporting It includes the steps involved in the process. In a preferred configuration of the method described in the invention, the input feature collector In the process step of collecting environmental information in real time via the module, Environmental information includes channel quality, media status, data type, user profile, and latency (20). At least one of the duration details is being collected. The structural and characteristic features and all the advantages of the invention are given in the figures below. This becomes clearer thanks to the detailed explanation written with references to these figures. This will be understood, and therefore the evaluation should also take these forms and detailed explanations into consideration. It needs to be done by taking precautions. 25 Figures that will help understand the invention. Figure 1 shows the elements of the system that enables the implementation of the method described in the invention. It is a drawing that shows something representatively. Explanation of Part References 1. Input feature collector module 30 4 2. Risk scoring engine 3. Retransmission decision block 4. Model update layer 5. Decision-making layer Detailed Description of the Invention 5 In this detailed explanation, the preferred configurations of the method in question are described only. This is explained to facilitate a better understanding of the subject. The invention in question is an intelligent re-enhanced system powered by a risk-scoring machine learning layer. Factors that enable the transfer method to be implemented: - input feature collector module (1), 10 - risk scoring engine (2), - retransmission decision block (3), - model update layer (4) and - is the decision-making explanatory layer (5). Input feature collector module (1) collects post-transmission medium information, namely channel quality, 15 This module collects information (attributes) such as latency, data priority, user type, etc. Risk scoring engine (2), packet retransmission system with machine learning model This module calculates the impact on system resources as a score. The retransmission decision block (3) is the module that makes the retransmission decision based on the risk score. Model update layer (4) based on real data, decisions made and system outputs 20 It retrains the machine learning model and adapts to the dynamic environment. The decision explanatory layer (5) explains the reason for the decisions taken, the parameters used and The decision explanatory layer (5) presents the effect of the system to the user. It generates explanatory information and creates feedback reports. In the method described in the invention, the recurrence that occurs in the HARQ process after transmission is 25 transmission requirement, classic SINR or CRC (Cyclic Redundancy Check) Instead of evaluating the result of the audit, it is evaluated using multidimensional parameters. The subject of the invention... The steps of this method are as follows: - real-time environmental information via the input feature collector module (1) gathering as - this collected information is used by the risk scoring engine (2) for machine learning 5 processing in a model based on and assigning a risk / benefit score to each package production, - this score will be the retransmission decision block (3) which will make the retransmission decision. interpreted by [the system / organization], retransmission is made if the score is high, score If the value is low, no retransmission will occur, 10 - via the model update layer (4), the system's real data by analyzing decision performance of the machine learning model updating, - through the decision explanatory layer (5), which attributes the decision taken has and its potential effects on the user or upper layers 15 reporting. Thanks to the method that includes the aforementioned steps, it is possible to learn, adapt, and A plausible retransmission mechanism is presented.

Claims

6 REQUESTS 1. Error correction and hybrid automatic data retransmission in wireless communication systems. used in demand processing, enhanced with a risk-scoring machine learning layer. It is an intelligent retransmission method; its feature is: - real-time environmental information via the input feature collector module (1) 5 gathering as - this collected information is used by machine learning by the risk scoring engine (2) processing in a model based on and assigning a risk / benefit score to each package production, - this score will be the retransmission decision block (3) 10 which will make the retransmission decision. interpreted by [the system / organization], retransmission is made if the score is high, score If the voltage is low, no retransmission will be performed. - via the model update layer (4), the system's real data by analyzing decision performance of the machine learning model updated, 15 - through the decision explanatory layer (5), which attributes the decision taken has and its potential effects on the user or higher layers reporting It includes the steps of the process.

2. The method is compliant with Request 1 and its feature is: via the input feature collector module (1), 20 In the process step of collecting environmental information in real time, environmental information factors such as channel quality, environmental conditions, data type, user profile, and latency. This involves gathering information on at least one of their pieces of data.

Images