| （英） |
In various services that require high-quality mobile communication, there is a challenge in ensuring that a single connection can always meet the service requirements. This issue can be addressed by enhancing communication quality through the redundant use of multiple connections. However, these methods transmit all information over multiple connections, without considering the overhead in data traffic. As a result, when a large number of users simultaneously use multiple connections, there is a concern that data traffic will significantly increase, leading to congestion. In this paper, we propose a method that predicts a deterioration in communication quality and sequentially switches connections. The proposed method detects signs of significant delays by estimating queuing delays based on throughput predictions using machine learning, and selects the optimal connection. Compared to using a single connection, the proposed method reduces the probability of delays exceeding 120 ms by more than 80% without increasing data traffic. These results demonstrate that the proposed method can improve communication quality while minimizing the increase in data traffic by utilizing multiple connections. |