US20260101248
2026-04-09
Electricity
H04W36/00838
The patent application discusses a system designed to optimize handover processes in heterogeneous wireless networks using AI/ML models. These models correlate various inputs, such as handover source and target information, with outputs like handover thresholds. The models are developed based on metrics from multiple handovers involving user equipment (UE). Upon receiving a request for handover threshold information, the system selects an appropriate model, matching the request's specific parameters, and provides the necessary threshold information for the UE to decide on handovers.
Wireless networks connect various UEs, including mobile phones and IoT devices, using different radio access technologies (RATs) like 5G, LTE, and Wi-Fi. Networks combining multiple RATs are termed heterogeneous wireless networks. For instance, Fixed Wireless Access (FWA) devices may connect to a network using one RAT and serve as access points using another. UEs capable of connecting via multiple RATs can switch between these technologies, enhancing connectivity options.
In heterogeneous networks, UEs often need to switch between different types of network access points, like moving from a base station to an FWA device. This switch, known as a handover, is crucial when the UE moves to areas with better signal strength or when the current connection becomes congested. Handover thresholds, based on signal strength indicators like SINR and RSSI, help determine when a UE should switch networks. These thresholds are vital for maintaining quality of service (QoS) and minimizing disruptions during handovers.
Manually setting handover thresholds is challenging due to the diverse environments and conditions in which network devices operate. Thresholds suitable for one setting might not work in another. The invention proposes using AI/ML to automatically adjust these thresholds, considering the specific characteristics of each environment. This approach ensures optimal handover performance, reducing disruptions and improving the overall user experience by tailoring thresholds to specific scenarios and times.
An example scenario involves a UE connected to a base station detecting a potential handover situation due to low signal quality. The UE requests smart handover thresholds from a Smart Handover System (SHS), which uses stored models to provide tailored thresholds. The request may include UE identifiers and attributes, helping the SHS refine its response. This automated process ensures the UE maintains optimal connectivity by dynamically adjusting to network conditions.