During the period of 2018 to 2020, typical LTE networks around the world will have five to ten frequencies. Inter-frequency measurements are required to obtain reference signal received power (RSRP) values of cells on neighboring E-UTRA frequencies for carrier selection in carrier aggregation (CA), mobility load balancing (MLB), and coverage-based handovers. When there are many frequencies, inter-frequency measurements by UEs may be delayed, causing the user-experienced data rate to decrease.
To accelerate carrier selection, inter-frequency handovers, or MLB and increase the user experienced data rate, Multi-carrier Optimization by AI-based Virtual Grid has been introduced. (AI is short for artificial intelligence.) With this solution, the eNodeB constructs AI-based virtual grid models for each of its served cells and uses these models to directly predict the RSRP values of cells on neighboring E-UTRA frequencies for each UE based on the UE's intra-frequency measurement results. The models for a cell require input of data associations between intra- and inter-frequency measurement reports (MRs) of previous UEs in the cell.
CA Optimization by AI-based Virtual Grid improves the uplink and downlink data rates of CA UEs by up to 25%. Inter-band Handover Optimization by AI-based Virtual Grid improves the uplink and downlink data rates of cell-edge UEs by up to 25%. MLB Optimization by AI-based Virtual Grid improves the uplink and downlink data rates of UEs transferred to other cells by up to 25%.