Antenna adjustment method based on reinforcement learning

Abstract

The invention discloses an antenna adjustment method based on reinforcement learning. The method comprises the following steps: acquiring MDT data reported by a user, and rasterizing a user cell; adjusting an antenna to align the antenna azimuth beam with the user clustering direction; calculating main cell signal coverage parameters based on the rasterized MDT data, and judging whether the antenna needs to be adjusted or not according to the main cell signal coverage parameters; on the basis of determining an antenna adjustment optimization target, constructing a state set and an action set which are respectively composed of main cell performance parameters and antenna adjustment actions, and realizing optimized adjustment of the antenna through reinforcement learning. According to the invention, the reinforcement machine learning based on the antenna adjustment optimization target is used to replace the manual calculation to realize the optimized adjustment of the antenna, the adjustment speed, efficiency and accuracy of 4G 3D-MIMO and 5G Massive MIMO antennas can be significantly improved, the 4G and 5G network performance indexes are improved, and the network experience of users is improved.

Description

technical field [0001] The invention belongs to the technical field of mobile communication network optimization, and in particular relates to an antenna adjustment method based on reinforcement learning. Background technique [0002] As one of the key 4G enhancement technologies for 5G evolution, the technical advantage of 3D MIMO (Multiple Input Multiple Output) is that it can improve the coverage and capacity of 4G networks at the same time, that is, using beamforming in horizontal and vertical dimensions to improve spectral efficiency. and throughput, meet the multi-level and differentiated capacity requirements of 4G hotspot areas and deep coverage of high-rise buildings, and improve 4G service carrying capacity; The previous implementation and experience preparation are fully applicable to the requirements of Massive MIMO antenna broadcast beamforming in the 5G network era. The corresponding weight optimization ideas for 3D MIMO can be accumulated and transformed into ...

AI Technical Summary

Problems solved by technology

[0003] With the development of 4G and 5G business requirements, the improvement of terminal technology and the surge in the number of users, the contradiction between network traffic and frequency coverage will lead to more and more technical difficulties in 3DMIMO and Massive MIMO network performance evaluation and antenna coverage optimization. , mainly manifested in two aspects: first, the complexity and diversification of user terminals, and the emergence of multi-network standard terminals, including 4GLTE terminals and 5G NR terminals, both single-mode and dual-mode working modes; the second is Different service characteristics of different users will be interwoven in the existing network where 4G and 5G are mixed, making network evaluation standards and antenna parameter dynamic adjustment methods more complicated

As the weight combination of 3DMIMO and MassiveMIMO becomes more and more complex, especially the sub-beam adjustment weight combination scale of MassiveMIMO can reach thousands or tens of thousands, resulting in a sharp increase in the change of network performance data and the change of air interface utilization efficiency. The rasterized evaluation of network performance data and the calculation of antenna weights bring incalculable complexity, which is far beyond what humans can achieve

Images