EBOOK
AI-driven accuracy in RAN planning Building Block 1
Only AI/ML-based RAN planning and design optimization can enable operators to reduce of the number of base stations without impacting their user experience. From static equations to learning models Legacy models such as Okumura-Hata, COST or Walfisch-Ikegami remain useful references, but they share constraints: • They assume environments can be sufficiently abstracted into a few tuned parameters • They require per-market, per-morphology calibration campaigns • They are poor at capturing complex 3D effects in dense urban or indoor/outdoor transitions • They struggle to absorb new data sources — crowdsourced data, call traces, rich GIS — at scale An AI-driven approach turns this problem on its head. Instead of a fixed formula with a few tweakable constants, machine learning trained on large volumes of real RF measurements and high-resolution 3D geodata means RAN planning processes become smarter, learning how radio behaves in different environments. The result is a modern AI-powered RF planning and optimization approach that is not static, but adaptive.
Accuracy has always been part of the fundamentals of RF planning. But where in relatively simple legacy macro networks “accurate enough” sufficed, in the more complex world of planning and deploying 5G, accuracy has become both more complex and more valuable. When designing dense urban C-band layers, industrial private networks, ports, airports or campus networks, small systematic errors in prediction rapidly translate into very real impacts. From unnecessary sites and uncovered indoor areas to missed SLAs and expensive post-deployment optimization, the costs of inaccurate RF planning are all too real. For example, independent research by analyst firm Mobile Experts reveals an improvement of 1 dB precision in RSRP leads to roughly 0.4 bps/Hz overall improvement in spectral efficiency; with the same numbers of radios and the same antennas, this is the equivalent of a 24% increase of capacity . For a nationwide U.S. mobile network, this potentially equates to approximately $2 billion over a 10-year period through a reduction in the number of base stations and smalls cells deployed.
Accurate RAN planning saves billions By Mobile Experts
5
Powered by FlippingBook