Each 6G base station serves the user devices (UE) nearest to it, but signal quality also depends on interference from every other cell. A Self-Organizing Network (SON) AI layer re-balances handovers using both distance and current cell load, instead of distance alone:
SINR_i = P/d_i² / ( N₀ + Σ_{j≠i} 0.12·P/d_j² ) (signal ÷ noise+interference)
C = B · log₂(1 + SINR) (Shannon capacity per link)
cost(UE→i) = d_i² + λ · load_i (AI handover cost, λ=0 when SON is off)
twin_error = mean| load_EMA − load_actual | / load_actual (digital-twin drift)
- Base-station rings — number of concentric hex rings of 6G cells (1→7, 2→19, 3→37 towers), mirroring a real cellular deployment grid.
- Connected devices — simulated UEs (phones/IoT) roaming the coverage area and constantly re-evaluating which cell serves them.
- SON AI — when ON, handover cost includes each cell's current load (λ>0), spreading devices across towers instead of letting the nearest one become congested; when OFF, devices greedily pick the nearest tower only.
- Digital Twin — an exponentially-smoothed forecast of each tower's load (its "digital twin") is rendered as a translucent holographic layer above the real network; the readout shows how far that prediction drifts from the instantaneous ground truth.