How to study handover in the upper mid-band (FR3)¶

The upper mid-band (FR3, about 7 to 24 GHz) is the range 3GPP studies for 5G-Advanced and beyond; TR 38.901 was re-examined for it in Rel-19. HandoverLab's default world already sits at its lower end (10 GHz). The built-in scenario fr3-urban-15ghz (since 1.0.0rc2) moves the same world to 15 GHz with a 400 MHz carrier and an antenna array of the same size, so that a rule can be compared across the band with everything else held equal.

1. One rule set, two carriers¶

# fr3.yaml
schema: 1
scenario: fr3-urban-15ghz
seeds: [1, 2, 3]
rules:
  - {rule: a3, off_db: 3.0, ttt_s: 0.16}
  - {rule: cho, o_prep_db: -2.0, o_exec_db: 3.0, ttt_exec_s: 0.08}
  - {rule: ltm, o_prep_db: -3.0, o_switch_db: 2.0}
output: results/fr3-15ghz.csv

handoverlab run fr3.yaml

Then set scenario: urban-grid-10ghz and run it again with --output results/fr3-10ghz.csv: the same seeds give the same users, services and site plan at both carriers. Or, in one study, as a condition (paired by seed; add radio.carrier_ghz to tuning.group_by for one tuning per carrier):

scenario: {base: urban-grid-10ghz}
conditions:
  radio.carrier_ghz: [10.0, 15.0]

A study condition changes only the key it names. To move the whole fr3-urban-15ghz set-up (carrier, bandwidth and beam gain together), use it as the base of a second study or vary the keys together in your own scenario files.

2. What to expect and what to check¶

• At 15 GHz the path loss is about 3.5 dB higher and the noise 3 dB higher (twice the bandwidth), while the larger array gains 3.5 dB: the signal-to-noise ratio drops by about 3 dB. Interference scales with the signal, so sinr_db falls less where the cells are interference-limited.
• Compare failures_pm and unnecessary_pm between the carriers for the same rule, not their absolute values (the failure model is not calibrated in absolute terms).
• If a rule tuned at 10 GHz loses at 15 GHz, tune it per carrier (group_by: [radio.carrier_ghz]) before drawing conclusions: an untuned baseline makes any other rule look good.

3. Other points in the band¶

Use the card as a base and override the carrier and the matching values:

scenario: {base: fr3-urban-15ghz, radio: {carrier_ghz: 22.0, bs_gain_db: 26.8}}

(26.8 dB = 20 + 20 log10(22/10) for an array of fixed size.) See add a scenario to give such a variant its own card.

Limits¶

The TR 38.901 UMi street-canyon formulas are used as they are; frequency-dependent penetration, foliage and body loss are not modelled, and the abstraction has no fast fading. The card lists these limits with the source of every value.