|
1 | 1 | # What's New in Simu5G |
2 | 2 |
|
| 3 | +## v1.4.5 (2026-07-09) |
| 4 | + |
| 5 | +This release is a collection of bug fixes to the physical-layer error model |
| 6 | +(CQI and BLER computation), the MAC layer, and the uplink scheduler. Several of |
| 7 | +these fixes change simulation results for the affected configurations. |
| 8 | + |
| 9 | +Tested with INET-4.5.4 and OMNeT++ 6.3, compatible with INET-4.6.0 and OMNeT++ |
| 10 | +6.1 through 6.4. |
| 11 | + |
| 12 | +### PHY error model fixes |
| 13 | + |
| 14 | +- **BLER table indexing**: `GetBLER_TU()`/`GetBLER_AWGN()` indexed the CQI/BLER |
| 15 | + tables one row too low, making throughput results slightly optimistic. |
| 16 | + |
| 17 | +- **CQI boundary condition**: An SNR exactly at the minimum (`minSnr`, -14 dB) |
| 18 | + wrongly yielded the maximum CQI 15 instead of CQI 0, scheduling a noise-floor |
| 19 | + UE at the highest MCS. |
| 20 | + |
| 21 | +- **CQI 0 handling**: `LteRealisticChannelModel::error()` no longer treats CQI 0 |
| 22 | + (a valid "channel unusable" value, e.g. just after handover) as a fatal error; |
| 23 | + the packet is simply dropped. |
| 24 | + |
| 25 | +### MAC and scheduler fixes |
| 26 | + |
| 27 | +- **HARQ process count**: The `harqProcesses` NED parameter (whose NR default is |
| 28 | + 5) was ignored by the C++ code, which used a hardcoded value of 8. The code now |
| 29 | + honors the parameter, so NR uses 5 HARQ processes as v1.4.4 already intended. |
| 30 | + Contributed by Esteban Egea Lopez (Universidad Politécnica de Cartagena). |
| 31 | + |
| 32 | +- **RAC grant sizing**: A UE completing RACH on a poor uplink channel could get a |
| 33 | + grant too small to even carry a Buffer Status Report, leaving it unable to |
| 34 | + report its buffer and re-RACHing forever. Grants are now sized to at least 56 B. |
| 35 | + |
| 36 | +### Other fixes |
| 37 | + |
| 38 | +- **PacketFlowObserverEnb**: An unknown grant ID on an uplink MAC PDU (normal |
| 39 | + during handover) now logs a warning instead of throwing a fatal error. |
| 40 | + |
| 41 | +- **PacketFlowObserver**: BSR-only MAC PDUs (no RLC SDUs) are now tolerated |
| 42 | + instead of triggering a fatal error. |
| 43 | + |
| 44 | +- **AmcPilotAuto**: Using it with a D2D direction now fails with a clear error |
| 45 | + message; scenarios with D2D should set `amcMode="D2D"`. |
| 46 | + |
| 47 | + |
3 | 48 | ## v1.4.4 (2026-05-24) |
4 | 49 |
|
5 | 50 | This release contains bug fixes and improvements, including more realistic MAC |
|
0 commit comments