Projecte llegit
Títol: Anàlisi del rendiment de la capa física en sistemes 6G sota limitacions de hardware no ideal
Estudiants que han llegit aquest projecte:
JAVAID SHARIF, MUHAMMAD HARIS (data lectura: 29-10-2026)- Cerca aquest projecte a Bibliotècnica
JAVAID SHARIF, MUHAMMAD HARIS (data lectura: 29-10-2026)Director/a: ALONSO ZÁRATE, LUIS
Departament: TSC
Títol: Anàlisi del rendiment de la capa física en sistemes 6G sota limitacions de hardware no ideal
Data inici oferta: 30-01-2026 Data finalització oferta: 30-09-2026
Estudis d'assignació del projecte:
GR ENG SIS TELECOMUN
| Tipus: Individual | |
| Lloc de realització: EETAC | |
| Paraules clau: | |
| Comunicacions sense fils, 5G, 6G, hardware | |
| Descripció del contingut i pla d'activitats: | |
| L'objectiu general del TFG és avaluar l'impacte de limitacions de hardware en el rendiment de la capa física de sistemes 6G, mantenint el compliment de les especificacions d'un estàndard global.
Objectius específics: 1. Modelatge d'imperfeccions de hardware Modelar i simular un transceptor 6G sota condicions de hardware no ideal (soroll de fase elevat, no-linealitats, resolució limitada d'ADC/DAC), representatiu de plataformes amb restriccions tecnològiques. 2. Anàlisi del pressupost d'enllaç (Link Budget) Analitzar com l'augment del consum energètic del processament i la menor eficiència del hardware poden afectar la potència de transmissió disponible, la cobertura i la qualitat de l'enllaç. 3. Avaluació de tècniques avançades del 6G Estudiar la viabilitat pràctica de tècniques com modulacions d'alt ordre, Massive MIMO i operació a freqüències elevades (mmWave/sub-THz) sota escenaris de hardware limitat. 4. Implicacions per a la interoperabilitat Analitzar si les diferències de rendiment observades poden requerir adaptacions a nivell de configuració o perfils d'implementació, sense comprometre la interoperabilitat definida per l'estàndard. |
|
| Overview (resum en anglès): | |
| The sixth generation of mobile communications (6G) is expected to operate in sub-THz bands, with large antenna arrays and high-order modulations. As a result, physical-layer performance depends more than ever on the quality of the radio-frequency hardware. At the same time, United States restrictions on exports of semiconductors and manufacturing equipment to China, together with China's response, are splitting the industry into two technological ecosystems. This thesis studies whether this divergence can translate into measurable performance differences between devices that comply with the same standard.
An OFDM transceiver at 140 GHz has been modelled with four hardware impairments: oscillator phase noise (Wiener process), power amplifier nonlinearity (Rapp model), IQ imbalance and finite ADC resolution. Three hardware scenarios are defined (high, medium and limited), with linewidths of 10 kHz, 100 kHz and 1 MHz, PA smoothness factors of 3, 2 and 1, IRRs of 44.7, 30.4 and 20.9 dB and ADCs of 12, 10 and 4 bits. The chain has been simulated with Sionna, and each impairment has been checked against an analytical model. In AWGN, the simulated BER matches the theoretical one within 0.1 dB. With the same numerology (7.81 MHz subcarrier spacing), the high scenario transmits up to 64-QAM at 2.33 Gb/s, the medium one is limited to QPSK at 0.78 Gb/s, and the limited one does not reach the target BER with any modulation. With an adapted numerology, the limited scenario achieves 0.44 Gb/s and a line-of-sight range of 226 m, compared with 412 m for the high scenario. Phase noise is the impairment that sets the maximum modulation order, and its SIR depends only on the ratio between the linewidth and the subcarrier spacing. The PA is the second source of degradation. IQ imbalance is compensated digitally at negligible cost, and in Massive MIMO the low resolution of the base station ADCs can be compensated with more antennas. The thesis also shows that the AQNM overestimates the SDR of a uniform ADC by up to 3.3 dB, and that its diagonal version for Massive MIMO overestimates the SINR by up to 3.8 dB. Regarding interoperability, a standard that fixes the numerology implicitly fixes a minimum oscillator quality. A terminal below this threshold forces the network to reduce the capacity of the other users, to reserve spectrum for it or to leave it without service. Finally, the most influential impairments depend mainly on RF design and on technologies such as SiGe BiCMOS or InP, rather than on access to the most advanced lithography nodes. |
|