Unmanned aerial vehicles (UAVs) control information delivery is a critical communication with stringent requirements in terms of reliability and latency. In this context, link adaptation plays an essential role in the fulfillment of the required performance in terms of decode error probability and delay. Link adaptation is usually based on channel quality indicator (CQI) feedback information from the user equipment that should represent the current state of the channel. However, measurement, scheduling and processing delays introduce a CQI aging effect, that is a mismatch between the current channel state and its CQI representation. Using outdated CQI values may lead to the selection of a wrong modulation and coding scheme, with a detrimental effect on performance. This is particularly relevant in ultra reliable and low latency communications (URLLC), where the control of the reliability can be negatively impacted, and it is more evident when the channel is fast varying as the case of UAVs. This paper analyzes the effects of CQI aging on URLLCs, considering transmissions under the finite blocklength regime, that characterizes such communications type. A deep learning approach is investigated to predict the next CQI from the knowledge of past reports, and performance in terms of decode error probability and throughput is given. The results show the benefit of CQI proposed prediction mechanism also in comparison with previously proposed methods.

Bartoli, G., Marabissi, D. (2022). CQI Prediction Through Recurrent Neural Network for UAV Control Information Exchange Under URLLC Regime. IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, 71(5), 5101-5110 [10.1109/TVT.2022.3152408].

CQI Prediction Through Recurrent Neural Network for UAV Control Information Exchange Under URLLC Regime

Bartoli, G;
2022-01-01

Abstract

Unmanned aerial vehicles (UAVs) control information delivery is a critical communication with stringent requirements in terms of reliability and latency. In this context, link adaptation plays an essential role in the fulfillment of the required performance in terms of decode error probability and delay. Link adaptation is usually based on channel quality indicator (CQI) feedback information from the user equipment that should represent the current state of the channel. However, measurement, scheduling and processing delays introduce a CQI aging effect, that is a mismatch between the current channel state and its CQI representation. Using outdated CQI values may lead to the selection of a wrong modulation and coding scheme, with a detrimental effect on performance. This is particularly relevant in ultra reliable and low latency communications (URLLC), where the control of the reliability can be negatively impacted, and it is more evident when the channel is fast varying as the case of UAVs. This paper analyzes the effects of CQI aging on URLLCs, considering transmissions under the finite blocklength regime, that characterizes such communications type. A deep learning approach is investigated to predict the next CQI from the knowledge of past reports, and performance in terms of decode error probability and throughput is given. The results show the benefit of CQI proposed prediction mechanism also in comparison with previously proposed methods.
2022
Bartoli, G., Marabissi, D. (2022). CQI Prediction Through Recurrent Neural Network for UAV Control Information Exchange Under URLLC Regime. IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, 71(5), 5101-5110 [10.1109/TVT.2022.3152408].
File in questo prodotto:
File Dimensione Formato  
CQI_Prediction_Through_Recurrent_Neural_Network_for_UAV_Control_Information_Exchange_Under_URLLC_Regime.pdf

non disponibili

Tipologia: PDF editoriale
Licenza: NON PUBBLICO - Accesso privato/ristretto
Dimensione 930.97 kB
Formato Adobe PDF
930.97 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11365/1216455