Luciano Miuccio
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Luciano Miuccio has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Daniela Panno
Università degli Studi di Catania
13 shared publications - Salvatore Riolo
Department of Electrical, Electonic and Computer Engineering, University of Catania
11 shared publications - Mehdi Bennis
University of Oulu
5 shared publications Salvatore Riolo
Department of Electrical, Electronics and Computer Engineering University of Catania Italy
4 shared publications- Sumudu Samarakoon
Centre for Wireless Communications, University of Oulu, Oulu, Finland
3 shared publications Daniela Panno
2 shared publications
Latest Publications
- Open Access
- 18 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Internet of Things
Resource-Efficient Deployment of Deep Reinforcement Learning-based MAC Protocols in Industrial IoT Devices
In future wireless Industrial Internet of Things (IIoT) networks, satisfying the stringent Quality of Service (QoS) requirements of industrial applications remains a major challenge. Multi-Agent Deep Reinforcement Learning (MADRL) has recently emerged as a promising solution for designing Medium Access Control (MAC) protocols capable of meeting these requirements. While state-of-the-art works show promising performance in simulations, real-world deployment is limited by the severe resource constraints of wireless IoT Devices (IoTDs), particularly with respect to executing Deep Neural Network (DNN) inference under tight latency and energy budgets. This paper addresses the feasibility of deploying MADRL-based MAC protocols during the operational phase by focusing on efficient DNN implementation on IoTDs while preserving communication performance. Using a MADRL-based MAC protocol available in the literature as a reference, we propose a resource-efficient selection criterion to identify the most suitable DNN implementation on constrained devices. The analysis considers multiple solutions, including general-purpose CPUs, quantization, and Domain-Specific Accelerators (DSAs). When accelerators are required, we introduce a methodology to evaluate different architectural configurations and select the most resource-efficient instance. Our real-world evaluation employs the STM32F401RE microcontroller as a case study, incorporating energy and inference latency measurements.
- 18 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Internet of Things
Open AccessResource-Efficient Deployment of Deep Reinforcement Learning-based MAC Protocols in Industrial IoT Devices
In future wireless Industrial Internet of Things (IIoT) networks, satisfying the stringent Quality of Service (QoS) requirements of industrial applications remains a major challenge. Multi-Agent Deep Reinforcement Learning (MADRL) has recently emerged as a promising solution for designing Medium Access Control (MAC) protocols capable of meeting these requirements. While state-of-the-art works show promising performance in simulations, real-world deployment is limited by the severe resource constraints of wireless IoT Devices (IoTDs), particularly with respect to executing Deep Neural Network (DNN) inference under tight latency and energy budgets. This paper addresses the feasibility of deploying MADRL-based MAC protocols during the operational phase by focusing on efficient DNN implementation on IoTDs while preserving communication performance. Using a MADRL-based MAC protocol available in the literature as a reference, we propose a resource-efficient selection criterion to identify the most suitable DNN implementation on constrained devices. The analysis considers multiple solutions, including general-purpose CPUs, quantization, and Domain-Specific Accelerators (DSAs). When accelerators are required, we introduce a methodology to evaluate different architectural configurations and select the most resource-efficient instance. Our real-world evaluation employs the STM32F401RE microcontroller as a case study, incorporating energy and inference latency measurements.
- Top Cited PapersOpen Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Machine Learning in Communications and Networking
- Vol. 4, 677-705
Reliable and Energy-Efficient MAC Protocols in Industrial IoT Networks via Multi-Agent Reinforcement Learning
The transition from wired to wireless communications in industrial Internet of Things (IIoT) networks introduces stringent challenges in terms of reliability and energy efficiency, aggravated by harsh propagation conditions and contention for a shared radio medium. These constraints require advanced medium access control (MAC) protocols capable of jointly managing channel access, packet retransmissions, and buffer operations while accounting for the battery limitations of IoT devices (IoTDs). This paper proposes a multi-agent reinforcement learning (MARL) framework for the autonomous design of energy-efficient and reliable MAC protocols in uplink wireless IIoT networks supporting time–frequency multiplexing. Moving away from conventional decentralized partially observable Markov decision process (Dec-POMDP)-based MARL designs, the framework adopts a partially observable Markov game (POMG), thereby enabling per-device policy learning. A novel reward mechanism is introduced, in which the base station broadcasts a resource-level feedback, and each device constructs a local reward based solely on its own observations and past actions, ensuring feasibility in real deployments. Simulation results show that the proposed framework achieves maximum reliability, whereas state-of-the-art MARL benchmarks based on global rewards fail to meet the required target, highlighting the importance of POMG modeling and local reward structures for reliable wireless IIoT networks. Furthermore, a comparison with conventional grant-based protocols, which inherently achieve maximum reliability, demonstrates that the proposed solution significantly reduces the active-mode duration, thereby improving overall energy efficiency.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Machine Learning in Communications and Networking
- Vol. 4, 677-705
Top Cited PapersOpen AccessReliable and Energy-Efficient MAC Protocols in Industrial IoT Networks via Multi-Agent Reinforcement Learning
The transition from wired to wireless communications in industrial Internet of Things (IIoT) networks introduces stringent challenges in terms of reliability and energy efficiency, aggravated by harsh propagation conditions and contention for a shared radio medium. These constraints require advanced medium access control (MAC) protocols capable of jointly managing channel access, packet retransmissions, and buffer operations while accounting for the battery limitations of IoT devices (IoTDs). This paper proposes a multi-agent reinforcement learning (MARL) framework for the autonomous design of energy-efficient and reliable MAC protocols in uplink wireless IIoT networks supporting time–frequency multiplexing. Moving away from conventional decentralized partially observable Markov decision process (Dec-POMDP)-based MARL designs, the framework adopts a partially observable Markov game (POMG), thereby enabling per-device policy learning. A novel reward mechanism is introduced, in which the base station broadcasts a resource-level feedback, and each device constructs a local reward based solely on its own observations and past actions, ensuring feasibility in real deployments. Simulation results show that the proposed framework achieves maximum reliability, whereas state-of-the-art MARL benchmarks based on global rewards fail to meet the required target, highlighting the importance of POMG modeling and local reward structures for reliable wireless IIoT networks. Furthermore, a comparison with conventional grant-based protocols, which inherently achieve maximum reliability, demonstrates that the proposed solution significantly reduces the active-mode duration, thereby improving overall energy efficiency.
- Open Access
- 26 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 6842-6865
Emerging Generalized Wireless MAC Communication Protocols via Abstraction
In this work, we propose an automated wireless medium access control (MAC) protocol design based on multi-agent reinforcement learning (MARL), in which multiple wireless devices (WDs) and a base station (BS) exchange control messages without prior knowledge of their meanings to coordinate data delivery across the network. In current approaches, the BS acts as a MAC expert, while the WDs are the only learning entities. As a result, the learned MAC protocols are constrained by the BS’s predefined knowledge. In contrast, our approach enables both WDs and the BS to act as learning agents, fostering the collaborative development of new ad hoc communication protocols. However, this approach introduces additional challenges in managing heterogeneous agents that learn simultaneously, alongside conventional issues such as the learned protocols being overfitted to training parameters, including the number of transmitted packets and the number of WDs managed by the BS. To overcome these limitations, we propose a novel learning framework that facilitates the emergence of MAC protocols with robust generalization capabilities. This framework incorporates innovative features such as a novel BS policy model, two distinct reward functions for the WDs and the BS, and different update routines for the agents’ policies. Moreover, we leverage the concepts of parameter sharing and state abstraction to enhance generalization in the WDs’ policies. The MAC protocols generated using the proposed framework are evaluated against state-of-the-art approaches. Simulation results show that the proposed solution significantly outperforms the benchmarks in terms of generalization capabilities.
- 26 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 6842-6865
Open AccessEmerging Generalized Wireless MAC Communication Protocols via Abstraction
In this work, we propose an automated wireless medium access control (MAC) protocol design based on multi-agent reinforcement learning (MARL), in which multiple wireless devices (WDs) and a base station (BS) exchange control messages without prior knowledge of their meanings to coordinate data delivery across the network. In current approaches, the BS acts as a MAC expert, while the WDs are the only learning entities. As a result, the learned MAC protocols are constrained by the BS’s predefined knowledge. In contrast, our approach enables both WDs and the BS to act as learning agents, fostering the collaborative development of new ad hoc communication protocols. However, this approach introduces additional challenges in managing heterogeneous agents that learn simultaneously, alongside conventional issues such as the learned protocols being overfitted to training parameters, including the number of transmitted packets and the number of WDs managed by the BS. To overcome these limitations, we propose a novel learning framework that facilitates the emergence of MAC protocols with robust generalization capabilities. This framework incorporates innovative features such as a novel BS policy model, two distinct reward functions for the WDs and the BS, and different update routines for the agents’ policies. Moreover, we leverage the concepts of parameter sharing and state abstraction to enhance generalization in the WDs’ policies. The MAC protocols generated using the proposed framework are evaluated against state-of-the-art approaches. Simulation results show that the proposed solution significantly outperforms the benchmarks in terms of generalization capabilities.
- Open Access
- 20 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Machine Learning in Communications and Networking
- Vol. 2, 298-317
On Learning Generalized Wireless MAC Communication Protocols via a Feasible Multi-Agent Reinforcement Learning Framework
Automatically learning medium access control (MAC) communication protocols via multi-agent reinforcement learning (MARL) has received huge attention to cater to the extremely diverse real-world scenarios expected in 6G wireless networks. Several state-of-the-art solutions adopt the centralized training with decentralized execution (CTDE) learning method, where agents learn optimal MAC protocols by exploiting the information exchanged with a central unit. Despite the promising results achieved in these works, two notable challenges are neglected. First, these works were designed to be trained in computer simulations assuming an omniscient environment and neglecting communication overhead issues, thus making the implementation impractical in real-world scenarios. Second, the learned protocols fail to generalize outside of the scenario they were trained on. In this paper, we propose a new feasible learning framework that enables practical implementations of training procedures, thus allowing learned MAC protocols to be tailor-made for the scenario where they will be executed. Moreover, to address the second challenge, we leverage the concept of state abstraction and imbue it into the MARL framework for better generalization. As a result, the policies are learned in an abstracted observation space that contains only useful information extracted from the original high-dimensional and redundant observation space. Simulation results show that our feasible learning framework exhibits performance comparable to that of the infeasible solutions. In addition, the learning frameworks adopting observation abstraction offer better generalization capabilities, in terms of the number of UEs, number of data packets to transmit, and channel conditions.
- 20 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Machine Learning in Communications and Networking
- Vol. 2, 298-317
Open AccessOn Learning Generalized Wireless MAC Communication Protocols via a Feasible Multi-Agent Reinforcement Learning Framework
Automatically learning medium access control (MAC) communication protocols via multi-agent reinforcement learning (MARL) has received huge attention to cater to the extremely diverse real-world scenarios expected in 6G wireless networks. Several state-of-the-art solutions adopt the centralized training with decentralized execution (CTDE) learning method, where agents learn optimal MAC protocols by exploiting the information exchanged with a central unit. Despite the promising results achieved in these works, two notable challenges are neglected. First, these works were designed to be trained in computer simulations assuming an omniscient environment and neglecting communication overhead issues, thus making the implementation impractical in real-world scenarios. Second, the learned protocols fail to generalize outside of the scenario they were trained on. In this paper, we propose a new feasible learning framework that enables practical implementations of training procedures, thus allowing learned MAC protocols to be tailor-made for the scenario where they will be executed. Moreover, to address the second challenge, we leverage the concept of state abstraction and imbue it into the MARL framework for better generalization. As a result, the policies are learned in an abstracted observation space that contains only useful information extracted from the original high-dimensional and redundant observation space. Simulation results show that our feasible learning framework exhibits performance comparable to that of the infeasible solutions. In addition, the learning frameworks adopting observation abstraction offer better generalization capabilities, in terms of the number of UEs, number of data packets to transmit, and channel conditions.
- Open Access
On Learning Intrinsic Rewards for Faster Multi-Agent Reinforcement Learning based MAC Protocol Design in 6G Wireless Networks
In this paper, we propose a novel framework for designing a fast convergent multi-agent reinforcement learning (MARL)-based medium access control (MAC) protocol operating in a single cell scenario. The user equipments (UEs) are cast as learning agents that need to learn a proper signaling policy to coordinate the transmission of protocol data units (PDUs) to the base station (BS) over shared radio resources. In many MARL tasks, the conventional centralized training with decentralized execution (CTDE) is adopted, where each agent receives the same global extrinsic reward from the environment. However, this approach involves a long training time. To overcome this drawback, we adopt the concept of learning a per-agent intrinsic reward, in which each agent learns a different intrinsic reward signal based solely on its individual behavior. Moreover, in order to provide an intrinsic reward function that takes into account the long-term training history, we represent it as a long shortterm memory (LSTM) network. As a result, each agent updates its policy network considering both the extrinsic reward, which characterizes the cooperative task, and the intrinsic reward that reflects local dynamics. The proposed learning framework yields a faster convergence and higher transmission performance compared to the baselines. Simulation results show that the proposed learning solution yields 75% improvement in convergence speed compared to the most performing baseline.
Open AccessOn Learning Intrinsic Rewards for Faster Multi-Agent Reinforcement Learning based MAC Protocol Design in 6G Wireless Networks
In this paper, we propose a novel framework for designing a fast convergent multi-agent reinforcement learning (MARL)-based medium access control (MAC) protocol operating in a single cell scenario. The user equipments (UEs) are cast as learning agents that need to learn a proper signaling policy to coordinate the transmission of protocol data units (PDUs) to the base station (BS) over shared radio resources. In many MARL tasks, the conventional centralized training with decentralized execution (CTDE) is adopted, where each agent receives the same global extrinsic reward from the environment. However, this approach involves a long training time. To overcome this drawback, we adopt the concept of learning a per-agent intrinsic reward, in which each agent learns a different intrinsic reward signal based solely on its individual behavior. Moreover, in order to provide an intrinsic reward function that takes into account the long-term training history, we represent it as a long shortterm memory (LSTM) network. As a result, each agent updates its policy network considering both the extrinsic reward, which characterizes the cooperative task, and the intrinsic reward that reflects local dynamics. The proposed learning framework yields a faster convergence and higher transmission performance compared to the baselines. Simulation results show that the proposed learning solution yields 75% improvement in convergence speed compared to the most performing baseline.
- 21 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 10 (9) , 7936-7959
An Energy-Efficient DL-Aided Massive Multiple Access Scheme for IoT Scenarios in Beyond 5G Networks
In view of the challenges foreseen in futuristic massive IoT (mIoT) scenarios, characterized by a huge deployment of energy-constrained IoT devices, we propose an efficient next-generation multiple access (NGMA) scheme where several innovative solutions combine and integrate. First, we aim to maximize the spectral efficiency of the physical uplink-shared channel (PUSCH) by adopting the most appropriate nonorthogonal multiple access (NOMA) technique and by introducing an efficient contention-based approach for data transmission in the unused PUSCH resources. Second, on the basis of the traffic load, the proposed scheme dynamically sets the UpLink (UL) radio resources dimensioning by searching the optimal tradeoff between lowering the collision probability in the physical random access channel (PRACH) and providing enough radio resources in the PUSCH to permit the data transmission to all the succeeded access requests. Third, it is enriched by a series of strategic procedures to make feasible the traffic load estimation, even when the PRACH is in overload condition. Fourth, we exploit artificial intelligence technologies to further optimize the performances obtained. The proposed scheme is compared with other benchmark schemes available in the literature. It outperforms them in the most important metrics for the mIoT scenario (i.e., spectral efficiency and energy consumption), especially in high load conditions, which is the expected requirement for Beyond 5G and 6G networks.
- 21 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 10 (9) , 7936-7959
An Energy-Efficient DL-Aided Massive Multiple Access Scheme for IoT Scenarios in Beyond 5G Networks
In view of the challenges foreseen in futuristic massive IoT (mIoT) scenarios, characterized by a huge deployment of energy-constrained IoT devices, we propose an efficient next-generation multiple access (NGMA) scheme where several innovative solutions combine and integrate. First, we aim to maximize the spectral efficiency of the physical uplink-shared channel (PUSCH) by adopting the most appropriate nonorthogonal multiple access (NOMA) technique and by introducing an efficient contention-based approach for data transmission in the unused PUSCH resources. Second, on the basis of the traffic load, the proposed scheme dynamically sets the UpLink (UL) radio resources dimensioning by searching the optimal tradeoff between lowering the collision probability in the physical random access channel (PRACH) and providing enough radio resources in the PUSCH to permit the data transmission to all the succeeded access requests. Third, it is enriched by a series of strategic procedures to make feasible the traffic load estimation, even when the PRACH is in overload condition. Fourth, we exploit artificial intelligence technologies to further optimize the performances obtained. The proposed scheme is compared with other benchmark schemes available in the literature. It outperforms them in the most important metrics for the mIoT scenario (i.e., spectral efficiency and energy consumption), especially in high load conditions, which is the expected requirement for Beyond 5G and 6G networks.
- Open Access
- 4 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2322-2327
Learning Generalized Wireless MAC Communication Protocols via Abstraction
To tackle the heterogeneous requirements of beyond 5G (B5G) and future 6G wireless networks, conventional medium access control (MAC) procedures need to evolve to enable base stations (BSs) and user equipments (UEs) to automatically learn innovative MAC protocols catering to extremely diverse services. This topic has received significant attention, and several reinforcement learning (RL) algorithms, in which BSs and UEs are cast as agents, are available with the aim of learning a communication policy based on agents' local observations. However, current approaches are typically overfitted to the environment they are trained in, and lack robustness against unseen conditions, failing to generalize in different environments. To overcome this problem, in this work, instead of learning a policy in the high dimensional and redundant observation space, we leverage the concept of observation abstraction (OA) rooted in extracting useful information from the environment. This in turn allows learning communication protocols that are more robust and with much better generalization capabilities than current baselines. To learn the abstracted information from observations, we propose an architecture based on autoencoder (AE) and imbue it into a multi-agent proximal policy optimization (MAPPO) framework. Simulation results corroborate the effectiveness of leveraging abstraction when learning protocols by generalizing across environments, in terms of number of UEs, number of data packets to transmit, and channel conditions.
- 4 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2322-2327
Open AccessLearning Generalized Wireless MAC Communication Protocols via Abstraction
To tackle the heterogeneous requirements of beyond 5G (B5G) and future 6G wireless networks, conventional medium access control (MAC) procedures need to evolve to enable base stations (BSs) and user equipments (UEs) to automatically learn innovative MAC protocols catering to extremely diverse services. This topic has received significant attention, and several reinforcement learning (RL) algorithms, in which BSs and UEs are cast as agents, are available with the aim of learning a communication policy based on agents' local observations. However, current approaches are typically overfitted to the environment they are trained in, and lack robustness against unseen conditions, failing to generalize in different environments. To overcome this problem, in this work, instead of learning a policy in the high dimensional and redundant observation space, we leverage the concept of observation abstraction (OA) rooted in extracting useful information from the environment. This in turn allows learning communication protocols that are more robust and with much better generalization capabilities than current baselines. To learn the abstracted information from observations, we propose an architecture based on autoencoder (AE) and imbue it into a multi-agent proximal policy optimization (MAPPO) framework. Simulation results corroborate the effectiveness of leveraging abstraction when learning protocols by generalizing across environments, in terms of number of UEs, number of data packets to transmit, and channel conditions.
- 20 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 72 (3) , 3288-3303
A Flexible Encoding/Decoding Procedure for 6G SCMA Wireless Networks via Adversarial Machine Learning Techniques
5 G networks may not be competent in fully supporting the high data rates, system capacity, and spectral efficiency expected in the upcoming 6 G networks. An enabling factor to achieve these requirements is the adoption of the Sparse Code Multiple Access (SCMA), which is one of the most promising code-domain Non-Orthogonal Multiple Access (NOMA) techniques. Nowadays, there are two major challenges for the implementation of SCMA in practical networks. First, the design of a near-optimal codebook set at the transmitter, that heavily impacts the network performance. Second, an efficient decoding algorithm at the receiver, that dominates the computational complexity of the system. The two aspects are not independent from each other; therefore the optimization of an SCMA network should address jointly both the Encoder and the Decoder side. Moreover, such solution is far from optimal when designed for general purpose use, since it does not exploit the channel statistical information known at the transmitter and the receiver side. However, most of the works available in the literature address only one aspect and also are not tailored for a given radio channel. In this paper, we design an end-to-end SCMA en/deconding structure based on the integration between a state-of-the-art AutoEncoder (AE) architecture and a novel Wasserstein Generative Adversarial Network (WGAN). The aim is to jointly create a near-optimal codebook design and a non-computationally heavy Decoder that are robust to the given channel statistics. Finally, we trained our model to be flexible to support different channel conditions and compared its performance against the state-of-the-art AE and the best conventional Decoders. The simulation results show that our innovative design exhibits decoding performance very similar to conventional Decoders with a reduced decoding latency.
- 20 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 72 (3) , 3288-3303
A Flexible Encoding/Decoding Procedure for 6G SCMA Wireless Networks via Adversarial Machine Learning Techniques
5 G networks may not be competent in fully supporting the high data rates, system capacity, and spectral efficiency expected in the upcoming 6 G networks. An enabling factor to achieve these requirements is the adoption of the Sparse Code Multiple Access (SCMA), which is one of the most promising code-domain Non-Orthogonal Multiple Access (NOMA) techniques. Nowadays, there are two major challenges for the implementation of SCMA in practical networks. First, the design of a near-optimal codebook set at the transmitter, that heavily impacts the network performance. Second, an efficient decoding algorithm at the receiver, that dominates the computational complexity of the system. The two aspects are not independent from each other; therefore the optimization of an SCMA network should address jointly both the Encoder and the Decoder side. Moreover, such solution is far from optimal when designed for general purpose use, since it does not exploit the channel statistical information known at the transmitter and the receiver side. However, most of the works available in the literature address only one aspect and also are not tailored for a given radio channel. In this paper, we design an end-to-end SCMA en/deconding structure based on the integration between a state-of-the-art AutoEncoder (AE) architecture and a novel Wasserstein Generative Adversarial Network (WGAN). The aim is to jointly create a near-optimal codebook design and a non-computationally heavy Decoder that are robust to the given channel statistics. Finally, we trained our model to be flexible to support different channel conditions and compared its performance against the state-of-the-art AE and the best conventional Decoders. The simulation results show that our innovative design exhibits decoding performance very similar to conventional Decoders with a reduced decoding latency.
- Open Access
Learning Generalized Wireless MAC Communication Protocols via Abstraction
To tackle the heterogeneous requirements of beyond 5G (B5G) and future 6G wireless networks, conventional medium access control (MAC) procedures need to evolve to enable base stations (BSs) and user equipments (UEs) to automatically learn innovative MAC protocols catering to extremely diverse services. This topic has received significant attention, and several reinforcement learning (RL) algorithms, in which BSs and UEs are cast as agents, are available with the aim of learning a communication policy based on agents' local observations. However, current approaches are typically overfitted to the environment they are trained in, and lack robustness against unseen conditions, failing to generalize in different environments. To overcome this problem, in this work, instead of learning a policy in the high dimensional and redundant observation space, we leverage the concept of observation abstraction (OA) rooted in extracting useful information from the environment. This in turn allows learning communication protocols that are more robust and with much better generalization capabilities than current baselines. To learn the abstracted information from observations, we propose an architecture based on autoencoder (AE) and imbue it into a multi-agent proximal policy optimization (MAPPO) framework. Simulation results corroborate the effectiveness of leveraging abstraction when learning protocols by generalizing across environments, in terms of number of UEs, number of data packets to transmit, and channel conditions.
Open AccessLearning Generalized Wireless MAC Communication Protocols via Abstraction
To tackle the heterogeneous requirements of beyond 5G (B5G) and future 6G wireless networks, conventional medium access control (MAC) procedures need to evolve to enable base stations (BSs) and user equipments (UEs) to automatically learn innovative MAC protocols catering to extremely diverse services. This topic has received significant attention, and several reinforcement learning (RL) algorithms, in which BSs and UEs are cast as agents, are available with the aim of learning a communication policy based on agents' local observations. However, current approaches are typically overfitted to the environment they are trained in, and lack robustness against unseen conditions, failing to generalize in different environments. To overcome this problem, in this work, instead of learning a policy in the high dimensional and redundant observation space, we leverage the concept of observation abstraction (OA) rooted in extracting useful information from the environment. This in turn allows learning communication protocols that are more robust and with much better generalization capabilities than current baselines. To learn the abstracted information from observations, we propose an architecture based on autoencoder (AE) and imbue it into a multi-agent proximal policy optimization (MAPPO) framework. Simulation results corroborate the effectiveness of leveraging abstraction when learning protocols by generalizing across environments, in terms of number of UEs, number of data packets to transmit, and channel conditions.
- 18 May 2022
- Published by Elsevier in Computer Communications
- Vol. 191, 299-314
A QoS-aware and channel-aware Radio Resource Management framework for multi-numerology systems
- 18 May 2022
- Published by Elsevier in Computer Communications
- Vol. 191, 299-314
A QoS-aware and channel-aware Radio Resource Management framework for multi-numerology systems