Omnidirectional videos have become a leading multimedia format for Virtual Reality applications. While live 360$^\circ$videos offer a unique immersive experience, streaming of omnidirectional content at high resolutions is not always feasible in bandwidth-limited networks. While in the case of flat videos, scaling to lower resolutions works well, 360$^\circ$video quality is seriously degraded because of the viewing distances involved in head-mounted displays. Hence, in this article, we investigate first how quality degradation impacts the sense of presence in immersive Virtual Reality applications. Then, we are pushing the boundaries of 360$^\circ$technology through the enhancement with multisensory stimuli. 48 participants experimented both 360$^\circ$scenarios (with and without multisensory content), while they were divided randomly between four conditions characterised by different encoding qualities (HD, FullHD, 2.5K, 4K). The results showed that presence is not mediated by streaming at a higher bitrate. The trend we identified revealed however that presence is positively and significantly impacted by the enhancement with multisensory content. This shows that multisensory technology is crucial in creating more immersive experiences.

Recent research has produced mixed results regarding the effectiveness of learning in VR. It has been suggested that the rich multisensory input in VR may induce cognitive overload that impedes the learning process. Cognitive load is typically measured by administering questionnaires. Although questionnaires are easily used, they imply the need to interrupt students during learning or to assess cognitive load in retrospect. In this work-in-progress paper, we argue that VR motion tracking data has the potential to provide unobtrusive, yet valid measures of cognitive load. We report preliminary results from a user study that aims at predicting cognitive load using the tracking data of a VR headset and two hand controllers. Using a recurrent neural network, we were able to distinguish between different levels of cognitive load with an accuracy of more than 88 percent. Based on this finding, we reflect on future research directions and practical considerations.

Most of the existing performance evaluation studies of various routing algorithms are done under limited experimental setups leading to an incomplete picture of the routing algorithm performance under dynamic network conditions. This paper presents a study that compares state-of-the-art routing algorithms over realistic multimedia-based Software Defined Networks (SDNs) with dynamic network conditions and various topology. Routing algorithms remain a key element of the networking landscape as they determine the path the data packets follow. The next-generation networking paradigm offers wide advantages over traditional networks through simplifying the management layer, especially with the adoption of SDN. However, Quality of Service (QoS) provisioning still remains a challenge that needs to be investigated especially for multimedia-based SDNs. This study investigates the impact of state-of-the-art centralized routing algorithms (e.g. MHA, WSP, SWP, MIRA) on multimedia QoS traffic under a realistic environment in terms of PSNR, Throughput, Packet Loss, Delay and QoS rejection.

The massive growth in mobile data traffic and the heterogeneity and stringency of Quality of Service (QoS) requirements of various applications have put significant pressure on the underlying network infrastructure and represent an important challenge even for the very anticipated 5G networks. In this context, the solution is to employ smart Radio Resource Management (RRM) in general and innovative packet scheduling in particular in order to offer high flexibility and cope with both current and upcoming QoS challenges. Given the increasing demand for bandwidth-hungry applications, conventional scheduling strategies face significant problems in meeting the heterogeneous QoS requirements of various application classes under dynamic network conditions. This paper proposes 5MART, a 5G smart scheduling framework that manages the QoS provisioning for heterogeneous traffic. Reinforcement learning and neural networks are jointly used to find the most suitable scheduling decisions based on current networking conditions. Simulation results show that the proposed 5MART framework can achieve up to 50% improvement in terms of time fraction (in sub-frames) when the heterogeneous QoS constraints are met with respect to other state-of-the-art scheduling solutions.

Previous research has shown that adding multisensory media-mulsemedia-to traditional audiovisual content has a positive effect on user Quality of Experience (QoE). However, the QoE impact of employing mulsemedia in 360° videos has remained unexplored. Accordingly, in this paper, a QoE study for watching a 360° video-with and without multisensory effects-in a full free-viewpoint VR setting is presented. The parametric space we considered to influence the QoE consists of the encoding quality and the motion level of the transmitted media. To achieve our research aim, we propose a wearable VR system that provides multisensory enhancement of 360° videos. Then, we utilise its capabilities to systematically evaluate the effects of multisensory stimulation on perceived quality degradation for videos with different motion levels and encoding qualities. Our results make a strong case for the inclusion of multisensory effects in 360° videos, as they reveal that both user-perceived quality, as well as enjoyment, are significantly higher when mulsemedia (as opposed to traditional multimedia) is employed in this context. Moreover, these observations hold true independent of the underlying 360° video encoding quality-thus QoE can be significantly enhanced with a minimal impact on networking resources.

The next generation of wireless networks targets aspiring key performance indicators, like very low latency, higher data rates and more capacity, paving the way for new generations of video streaming technologies, such as 360° or omnidirectional videos. One possible application that could revolutionize the streaming technology is the 360° MULtiple SEnsorial MEDIA (MULSEMEDIA) which enriches the 360° video content with other media objects like olfactory, haptic or even thermoceptic ones. However, the adoption of the 360° Mulsemedia applications might be hindered by the strict Quality of Service (QoS) requirements, like very large bandwidth and low latency for fast responsiveness to the user's inputs that could impact their Quality of Experience (QoE). To this extent, this paper introduces the new concept of 360° Mulsemedia as well as it proposes the use of Reinforcement Learning to enable QoS provisioning over the next generation wireless networks that influences the QoE of the end-users.

Dominated by delay-sensitive and massive data applications, radio resource management in 5G access networks is expected to satisfy very stringent delay and packet loss requirements. In this context, the packet scheduler plays a central role by allocating user data packets in the frequency domain at each predefined time interval. Standard scheduling rules are known limited in satisfying higher quality of service (QoS) demands when facing unpredictable network conditions and dynamic traffic circumstances. This paper proposes an innovative scheduling framework able to select different scheduling rules according to instantaneous scheduler states in order to minimize the packet delays and packet drop rates for strict QoS requirements applications. To deal with real-time scheduling, the reinforcement learning (RL) principles are used to map the scheduling rules to each state and to learn when to apply each. Additionally, neural networks are used as function approximation to cope with the RL complexity and very large representations of the scheduler state space. Simulation results demonstrate that the proposed framework outperforms the conventional scheduling strategies in terms of delay and packet drop rate requirements.

The expected diversity of services and the variety of use cases in 5G networks will require a flexible Radio Resource Management able to satisfy the heterogeneous Quality of Service (QoS) requirements. Classical scheduling strategies have been designed to deal mainly with some particular QoS requirements for specific traffic types. To improve the scheduling performance, this paper proposes an innovative scheduler framework, that selects at each transmission time interval, the appropriate scheduling strategy capable to maximize the users' satisfaction measure in terms of distinct QoS requirements. Neural networks and the Reinforcement Learning paradigm are jointly used to learn the best scheduling decision based on the past experiences. The simulation results show very good convergence properties for the proposed policies, and notable QoS improvements with the respect to the baseline scheduling solutions.

Maintaining a desired trade-off performance between system throughput maximization and user fairness satisfaction constitutes a problem that is still far from being solved. In LTE systems, different tradeoff levels can be obtained by using a proper parameterization of the Generalized Proportional Fair (GPF) scheduling rule. Our approach is able to find the best parameterization policy that maximizes the system throughput under different fairness constraints imposed by the scheduler state. The proposed method adapts and refines the policy at each Transmission Time Interval (TTI) by using the Multi-Layer Perceptron Neural Network (MLPNN) as a non-linear function approximation between the continuous scheduler state and the optimal GPF parameter(s). The MLPNN function generalization is trained based on Continuous Actor-Critic Learning Automata Reinforcement Learning (CACLA RL). The double GPF parameterization optimization problem is addressed by using CACLA RL with two continuous actions (CACLA-2). Five reinforcement learning algorithms as simple parameterization techniques are compared against the novel technology. Simulation results indicate that CACLA-2 performs much better than any of other candidates that adjust only one scheduling parameter such as CACLA-1. CACLA-2 outperforms CACLA-1 by reducing the percentage of TTIs when the system is considered unfair. Being able to attenuate the fluctuations of the obtained policy, CACLA-2 achieves enhanced throughput gain when severe changes in the scheduling environment occur, maintaining in the same time the fairness optimality condition.

As a typical data aggregation technique in wireless sensor networks, the spanning tree has the ability of reducing the data redundancy and therefore decreasing the energy consumption. However, the tree construction normally ignores some other practical application requirements, such as network lifetime, convergence time and communication interference. In this case, the way how to design a tree structure subjected to multi-objectives becomes a crucial task, which is called as multi-objective steiner tree problem (MOSTP). In view of this kind of situation, a multi-objective optimization framework is proposed, and a heuristic algorithm based on jump particle swarm optimization (JPSO) with a specific double layer encoding scheme is introduced to discover Pareto optimal solution. Furthermore, the simulation results validate the feasibility and high efficiency of the novel approach by comparison with other approaches.

In LTE-A cellular networks there is a fundamental trade-off between the cell throughput and fairness levels for preselected users which are sharing the same amount of resources at one transmission time interval (TTI). The static parameterization of the Generalized Proportional Fair (GPF) scheduling rule is not able to maintain a satisfactory level of fairness at each TTI when a very dynamic radio environment is considered. The novelty of the current paper aims to find the optimal policy of GPF parameters in order to respect the fairness criterion. From sustainability reasons, the multi-layer perceptron neural network (MLPNN) is used to map at each TTI the continuous and multidimensional scheduler state into a desired GPF parameter. The MLPNN non-linear function is trained TTI-by-TTI based on the interaction between LTE scheduler and the proposed intelligent controller. The interaction is modeled by using the reinforcement learning (RL) principle in which the LTE scheduler behavior is modeled based on the Markov Decision Process (MDP) property. The continuous actor-critic learning automata (CACLA) RL algorithm is proposed to select at each TTI the continuous and optimal GPF parameter for a given MDP problem. The results indicate that CACLA enhances the convergence speed to the optimal fairness condition when compared with other existing methods by minimizing in the same time the number of TTIs when the scheduler is declared unfair.

The tradeoff concept between system capacity and user fairness attracts a big interest in LTE-Advanced resource allocation strategies. By using static threshold values for throughput or fairness, regardless the network conditions, makes the scheduler to be inflexible when different tradeoff levels are required by the system. This paper proposes a novel dynamic neural Q-learning-based scheduling technique that achieves a flexible throughput-fairness tradeoff by offering optimal solutions according to the Channel Quality Indicator (CQI) for different classes of users. The Q-learning algorithm is used to adopt different policies of scheduling rules, at each Transmission Time Interval (TTI). The novel scheduling technique makes use of neural networks in order to estimate proper scheduling rules for different states which have not been explored yet. Simulation results indicate that the novel proposed method outperforms the existing scheduling techniques by maximizing the system throughput when different levels of fairness are required. Moreover, the system achieves a desired throughput-fairness tradeoff and an overall satisfaction for different classes of users.