Random Access with Joint Uplink/Downlink Resource Allocation for Multimedia Tactile Internet

The multimedia Tactile Internet is a network that provides ultra-low latency and ultra-high reliability for two-way multimedia services. Since random access in uplink dominates the delay and there exist variations in target delays and traffic volumes among connections, it is much more efficient to perform a joint uplink and downlink resource allocation. In this paper, we provide uplink random access schemes with joint uplink and downlink resource allocation with multichannel architecture for multimedia Tactile Internet. By doing so, one can significantly increase the resource utilization efficiency and improve the reliability of the multimedia Tactile Internet as well. The cores of resource management are two-fold. First, the joint resource allocation dynamically adjusts the ratio of the number of uplink channels to the number of downlink channels and performs access control based on the collision ratios and the queue lengths. Machine learning techniques such as deep reinforcement learning can be applied to maximize the throughput. Secondly, the efficient uplink random access for multichannel architecture increases the throughput, where each transmitting node selects several time slot/channel pairs based on either theory of the Finite Projective Plane, the random numbers, or the hybrid. In the performance evaluation, we compare the performance of different selection schemes for the multichannel architecture in terms of throughput for the multimedia Tactile Internet through simulation runs and analysis. The hybrid selection scheme performs the best since it has the advantages of both the Finite Projective Plane structure and random numbers. Lastly, we compare and contrast the proposed joint uplink/downlink random access scheme with other known schemes.