Simulation Results MmWave Communications: Opportunities MmWave Communications: Challenges Integrated Millimeter Wave and Sub-6 GHz Resource Management for Low-Latency Communications in 5G Cellular Networks Omid Semiari 1 and Walid Saad 2 1 Department of Electrical Engineering, Georgia Southern University, Statesboro, GA, Email: [email protected] 2 Bradley Department of Electrical and Computer Engineering, Virginia Tech, Blacksburg, VA, Email: [email protected] Integrated mmWave and sub-6 GHz Cellular Networks Coverage High latency Low complexity Capacity Low latency High complexity Sub-6 GHz mmWave UltraGig 6400 60-GHz transceiver TP-Link Tri-band access point Large available bandwidth, Small size antenna array, Massive MIMO Large Path Loss Directional Transmission (PHY and MAC layers reconsideration) Blockage 20 dB attenuation Integrated Radio Interface Protocols Frame Structure and Problem Formulation Minimum unsatisfied relations (min-UR) NP-hard Number of satisfied UAs at sub-6 GHz RAT Number of satisfied UAs at mmW RAT Proposed Dual-Mode Scheduler Inference of Line-of-Sight Probability Three-states Q-learning model is proposed to extract LoS probability for each user. Two-sided stability of the proposed scheduling framework is mathematically proved. We have shown that the proposed scheme has polynomial time complexity with respect to the number of UAs and resources. 1.Tighter integration features 2.Fast switching between RATs 3.Low latency 4.No backhaul Quality-of-Experience: Problem: Dynamic allocation of mmWave time slots and microwave sub- channels to minimize service outage, subject to delay constraints. Impact of beam-training overhead: Number of iterations: Integrated mmWave-μW μW only mmWave only Not reliable: 15% outage Low QoS High QoS for delay- sensitive services [1] IEEE TWC, July 2017 [2] IEEE ICC, May 2016