2473-2400 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TGCN.2018.2881242, IEEE Transactions on Green Communications and Networking IEEE TRANSACTIONS ON GREEN COMMUNICATIONS AND NETWORKING 1 Abstract— Heterogeneous cellular networks (HCNs) provide a promising paradigm for supporting different types of devices that have diverse quality of service (QoS) requirements such as Internet of things devices (IoTDs) and human-to-human devices (H2HDs). In this paper, a distributed cell association algorithm is developed to consider the dissimilar association requirements for IoTDs and H2HDs coexisting in HCNs. The device association process is formulated as a multi-objective optimization problem to minimize the uplink (UL) transmit power for IoTDs and maximize the downlink (DL) data rate for H2HDs while considering the various devices’ QoS requirements. To solve this problem, an approach dependent on matching theory is proposed to model the interactions between devices and base stations (BSs) in the network. Using this approach, a distributed algorithm is developed to provide the device association solution. The proposed algorithm is then proved to converge to a stable matching. Simulation results validate the effectiveness of the proposed algorithm in improving the UL power performance for IoTDs as well as DL rate performance for H2HDs compared to other association algorithms. Index Terms— Heterogeneous cellular networks, cell association, human-to-human, Internet of things, matching game. I. INTRODUCTION nternet of things (IoT) has recently been deemed as one of the most promising communication paradigms, which allows smart devices to communicate with Internet for requiring services [1]. Wireless communications solutions play a critical role for enabling IoT because of ubiquity of emerging mobile IoT devices (IoTDs). However, most of wireless technologies that operate in unlicensed bands, such as Zigbee, WiFi, and Bluetooth, cannot provide the seamless, ubiquitous, and quality of service (QoS) requirements for IoT communications [2]. In comparison with unlicensed band technologies, cellular networks can provide mobility support, global coverage, and guaranteed QoS [2]. With the proliferation of emerging IoT applications and video streaming, the data traffic demand in cellular networks grows enormously [3]. In order to accommodate this demand, mobile networks are strongly trending towards deploying of small cell-based heterogeneous cellular networks (HCNs). However, one of the main challenges in HCNs is the device association process in which the devices need to be attached to M. K. Elhattab is with Department of Electrical and Computer Engineering, Concordia University, Montreal, QC, Canada, H3G 1M8, e-mail m_elhatt@encs.concordia.ca. Mahmoud M. Elmesalawy, Fatty M. Salem, and I. I. Ibrahim are with Electronics, Communications and Computers Dept., Helwan University, Egypt, 11792. (E-mails: {melmesalawy, faty_ahmed, ibrahim.ismail}@h- eng.helwan.edu.eg). different types of base stations (BSs) existing in the HCNs [4]. In cellular networks, the conventional association method is dependent on the maximum signal-to-interference-plus-noise ratio (Max-SINR) criteria. Where the BS that provides the Max-SINR is always designated [4]. However, one of the main problems arises from applying the Max-SINR scheme on HCNs is the wide disparity in the transmit power for the different classes of BSs, which leads to an extremely unbalanced load distribution among the heterogeneous BSs. As a result, the overloaded BS cannot achieve the requested QoS of all associated devices due to having a limited number of resources; meanwhile, the lightly loaded BS cannot achieve the maximal efficiency because of the lower number of associated devices. Therefore, the user association strategy with the ability to take the benefits of the HCNs’ architecture is needed. Consequently, there are a lot of recent studies that have been done regarding the cell association in HCNs under various scenarios including [5]-[21]. A. Related Work The joint optimization of cell association and resource allocation is considered in HCNs [5]. The aim of this work is to maximize the network utility while considering both the BSs’ energy and backhaul constraints. In [6], the QoS provisioning of HCNs is taken into consideration for user association design. The authors investigate a unified framework for downlink (DL) data rate maximization and outage minimization, and then distributed user association algorithms are proposed. A message-passing framework is used to achieve load-balancing solutions [7]. The cell association in [7] aims at maximizing the sum-rate of the network, while guaranteeing the minimum device QoS of DL data rate. In [8], the authors investigate joint cell association and bandwidth allocation to maximize the DL sum-rate of the network. However, the out-of-band wireless backhaul is assumed that decreases the spectral efficiency of the network. An association strategy is developed in [9] to maximize the weighted sum-energy efficiency for massive multiple-input and multiple-output-enabled HCNs while achieving the users’ QoS in terms of the minimum SINR. A cell association scheme with QoS support and load balancing for HCNs is jointly proposed considering both the load level of each BS and the DL achievable rate for user equipment (UE) [10]. Then, a distributed user association algorithm based on the dual decomposition is developed. An approximation algorithm is developed in [11] to solve the intractable cell association problem efficiently, dependent upon which an efficient search procedure is presented to minimize the total power consumption of the HCNs by controlling the density of the small cells dynamically. In [12], Device-aware Cell Association in Heterogeneous Cellular Networks: A Matching Game Approach M. K. Elhattab 1 , Mahmoud M. Elmesalawy, Member, IEEE, Fatty M. Salem, and I. I. Ibrahim I