Electronic Thesis and Dissertation Repository

Thesis Format

Monograph

Degree

Doctor of Philosophy

Program

Electrical and Computer Engineering

Supervisor

Xianbin Wang

2nd Supervisor

Hongbo Zhu

Affiliation

Nanjing University of Posts and Telecommunications

Co-Supervisor

Abstract

Internet-of-Things (IoT), which aims to integrate machines, industrial process, and people together, has become one of the most critical technology for the next wave of industrial revolution. Among many related technical fields of IoT, machine-to-machine (M2M) communication is the most important paradigm to realize the seamless information exchange between different machine-type devices (MTDs) with minimal human interference and has experienced rapid development in the past few decades. However, features of M2M communication such as diverse applications, small data packets, and large connections, would cause quality-of-service (QoS) dissatisfaction, channel utilization deterioration and severe collisions to the random access procedure, which is adopted to establish initial connections between MTDs and the base station. To overcome these drawbacks, this thesis aims to design effective access traffic control and collision resolution methods for the massive M2M communication oriented random access procedure.

Towards the feature of diverse applications, a multi-group analytical framework for massive random access of M2M communication in IoT is proposed. Specifically, we consider delay-sensitive MTDs and delay-tolerate MTDs coexist in the network, and those MTDs are divided into multiple groups according to their delay requirements. The throughput and the mean access delay of each group of MTDs are characterized based on queue theory. Then, the access traffic control method is formulated as an optimization problem to maximize the throughput of delay-tolerate MTDs under delay constraints of delay-sensitive MTDs. Simulations show that the proposed access traffic control method can achieve the maximum throughput of both the whole network and the delay-tolerate MTDs while satisfy the delay requirements of delay-sensitive MTDs.

The channel utilization will decrease dramatically due to frequent signaling exchange, especially with the small data packets in M2M communication. To solve this drawback, two distributed access traffic control algorithms are proposed for homogeneous and heterogeneous scenarios, respectively. In particular, each MTD only needs to observe its own number of successful access requests and the total number of access request attempts within a certain estimation period, and the essential parameters to control the access traffic can be obtained from the observations. Simulations show that when the number of MTDs in the network does not change or varies slowly, the network throughput can be maximized if the estimation interval is chosen properly. Compared to existing centralized methods, the proposed method can achieve similar performance with much lower signaling overhead.

In addition, to improve the random access efficiency in massive M2M communication where the number of MTDs is huge, a deep neural networks based double-contention random access (DCRA) mechanism is proposed. In particular, the base station first adopts a deep neural network to detect random access collisions by learning the features of the received signals. Based on the collision-detection results, a DCRA scheme is proposed, with which the base station can schedule one more contention process for devices experiencing collisions to resolve these collisions. To fully harness the collision-resolution capability of the proposed DCRA scheme, we further analyze its performance, and illustrate how to control the access traffic to optimize the network throughput. Simulation results show that with a high collision recognition accuracy, the proposed scheme can achieve significant throughput improvement.

In the final part of this thesis, a hybrid channel access by admission and contention (HCAAC) scheme is designed to achieve the coexistence of real-time applications (RTA) and non-RTA in Wi-Fi networks while optimizing the delay performance of RTA. With this scheme, the base station can schedule the transmission of RTA packets and temporarily interrupt the transmission procedure of non-RTA packets to guarantee a quick channel access for RTA packets, the delay of non-RTA packets can also be improved without the contention from RTA packets. An optimization problem is formulated to illustrate how to control the access traffic to minimize the delay performance of RTA. Simulations show that with the proposed HCAAC scheme, the delay performance of RTA can achieve a dramatic improvement, while the performance of non-RTA does not deteriorate.

Summary for Lay Audience

Machine-to-Machine (M2M) communication is seen a key enabler of the Internet-of-Things (IoT) and has experienced tremendous development. However, diverse applications, small data packets, and large connections, of M2M communication bring new challenges, especially in the random access stage. In order to address these challenges, four M2M communication-oriented random access schemes are proposed to optimally control the access traffic and resolve collisions in the random access procedure.

First of all, to satisfy diverse requirements of different applications, a multi-priority random access analytical framework is proposed. Based on which the throughput and average access delay of each group can be obtained explicitly. Then, an optimization problem is formulated to both satisfy the requirements of diverse applications and maximize the throughput of the whole network.

Moreover, to reduce the signaling consumption during the random access, two distributive optimal access parameters determination algorithms are proposed towards homogeneous and heterogeneous networks, respectively. Based on which each machine-type devices (MTD) can aware the congestion of the whole network based on their history access experiences and can control the access traffic by themselves to reduce the signaling exchange between MTDs and the base station.

Additionally, to improve the efficiency of the random access procedure, a novel double-contention random access (DCRA) scheme is designed. With which the base station first detect the random access collisions based on a machine-learning method, then resolve collisions by arranging further contention chances for the collided MTDs.

Finally, to guarantee the delay performance of real-time applications (RTA), a hybrid channel access by admission and contention (HCAAC) scheme is presented. With this scheme, the base station can interrupt the transmission of non-RTA to guarantee immediate transmission of RTA. In addition, the non-RTA performance deterioration caused by the transmission of RTA packets can be alleviated by improving the retransmission efficiency of non-RTA packets.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS