This book shares valuable insights into high-efficiency data transmission scheduling and into a group intelligent search and rescue approach for artificial intelligence (AI)-powered maritime networks. Its goal is to highlight major research directions and topics that are critical for those who are interested in maritime communication networks, equipping them to carry out further research in this field. The authors begin with a historical overview and address the marine business, emerging technologies, and the shortcomings of current network architectures (coverage, connectivity, reliability, etc.). In turn, they introduce a heterogeneous space/air/sea/ground maritime communication network architecture and investigate the transmission scheduling problem in maritime communication networks, together with solutions based on deep reinforcement learning. To accommodate the computation demands of maritime communication services, the authors propose a multi-vessel offloading algorithm for maritime mobile edge computing networks. In closing, they discuss the applications of swarm intelligence in maritime search and rescue.
Author(s): Tingting Yang, Xuemin (Sherman) Shen
Series: Springer Briefs in Computer Science
Edition: 1
Publisher: Springer
Year: 2020
Language: English
Pages: 86
Preface
Contents
1 Introduction
1.1 Mission-Critical Applications and Services at Sea
1.2 Challenges to Maritime Communications
1.3 Our Contributions
References
2 Background and Literature Survey
2.1 SDN-Based Maritime Heterogeneous Networks
2.2 Mobile Edge Computing
2.3 Search and Rescue Under Maritime Communications
2.4 Summary
References
3 Intelligent Transmission Scheduling Based on Deep Reinforcement Learning
3.1 Software-Defined Maritime Communication Networks
3.1.1 Channel State Model
3.1.2 Cache State Model
3.1.3 Energy Consumption Model
3.2 Markov Decision Processes
3.2.1 System State Transition
3.2.2 System Reward Function
3.3 Software Defined Network Deep Q-Learning Algorithm for Data Transmission Scheduling
3.3.1 Optimal Channel Allocation Decision Based on MDPs
3.3.2 Improved Deep Q-Learning
3.4 Simulations of S-DQN Algorithm
3.4.1 Key Points: State Transition Process Simulation
3.4.2 Algorithm Performance Simulation and Comparisons
3.5 Summary
References
4 Multi-vessel Computation Offloading in Maritime Mobile Edge Computing Network
4.1 Multi-vessel Computation Offloading
4.1.1 Computation Offloading Technology
4.1.2 Offloading Judgement
4.2 Minimize Time Delay and Energy Consumption
4.2.1 Time Delay
4.2.2 Energy Consumption
4.2.3 Optimization Target
4.3 Optimal Energy Consumption Algorithm
4.3.1 Improved Hungarian Algorithm
4.3.2 Optimal Energy Consumption
4.4 Simulations of Different Scenarios
4.4.1 Different Scenarios Comparisons
4.4.2 Performance and Comparisons
4.4.3 The Saturation Time
4.5 Summary
References
5 Mission-Critical Search and Rescue Networking Based on Multi-agent Cooperative Communication
5.1 Model of Multi-agent Search and Rescue
5.1.1 Sea Search and Rescue Process Based on Multi-agent
5.1.2 Multi-agent Collaborative Networking
5.2 Establishment of Temporary Communication Network
5.2.1 Planning of Route to Reach Search and Rescue Area
5.2.2 Search Planning in the Wrecked Area
5.2.3 Establishment of Temporary Communication Network
5.3 The Improved Swarm Intelligence Algorithms
5.3.1 Classical ACO Algorithm
5.3.2 The Improved ACO Algorithm
5.3.3 Optimizing Packet Scheduling Order in Node
5.4 Simulations of the Maritime Search and Rescue Mission Algorithms
5.4.1 Shortest Path Planning
5.4.2 The Fastest Traversal in the Region
5.4.3 Forwarding Sequence Scheduling of Data Packets Within a Node
5.5 Summary
References
6 Conclusions and Future Directions
