검색 상세

Learning-Based Anti-Jamming and Routing Misbehavior Detection for Ad Hoc Networks

초록/요약

Concerning ad hoc network properties, the implementation of some complex security systems with more computing resources appears troublesome in most circumstances. Thus, the usage of anomaly and intrusion detection systems has drawn considerable attention. Detection systems are achieved as host-based (each node) or network-based (cluster head). These implementations exhibit advantages and drawbacks. For example, when cluster-based is used alone, it faces keeping protection when nodes delay to choose or substitute a cluster head. Notwithstanding different heuristics that have been introduced, there is still room for improvement. This dissertation proposes a detection system that can run as host- or as cluster-based to detect routing misbehavior. The detection operates on datasets built using proposed routing information-sharing algorithms. The detection system uses supervised learning to train when previous network status or exploratory network is available. Otherwise, it uses unsupervised learning. The testbed is stretched to evaluate the effects of mobility and network size. The simulation results show promising performance, even with limiting factors. The performance of the proposed detection system relies on neighboring nodes' communication. This communication can be heavily affected, at the data-link layer, by the presence of a jammer. In this dissertation, we analyze the efficiency of using a single-task reinforcement learning algorithm to mitigate jamming attacks with frequency hopping strategy. Our findings show that single-task learning implementations do not always guarantee a better cumulative reward. Hence come the possibilities of using multi-task learning instead. Multi-task reinforcement learning provides room for performance improvement to single-task learning when the tasks are related and learned with mutual information. Therefore, to improve the communication despite a jammer's presence, we propose deep multi-task conditional and sequential learning (DMCSL), a multi-task learning algorithm that develops a transition policy to resolve conditional and sequential tasks. The tasks are sensing time and transmission channel selection. DMCSL is a composite of state-of-the-art reinforcement learning algorithms, multi-armed bandit, and an extended deep-Q-network. To ensure the convergence and optimal cumulative reward of the algorithm, DMCSL is proposed with a continuous update algorithm for the sensing time action-space. The simulation results show that DMCSL with logarithmically increased action space guarantees better performance than single-task learning.

more

목차

1. Introduction 1
1.1 Machine Learning 4
1.1.1 Machine Learning for Routing Misbehavior Detection 4
1.2 Reinforcement Learning 7
1.2.1 Multi-Task Reinforcement Learning 9
1.2.2 Reinforcement Learning for Jamming Mitigation 10
1.3 Notations and Organization 11
2. Background Study 14
2.1 Routing Misbehavior Attacks 18
2.1.1 Routing Misbehavior: Attack model 19
2.1.2 Anti-Routing Misbehavior: Related works 20
2.2 Jamming Attacks 24
2.2.1 Jamming: Attack model 26
2.2.2 Anti-Jamming Systems: Related works 27
2.2.3 Anti-Jamming as an RL Problem 35
2.2.4 Single-Task RL Anti-Jamming Performance 38
3. Data-Link Layer Jamming Mitigation 42
3.1 Conditional and Sequential Tasks 45
3.2 Baseline Models 46
3.2.1 Multi-Armed Bandit 47
3.2.2 Deep Q-Network 49
3.3 Deep Multi-Task Conditional and Sequential Learning Algorithm 52
3.3.1 Action-Space Update Condition 58
3.3.2 Action-Space Update Algorithm 62
3.3.3 Action-Space Increase Analysis 63
3.3.4 Complexity Analysis 68
3.4 DMCSL Simulation Setup 70
3.5 DMCSL Anti-Jamming Simulation Performance 72
3.5.1 Static Jammer Time-Slot Duration 73
3.5.2 Dynamic Periodic Jammer Time-Slot Duration 74
3.5.3 Dynamic Random Jammer Time-Slot Duration 75
4. Network Layer Routing Misbehavior Detection 77
4.1 Routing Information-Sharing Algorithms 79
4.1.1 Neighbors Route Cache Sharing Algorithm 80
4.1.2 Cluster Route Cache Sharing Algorithm 82
4.1.3 Potential Attacks on RISA 83
4.2 RISA-based Detection Systems 85
4.2.1 Supervised LIDS . 86
4.2.2 Supervised LIDS Algorithms 88
4.2.3 Unsupervised LIDS 91
4.2.4 Unsupervised LIDS Algorithms 92
4.3 LIDS Simulation Environment Setup 95
4.3.1 LIDS Setup . 95
4.3.2 LIDS Supervised Model Selection 97
4.4 RISA Overheads Analysis 100
4.4.1 RISA Memory Overhead 100
4.4.2 RISA Bandwidth Overhead 103
4.5 LIDS Anti-Routing-Misbehavior Performance 105
4.5.1 Formerly Supervised Detection 105
4.5.2 Exploratory Supervised Detection 109
4.5.3 Unsupervised Detection 109
4.5.4 Increase in Malicious Nodes 111
4.5.5 Uncooperative Malicious Nodes 112
4.5.6 Packet Loss Effect 113
4.5.7 Comparison of Detection Schemes 113
5. Final Remarks 117
5.1 Limitations 117
5.2 Conclusions 119
Bibliography 120
Appendix A. Publications 131
Appendix B. Action-Space Update Example 132
Nomenclature 135

more
반출 Meta View 목록