Intelligent Routing In Wireless Ad Hoc Network

With the development of the Internet of Things and Ubiquitous Network, wireless Ad Hoc network become a key technology in wireless communications. The increasing requirements of network intelligence make new requests on the intelligence of routing protocols. The new intelligent routing protocols not only need to be self-awareness, self-learning, self-decision-making, self-configuration capability, but also need to be able to combine with service model, modulation, power control, spectrum sensing technology for joint optimization. This dissertation focuses on the design of routing algorithms and routing mechanisms, aiming at optimizing the QoS requirement of users, energy efficiency and load balance. Optimization theory, multi-objective decision-making, machine learning, intelligent search algorithm are employed as academic rudiment to research on the intelligent routing of wireless Ad Hoc networks.Supported by several project of National High Tech Research and Development Plan of China, National Natural Science Foundation of China, National International Science and Technology Cooperation Project with Sweden and Canada, the dissertation aims at the intelligent mechanism of the routing protocols, exploiting new solving methods.The main contributions of the dissertation can be described as follows:(1) The dissertation investigates a framework of intelligent routing based on OODA innovatively, which is aim at the modeling of intelligent routing. The intelligent routing protocol can make full use of self-cognitive informations adaptively. It is also able to complete the self-configuration process, at the same time, it has good robustness, load balancing capabilities, QoS guarantee capabilities. By the method of cross-layer design, based on cognitive radio and cognitive packet technology, it can actualize the collection, interaction and harvest the information of network status and node status, then it obtain a routing information data-base. With the self-learning ability, it can filter, reasoning, judge and update the information, then construct knowledge data base of routing. It can obtain the network optimization goals different from Network structure, user class and service types, and against the goal choose different routing policy adaptively from multiple of optimization strategies. According to the optimization results, it completes the routing process base on self-organizing and self-configuration.(2) The dissertation proposes a reinforcement learning mechanism to learn and predict the time-varying characteristics of wireless channels and the status of the network. The mechanism completed the learning and predict of the environment through the interaction with the environment with a low routing overhead. First, the dissertation studies the method of reinforcement learning and Q-learning, then design the multicast routing protocol and opportunistic routing protocol based on Q-learning. The design of multicast routing protocols brings Q-learning into the field of multicast innovatively. It can complete the multicast transmission with QoS guarantees under fewer transmit opportunities. The design of opportunistic routing protocol introduces the Q-learning into that field innovatively. With the ability to adapt to the dynamic environment of Q-learning, it can be more adaptable to the fast change scenarios under the premise of ensuring the high-performance of the transmission rate and delay.(3) The dissertation designs the green routing mechanism based on the joint optimization and simulated annealing algorithm to meet the needs of green communications. The dissertation first completes the survey of green communications and energy efficient routing, and then it proposes a joint modulation, power control and routing cross-layer optimization innovatively. It can obtain the most energy-efficient modulation, transmit power and transmission path by solving the optimization problem of optimization objectives of energy saving and flow constraints. Finally, focusing on the problem that the network topology is too large and the nodes are too dense, the dissertation proposes a routing search method based on simulated annealing innovatively. The result shows that the mechanism can obtain better data transmission performance with fewer computing resources.(4) For the application of cognitive radio in wireless Ad Hoc network, the dissertation design a collaborative spectrum sensing algorithm and a load balancing routing protocol considering the spectrum sensing results. In cognitive radio scenarios, the dissertation designs a distributed spectrum sensing algorithm based on channel estimation and incentive innovatively. The algorithm improves the detection probability under the premise of ensuring the false alarm radio. Based on the collaborative spectrum sensing algorithm, the dissertation designs a Multi-policy routing considering the spectrum stability, load balancing, QoS guarantee performance innovatively. The routing adjusts the weight of routing metrics dynamically to improve the routing efficiency in the cognitive radio environment.(5) Facing the demand for physical experiments of intelligent routing technology, Set up a testbed with embedded development board, desktops, laptops, wireless Ad Hoc network. Follow the OODA intelligent routing framework, we design self-cognitive module, self-learning module, self-decision making module and self-configuration module in this testbed. Through the experiment, we verify the feasibility of the intelligent routing framework based on OODA. And the testbed provide the test sites and the experiment basis for further improvement of other research.In summary, this dissertation researches the intelligent routing protocol in-depth. The researche fruits explore the design of new intelligent routing for wireless Ad Hoc network, the works even have meaning to the whole field of communications.