Font Size: a A A

Research On Random Mobile Agent Based Wireless Sensor Network Model

Posted on:2015-07-29Degree:DoctorType:Dissertation
Country:ChinaCandidate:T YangFull Text:PDF
GTID:1108330473956023Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
With versatile, low cost, low power consumption and self-organization characteristics, Wireless Sensor Network(WSN) has been widely used in real life. The technology has great impact on national economy and military fields. WSN gradually became the access point of cyber world to the physical world. It changes the way people interacts with the world, and impacts on human activities such as the way of dealing with the problems and perspective to understand the world.WSN applications still have many key technical bottlenecks waiting to be solved, they have been attracted attention from many scholars in this decade. Mobile Agent technology has several advantages in WSN. In addition to reduce network traffic overhead, mobile agent can carry, migrate, fusion data independently and transmite data after network recovered from failure. The mobility of mobile agent can simulate the behavior of group intelligent optimization algorithms, and the scale of WSN has less impact on the performance of distributed mobile agents. So, compared with the client/server model, mobile agent provides a more promising solution for WSN applications.This dissertation presents a Random Mobile Agent-based Wireless Sensor Network(RMAWSN) model to organically integrate the mobile agent and the group intelligent techniques into WSN. Through the mobility provided by the agent, this model uses random self-activate mobile agent to simulate swarm intelligence for event monitoring and optimizes network coverage. The main contributions of this paper include the following aspects:Firstly, in the literatures, sink nodes process itinerary planning for mobile agent, and then distribute them to the remote areas. To reduce the energy consumption in this procedure, this dissertation proposes RMAWSN: Mobile agents are randomly activated at local monitoring area and then migrate randomly to cover the area. This model was validated by two aspects: Firstly, this model proves energy consumption is less than that of the other two models, i.e., the Mobile agent-based WSN model,(MAWSN) and Mobile agent-based distributed sensor network model(MADSN), especially in the case of large scale WSN. Secondly, developed event-detection probability of local randomly-moving nodes, and proved that by trading time for space, the proposed strategy of random nodes deployment and random nodes movement needs less sensor nodes to cover the monitoring area than the strategy using fixed nodes. Consequently, the proposed strategy can reduce the network energy consumption and extend the network serve period.RMAWSN defines a variety of process migration patterns for mobile agent, enable sensor nodes to search access and return the data autonomously. RMAWSN also establish basic travel itinerary, and provide related algorithms.Secondly, based on the RMAWSN model, this dissertation proposed a new algorithm, i.e., the Short-life Artificial Fish Swarm Algorithm(SLAFSA) to dynamically cover an area by monitoring only the regions being visited by targets. SLAFSA uses artificial fish cruising within wireless sensor networks to effectively detect and track targets. According to the random mechanisms of RMAWSN, SLAFSA is designed to have short-cycle attributes and random behavior, in order to cover a larger area with less active nodes. Simulation results show that SLAFSA can achieve better network coverage with less active nodes and shorter communication range.To further reduce the number of active nodes in the network, we design the chasing be-havior for artificial fish in SLAFSA. Based on the chasing behavior, two adjacent artificial fish can perform pair-wise sensing, which generates larger coverage area than using individual sensors(i.e., individual artificial fish). Finally, based on the autonomous itinerary planning mechanism of RMAWSN, SLAFSA defines the foraging behavior for the artificial fish to drive them to approach the targets. As a result, the artificial fish can detect the events quickly. Experimental results show that this method can achieve higher coverage probability but with fewer active nodes.Third, as a key parameter, the distance between two nodes needs to be measured when performing pair-wise sensing in SLAFSA. By exploring the correlation between the Round-Trip Time(RTT) and the distance, this parameter can be estimated based on the IEEE 802.11 g RTT and the ZigBee LQI(Link Quality Indicator) data.This dissertation presents RTT time measurement from driver layer and LQI-based ranging technique. With minimal modifications, difference of RTT data is distinguished by Euclidean distance(ED) which provides distance information without additional hardware or topology changes.In ZigBee communication, experimental data show intrinsic error in LQI data at certain distance. With a novel LQI-based ranging technique by pre-correction, error compensation and mixed regression analysis, experiment proved that proposed technique achieved better ranging accuracy.
Keywords/Search Tags:random mobile agent, wireless sensor network, coverage, distance measurment, measurement short life aritficial fish swarm algorithm
PDF Full Text Request
Related items