Research On Sensing Data Falsification Attack And Defense Method In Cognitive Vehicular Network

With the advent of intelligent transportation and driverless technology,the shortage of spectrum resources in the vehicular network has become increasingly prominent.Cognitive radio technology can wait to find and use the idle spectrum,so it is introduced into the vehicular ad hoc network,generating the cognitive vehicular network.Spectrum sensing is the basis for discovering and accessing idle frequency bands.The distributed cognitive vehicular network spectrum sensing process mainly relies on sensing data interaction between neighbor nodes to obtain unified sensing results.Accurate sensing data is an important guarantee for the reuse of spectrum resources without disturbing the primary users.Attackers can influence the decision-making and allocation of spectrum by falsifying the spectrum sensing data to occupy the idle frequency band alone or interfere with the other users.In the existing distributed network spectrum sensing process,the researches on spectrum sensing data falsification attack rarely consider the mobility of attackers.At the same time,it mainly uses the historical data of neighbor nodes to detect abnormal data,which is not applicable to dynamic topology of cognitive vehicular network.As an important feature of the cognitive vehicular network,high-speed mobility of nodes can lead to rapid changes in network topology and network fragmentation.For distributed cognitive vehicular network,the high-speed mobility of nodes has an impact on the data fusion results and convergence time of distributed spectrum sensing algorithm,and also provides more opportunities for attackers.Meanwhile,the defense schemes based on historical data of neighbor nodes may have serious performance problems due to rapid changes in the set of neighbor nodes.The high-speed mobility of nodes makes the cognitive vehicular network face many challenges.In this thesis,the network topology change rate and the performance of distributed spectrum sensing algorithm in distributed cognitive vehicular network are analyzed.Meanwhile,we propose a new spectrum sensing data falsification attack based on mobility of vehicles and defense scheme for this attack.The main contents are as follows:(1)Through theoretical analysis and simulation verification,the relationship among the topological change rate of vehicular network with vehicular speed,wireless signal coverage,time and initial distance between vehicles is analyzed,and the overall characteristics of network topology change rate are characterized.Then,we make curve fitting of relationship between the main influencing factors and the change rate of network topology.(2)The performance of distributed spectrum sensing algorithm in high dynamic cognitive vehicular network is verified in both efficiency and accuracy.Based on this algorithm,combined with the mobility characteristics of cognitive vehicular network nodes,a new spectrum sensing data falsification attack based on speed adjustment of vehicles is proposed.Meanwhile,the effectiveness of this attack is analyzed in terms of attack efficiency and defense difficulty.The simulation results show that the proposed attack can reduce the number of iteration rounds,and effectively improve the efficiency of collusion attack.At the same time,it can increase the defense difficulty through the rapid change of neighbor nodes.(3)Considering the dynamic characteristics of cognitive vehicular network,for the proposed spectrum sensing data falsification attack based on vehicular speed adjustment,a dynamic update weighting scheme based on node data difference and neighbor nodes changes is explored.Meanwhile,the defense effect of the scheme is verified,and the simulation results show that the weighting scheme can complete the spectrum sensing process quickly and accurately in the presence of different attack modes.