In the edge computing, the communication between edge computing nodes and terminal devices is vulnerable to impersonation attacks, therefore a physical-layer impersonation attack detection algorithm based on Deep Q-Network (DQN) was proposed. Firstly, an impersonation attack model was built in the edge computing network, a hypothesis test based on the physical-layer Channel State Information (CSI) was established by the receiver, and the Euclidean distance between the currently measured CSI and the last recorded CSI was taken as the test statistics. Secondly, for the dynamic environment of edge computing, the DQN algorithm was used to adaptively select the optimal test threshold with the goal of maximizing the gain of the receiver. Finally, whether the current sender was an impersonation attacker was determined by comparing the statistics with the test threshold. The simulation results show that the Signal-to-Interference plus Noise Ratio (SINR) and channel gain ratio have certain effect on the performance of the detection algorithm, but when the relative change of channel gain is lower than 0.2, the false alarm rate, miss rate and average error rate of the algorithm are less than 5%. Therefore, the detection algorithm is adaptive to the dynamical environment of edge computing.

Aimming at the problem that the traditional physical layer security algorithm does not make full use of the channel,a new physical layer channel detection algorithm was proposed. In view of the essential properties of 4G wireless channel, combined with the hypothesis testing, Support Vector Machine (SVM) was used to analyse the metrics of channel vector to decide whether there are counterfeit attackers or not. Simulation experiments show that the accuracy of the proposed algorithm based on linear kernel is more than 98%, and the accuracy of the proposed algorithm based on Radial Basis Function (RBF) is more than 99%. The proposed algorithm can make full use of the wireless channel characteristics of different spatial locations to implement authenticaton of information source one by one, and hence enhances the security of the system.

Memory dependence prediction plays a very important role to reduce memory order violation and improve microprocessor performance. However, the traditional methods usually have large hardware overhead and poor realizability. Through the analysis of memory dependence's locality, this paper proposed a new memory predictor based on instruction distance. Compared to other memory dependence predictors, this predictor made full use of memory dependence's locality on instruction distance, predicted memory instruction' violation distance, controlled the speculation of a few instructions, finally deduced the number of memory order violation and improved the performance. The simulation results show that with only 1KB hardware budget, average Instruction Per Cycle (IPC) get a 1.70% speedup, and the most improvement is 5.11%. In the case of a small hardware overhead, the performance is greatly improved.