A Physical Layer Security Approach For Indoor Visible Light Communications

Visible light communication(VLC)systems are gradually entering people's lives,but the communication security problems at the same time cannot be ignored.The traditional mechanism of secure communication is implemented in the upper layer of the network,which is effective only when the eavesdropper has limited storage capacity and computing power.In contrast,the physical layer secure communication(PLSC)method only makes use of channel characteristics to hide information from unauthorized persons.Common PLSC methods include artificial noise,multi-input single-output(MISO)technology,etc.Among them,MISO technology utilizes the enormous spatial freedom provided by the LED arrays to protect communication security.However,MISO technique also has the problem that the transmission signal is not easily optimized and the main lobe cannot be adjusted.For the indoor VLC with MISO technology,the secure communication problem is abstractly modeled as a constrained optimization problem,and the group intelligent optimization algorithm is used to solve the difficulty of signal optimization.The traditional method of protecting indoor VLC security usually needs to know the accurate position of the eavesdroppers in advance,and the feasibility is low.This article defines a circular insecure area centered on the target receiver and assumes that all eavesdroppers are outside this unsafe area.Using the power adjustment factors(PAFs),the distribution of optical power can be adjusted through controlling the direct channel gain of each LED,and the PAFs of all LEDs are integrated into a vector as a solution of the group intelligent optimization algorithm.The difference between the maximum signal strength received by the eavesdroppers and the received signal strength of the target receiver is taken as the objective function,which converges with the iteration.The solution vector obtained at the completion of the iteration is the optimal set of PAFs,and the main lobe is controlled toward the target receiver.Thus,the signal strength received by the eavesdroppers is much lower than that of the target receiver,which achieves the purpose of secure communication and compensates the lack of the insufficient adjustment of main lobe in multi-input single-output technology.According to the different definition domain of PAFs,two PLSC schemes are further proposed.Scheme I uses the improved continuous cuckoo search algorithm to solve the continuous optimal PAF vector with continuous definition domain.All LEDs are at “ON” state,and the PAF factor is used to optimize the distribution of optical power.To further improve the performance of the cuckoo search algorithm,the fixed parameters controlling position updating are modified to adaptively adjust with the iteration,which can also improve localconvergence ability of the cuckoo.In the simulation,the improved cuckoo search algorithm is compared with the standard one to verify its advantages.Compared with the bat algorithm,invasive weed optimization algorithm,particle swarm optimization algorithm and firefly algorithm,the improved cuckoo search algorithm exhibits a better convergence speed and convergence level.Scheme II uses the discrete biogeography-based optimization algorithm to solve the optimal PAF vector with a component value of 0 or 1,in order to optimize the layout of the LED arrays.The optimized biogeography-based optimization algorithm is more efficient than the cuckoo search algorithm and invasive weed algorithm.The optimization effect of Scheme II is further improved by combining Scheme I: the optical power distribution of the LED with the PAF of 1 is continuously optimized using the continuous group intelligent optimization algorithm after finishing the optimization of the LED layout.Simulations show that the improved Scheme II achieves the same PLSC level as Scheme I.