Internet of Multimedia Things (IoMT) faces the challenge of how to realize low-cost data acquisition while still preserve data confidentiality. In this work, we present a low-cost and confidentiality-preserving data acquisition framework for IoMT. Firstly, we harness chaotic convolution and random subsampling to capture multiple image signals. The measurement matrix is under the control of chaos, ensuring the security of the sampling process. Next, we assemble these sampled images into a big master image, and then encrypt this master image based on Arnold transform and single value diffusion. The computation of these two transforms only requires some low-complexity operations. Finally, the encrypted image is delivered to cloud servers for storage and decryption service. Experimental results demonstrate the security and effectiveness of the proposed framework.

According to the DNA encryption algorithm and the double-chaotic system which contains the optical chaos and the coupled map lattice chaotic system, a novel image encryption-then-transmission system is proposed. In the system, with identical chaotic injection from a master laser with two optical feedbacks, two slave lasers (SL1 and SL2) can output similar chaotic signals served as chaotic carrier to transmit image and used to generate the core part of the encryption scheme. A 128bit key is selected to generate the original value of the double-chaotic system, which decides the DNA complementary rule. The security analysis demonstrates the effectiveness of the proposed encryption system. The simulation results verify that the cryptosystem is enough to against the traditional attacks such as statistical attack, differential attack, brute force attack and entropy attack. Moreover, the encrypted image can be the optical message and transmitted in 10 km single mode fiber channel from SL1 to SL2. In order to ensure the security, we use the chaos masking technique to modulate and demodulate the optical message. Through numerical simulations of the cross-correlation function, the chaos synchronization between SL1 and SL2 is desired. The Q-factor is 9.559 and the bit error rate is 5.771$\times$10$^{-22}$.

In this paper, the secrecy performance of a wireless-powered multi-antenna relaying system, which consists of a source $\boldsymbol {(S)}$ , a destination $\boldsymbol {(D)}$ , $\boldsymbol {N (N>1)}$ relays (Rs), a power beacon $\boldsymbol {(B)}$ , and an eavesdropper $\boldsymbol {(E)}$ , is studied in the presence of imperfect channel state information at Rs. Both S and Rs are assumed to be with nonlinear energy harvesters, and harvest energy from radio-frequency signals from B through a time-switching harvesting scheme. Assuming Rs deploy decode-and-forward scheme, a general $\boldsymbol {K}$ th best relay in terms of instantaneous secrecy capacity for S-R-D link is selected for information transmission. An analytical expression for secrecy outage probability (SOP), validated via Monte Carlo simulation, is derived when all channels experience independent Rayleigh fading. Numerical results show that a higher correlation coefficient between the estimated and real channel values leads to better SOP performance when the ratio of the channel gains of R-D and R-E links is in the high region for the first hop or in the low-to-medium region for the second hop. Furthermore, there exists a special value of the saturation threshold of the energy harvester at S or R, and SOP performance cannot be improved when the saturation threshold is beyond that value.

A plasmonic temperature-sensing structure, based on a metal-insulator-metal (MIM) waveguide with dual side-coupled hexagonal cavities, is proposed and numerically investigated by using the finite-difference time-domain (FDTD) method in this paper. The numerical simulation results show that a resonance dip appears in the transmission spectrum. Moreover, the full width of half maximum (FWHM) of the resonance dip can be narrowed down, and the extinction ratio can reach a maximum value by tuning the coupling distance between the waveguide and two cavities. Based on a linear relationship between the resonance dip and environment temperature, the temperature-sensing characteristics are discussed. The temperature sensitivity is influenced by the side length and the coupling distance. Furthermore, for the first time, two concepts—optical spectrum interference (OSI) and misjudge rate (MR)—are introduced to study the temperature-sensing resolution based on spectral interrogation. This work has some significance in the design of nanoscale optical sensors with high temperature sensitivity and a high sensing resolution.

Wireless communications can be improved by employing free space optical (FSO) channels. Since optical signals can only be transmitted via line-of-sight paths, UAVs are employed to forward data from a base station (BS) to remote users for urgent tasks using multi-hop mixed FSO/RF links. The UAVs employ the decode and forward protocol to relay data. The last UAV decodes and forwards the data to multiple users through RF links using non-orthogonal multiple access (NOMA). To improve fairness, a modified deep reinforcement learning (DRL) algorithm is used to optimize the transmit power allocation in real-time to minimize the maximum user decoding outage probability. Numerical results are presented to illustrate the system design tradeoffs. In addition, the validity of the proposed approach are verified by comparing it with exhaustive search algorithm.