論文

Deep joint source-channel coding (DeepJSCC) has attracted attention as a type of semantic communication that shares not only information but also meaning and intent, and it is a type of deep learning that uses an autoencoder instead of conventional source and channel coding. DeepJSCC with convolutional neural network (CNN) layers has been studied specifically for image transmission, where it avoids cliff effects and achieves higher image quality even in the low signal-to-noise power ratio (SNR) range. In addition, DeepJSCC is suit for services related to the Internet of things (IoT), and is likely to be widely applicable to monitoring-related applications such as surveillance cameras. For this reason, it may not be possible to provide sufficient GPU resources, especially for the transmitter where the encoder is installed. Therefore, we propose a fast Fourier transform (FFT)-based DeepJSCC that replaces CNN with FFT and element-wise product to reduce the computation time of DeepJSCC. This paper evaluates the impact of changing the layer structure on image quality.

Asynchronous Quick Link (A-QL) is the tri-color-band screen code specified as one of the transmission symbol formats in IEEE802.15.7. An edge detection algorithm is required to extract information from A-QL symbol. However, edge detection wastefully recognizes unwanted objects and it results in communication failure. To improve its detection performance, this paper proposes to apply YOLO as deep learning-based object detection and demonstrates its effectiveness.

Optical Camera Communication (OCC) is one of the types of visible light communication which enables low-cost and license-free communication by using general-purpose devices, such as LEDs or displays as transmitters and optical cameras as receivers. It is expected to have variety of applications. Its one of the challenges is constrained communication speed; limited by the frame rate of display and image sensors. Employing multi-level Color Shift Keying (CSK) is being considered as an effective approach to achieve higher data capacity. However, demodulation errors occur when the constellation of the received signal deviates from the reference point due to hue changes of the light source in the shooting environment or internal processing of the image sensor device. This paper proposes an equalization method that uses deep learning to compensate for received constellations and improve communication performance. A simple indoor experiment using LEDs as a light source demonstrates its fundamental effectiveness.

This paper proposes an autoencoder-based nonlinear quantizer to conduct IQ data compression for the Fronthaul link employing digitized radio over fiber (D-RoF). The numerical simulation reports a 6-bit nonlinear quantizer to satisfy the EVM requirement.

This demonstration illustrates broadband and high-efficiency OFDM-based massive MIMO transmission aimed at stabilizing and increasing the capacity of underwater acoustic communication (UWAC). In UWAC, mitigating doubly selective propagation channels with long delay spreads and severe time variations remains a critical challenge. Conventional approaches have accommodated delayed paths at the cost of efficiency and have relied on complex signal processing to cope with Doppler variations. In our study, we allow inter-symbol interference by employing broadband transmission with shorter symbol durations, while suppressing its impact through the diversity gain of large-scale antenna arrays. At the same time, spatial filtering based on CSI separates multiple streams associated with dominant propagation paths.

Optical camera communication (OCC) provides license-free communication channels without radio waves. The rolling shutter mechanism is expected to enhance the capacity of wireless links in OCC. However, signal erasure occurs due to waiting times between frames in consecutive signal transmission. To address this issue, we propose a symbol timing synchronization for OCC using a rolling shutter camera. The receiver camera synchronizes to the transmitter with additional waiting time which is computed based on the defined preamble signals. We analyzed the performance to clarify the theoretical limit. The feasibility of the proposed idea was confirmed with experimental results.

This paper proposes an optical circuit-based delta-sigma modulator (ODSM) to achieve high-speed operation. The proposed ODSM uses optical devices and an electrical 1-bit analog-to-digital convertor. We report the impact of noise in the ODSM circuit and the optical link on the signal quality by numerical simulation.

The world is often stricken by catastrophic disasters. On-demand drone-mounted visible light communication (VLC) networks are suitable for monitoring disaster-stricken areas for leveraging disaster-response operations. The concept of an image sensor-based VLC has also attracted attention in the recent past for establishing stable links using unstably moving drones. However, existing works did not sufficiently consider the one-to-many image sensor-based VLC system. Thus, this paper proposes the concept of a one-to-many image sensor-based VLC between a camera and multiple drone-mounted LED lights with a drone-positioning algorithm to avoid interference among VLC links. Multiple drones are deployed on-demand in a disaster-stricken area to monitor the ground and continuously send image data to a camera with image sensor-based visible light communication (VLC) links. The proposed idea is demonstrated with the proof-of-concept (PoC) implemented with drones that are equipped with LED panels and a 4K camera. As a result, we confirmed the feasibility of the proposed system.

Optical Camera Communication (OCC) is a promising solution for long-range point-to-multipoint communication between drones and a ground camera. OCC requires line-of-sight (LoS) channels for a camera to receive optical signals transmitted from drone-mounted LED lights or panels. When multiple drones are deployed in a certain area to transmit optical signals simultaneously, inter-light interference avoidance is a significant issue. The inter-light interference has been modeled in the previous works. However, existing works have not sufficiently investigated the trajectory control of drones. To address this issue, in this paper, we propose a distributed trajectory control algorithm for drones to avoid inter-light interference. Based on the approximate interference model in the image plane, each drone controls its trajectory to ensure LoS communication links of other drones. The performance of the proposed algorithm is confirmed via intense multi-agent simulation. The proposed algorithm contributes to establishing stable point-to-multipoint OCC links between drones and a ground camera.

Deep learning-based image recognition systems have rapidly evolved. Due to the extensive processing load of the deep neural network (DNN) on graphic processing units (GPUs), the DNN model is deployed on the cloud server. Images or videos are forwarded from user terminals through the network to the server. In recent years, edge computing has gained popularity as a means of reducing the data traffic in the backbone network. However, the last one-mile access network between an edge server and user terminals will still be congested because a large amount of data such as video/image files must be forwarded. In particular, when computer vision applications such as image recognition are loaded in the edge network, a large amount of data is forwarded although the edge server always may not need the high-definition image. This paper proposes an image compression and progressive retransmission scheme for deep learning-based image recognition systems to reduce image data traffic and alleviate network congestion. The proposed method introduces an entropy-based threshold calculated from posterior probabilities from a deep learning model's output layer. Entropy is an extremely effective metric because it can be used as an indicator independent of the number of classification labels in the DNN model. The thresholding can control the image retransmission and reduce traffic while maintaining image recognition accuracy. We implement the proposed scheme on the edge server and reveal the relationship between the data compression and the recognition accuracy through simulation evaluation. As a result, we indicate that an entropy-based threshold reduces the overall ambiguity of the accuracy of image recognition. Moreover, when a higher accuracy recognition model with more accuracy is combined with a retransmission scheme, it becomes the more effective.