Reservoir computing (RC) has attracted much attention in recent years as a model suitable for machine learning of time series information and requiring low computational resources. Among these, physical reservoir computing (PRC), in which RC is implemented in hardware devices, has been actively studied using various devices as it is applicable to edge computing.

In this paper, we construct a reservoir computing (RC) system using Pulse-type Hardware Chaotic Neuron Model (P-HCNM), which can be implemented in integrated circuits, and perform a benchmark task. A second order Nonlinear Auto Regressive Moving Average (NARMA) task, which is widely used as a benchmark task for RC, is evaluated using Normalized-Mean-Square-Error (NMSE), and the result is 0.0456. The score is comparable to other PRCs in previous studies. In the Delay task that evaluates short-term memory, the recall results of the RC showed a forgetting curve. The result indicates that RCs have short-term memory, which is a necessary component for RCs. Using these benchmark tasks, we suggest that a P-HCNM-implemented neural network can be the hardware basis for a neuromorphic reservoir computing system.

In analog computation, there are very few types of activation functions. In this study, we created a circuit that generates the difference between the Swish and the ReLU activation function, and aimed to mimic the behavior of the Swish function by combining ReLU and the difference. We obtained outputs close to the difference between the swish function and the ReLU function in simulation. We also confirmed that the sum of ReLU and the output works as activation function in neural networks on software.

The prototype of Rotating Polarization Wave (RPW) radio is proposed. The prototype consists of two evaluation boards of commercial Radio Frequency (RF) chips, an evaluation platform of general-purpose microprocessor, and the 90-degree hybrid coupler. These boards are categorized into Software Defined Radio (SDR) that is performed by embedded C language programs. A RPW communication system using the proposed prototype radios achieves to enhance not only quality but also range of communication in various wireless environments by developing a novel embedded program on a free c programming language platform.

Atherosclerosis is a key factor in lifestyle-related diseases, characterized by near-irreversible arterial thickening and hardening. Early detection is crucial for effective intervention, however current wearable technology lacks non-invasive arterial stiffness assessment capabilities. This study explores the feasibility of using Photoplethysmogram (PPG) at the radial artery for arterial stiffness evaluation, focusing on green-light PPG. PPG signals were collected from 18 healthy young participants (9 males: 23.3±3.7 years; 9 females: 24.7±7.3 years) at the fingertip and wrist. The second derivative PPG (SDPPG) obtained from fingertip with green and near-infrared light and that from radial artery with green light was analyzed. Preliminary results showed that SDPPG aging indices obtained from fingertip with green and near-infrared light were highly correlated in 4 participants, suggesting its feasibility for vascular age estimation. However, at the radial artery, individualized vascular age estimation models may be required. Further studies are needed to validate this approach and establish standardized evaluation criteria.

Objective and convenient monitoring of health status is increasingly important for maintaining and enhancing well-being. Recent studies have attempted to estimate physiological and psychological states using facial thermal images, focusing on blood flow controlled by the autonomic nervous system. However, these studies often overlook the mechanisms underlying skin temperature formation. Skin temperature distribution is expected to exhibit spatial autocorrelation due to heat transfer between adjacent regions via blood flow. The objective of this study is to extract health-related features from facial thermal images by considering spatial autocorrelation. Facial thermal images linked to health questionnaire responses were analyzed to identify regions where spatial autocorrelation varies with health status. The results showed that poorer health correlated with lower spatial autocorrelation values around the eyes and nose, and a significant reduction in overall spatial autocorrelation. These findings suggest that spatial autocorrelation in facial thermal images may serve as a useful indicator of health status.

In this paper, we propose a hierarchical control method for the motion of transportation robots with the aim of solving problems related to the transportation of cargo in the logistics/transportation industry by introducing transportation robots. Specifically, multiple robots simultaneously achieve transportation of cargo to a target position, avoidance of collisions with obstacles, and coverage of the area in a two-dimensional region with obstacles. At first, the administrator specifies a transportation or covering task for each robot as a global control. Then, one robot that is assigned a transportation task carries out the transportation of cargo while avoiding collisions with obstacles by local tangent bug-based motion control. On the other hand, the remaining robots without the transportation task continue to cover the region. Finally, the effectiveness of the proposed control method is verified by numerical simulations assuming a train station and the other for a shopping mall.

To reduce traffic fatalities, it is effective to detect driver's cognitive distraction. In this study, to detect driver's cognitive distraction, we focus on the difference between the measured gaze points and the visual attention maps indicating the objects to be gazed at estimated from the deep learning model. In order to verify the effectiveness of the proposed method, the BDD-A dataset of in-vehicle camera videos is used. As a result, it is confirmed that the proposed method can detect driver's cognitive distraction with high accuracy.

Continuous hidden Markov model with Gaussian mixture model as the emission distribution is prominent in modelling sequential data. It is well known that Gaussian distribution confronts the bottleneck of sensitivity to outliers. However, student’s t-distribution is heavily tailed and more robust. Therefore, the student's t-distribution based hidden Markov model (SHMM) is useful for analyzing the sequential data with outliers. In this paper, the dynamic texture is considered as a set of the observed data sequences generated with an SHMM model. The motion and appearance properties of the dynamic texture are characterized with the hidden state sequence and the emission distribution, respectively. Then the parameters of the model SHMM can serve as the features to describe the dynamic texture. Finally, we use the maximum likelihood criterion to categorize the dynamic texture to be classified. The classification experiments on DynTex dataset and UCLA dataset demonstrate the superiority of the proposed method.

In depth estimation for 360-degree images, it is common to unfold the 360-degree images into planar images using the projection methods such as Equirectangular Projection (ERP) and Cubemap Projection (CMP) before estimating the depth. The projected images using each projection method exhibit unique distortions, which complicate depth estimation. To address this, some methods combine both projection methods to mitigate the effects of both distortions. However, in such methods, to realize the high depth estimation performance, an additional image processing (Spherical Padding) is required, which need to pay high computational cost. To solve this problem, in this research, we propose a novel depth estimation method with Gnomonic Projection (GP), which does not require the additional image processing and provides the projected images having less distortions compared to CMP. In the proposed method, we employ GP instead of CMP to reduce the computational cost and combine ERP and GP to improve the depth estimation performance. Extensive experiments demonstrate that our proposed method not only achieves higher depth estimation accuracy but also delivers faster estimation times.

This letter proposes the concept of a multifunctional line trace machine focusing on the use of sensors, which are important for the realization of IoT systems. As an example of the proposed system, we designed and developed a line trace machine based on an automotive concept. The effectiveness of the proposed system was demonstrated through demonstration experiments.