In a crisis situation, it is expected that a user's mobile device may not be connected to the Internet or advanced AI services, and there is a need for technology that can be used even in such cases. In this paper, we demonstrate through computer experiments that a method using BLIP-2, a visual language technology, is promising in terms of accuracy as a method for locally determining whether a subject in a photo is related to a disaster. The proposed method focuses on embedding technology, which is a preprocessing method for LLM. When LLM is not available, we thought that using its vectors as the basis for processing is one of the optimal methods. BLIP-2 is used to vectorize the photo, and the judgment is made by comparing it with the built-in database using the k-nearest neighbor method. Computer experiments show that when an appropriate database is used, this method outperforms other methods in terms of accuracy and F1 value.

Methods for wirelessly providing power from an external source to sensors and actuators have been required in various scenes. In this research, we aim to transfer power from an external energy source using ultrasound. In particular, we attempt the sound-to-heat energy conversion, which is more straightforward than the electroacoustic transduction in specific applications. By attaching a thin layer of porous material to the heating target, the irradiated ultrasound is absorbed as heat by the sound absorption effect and transmitted to the target. In the experiment, urethane foams with different widths were compared to maximize the energy conversion efficiency from acoustic energy to heat. As a result, urethane foam with a width of 5mm could transmit 26% of the presented sound power to the heating target. In this paper, details and results of our method are introduced.

An important issue in permanent-magnet synchronous motors (PMSMs) for electric vehicle (EV) traction is estimating the temperature of the rotor permanent-magnet in real-time. This paper proposes a novel rotor flux amplitude estimation method to enable real-time rotor magnet temperature estimation of the PMSMs with flux dq-cross coupling. The proposed method has the following features: (a) A novel minimum-order D-flux state observer, included in the D-filter, is used for the real-time rotor flux-amplitude estimation in the PMSMs with flux dq-cross coupling. (b) The three inductances of d-axis, q-axis, and cross-coupling required for the D-filter are identified simultaneously using an adaptive parameter identifier. (c) The adverse effect caused by the current detection offset on the flux-amplitude estimate and the indentified inductances is overcome by utilizing the characteristics unique to each system. (d) The instability caused by the coupling of two real-time systems is addressed by adopting a special system structure that prevents such instability. (e) Using the novel method, the real value of the rotor flux-amplitude can be estimated in 0.9s after the zero initial-state start for the PMSMs with flux dq-cross coupling.

Control performance often deteriorates when conventional methods based on constant sampling intervals are applied to systems with varying sampling intervals, due to the mismatch between assumed and actual sampling characteristics. To address this challenge, we propose a novel design method for a Model Error Compensator (MEC) that mitigates the adverse effects caused by sampling interval variations. The MEC dynamically compensates for changes in the input-output characteristics of the controlled system, which occur due to aperiodic sampling, thereby enhancing the robustness and effectiveness of control methods designed for constant sampling intervals. Numerical simulations demonstrate that the proposed MEC significantly improves control performance in systems with aperiodic sampling by reducing model errors. Compared to conventional control approaches, the proposed method offers superior performance and contributes to the broader field of sampled-data system control by addressing the practical challenges associated with sampling variability.