Vision–language models learn visual concepts from the supervision of natural language. It can significantly enhance the generalizability of real-time intelligent sensing, such as analyzing camera-captured real-time images for visually impaired users. However, adapting vision–language models to distribution shifts at test time, caused by several factors such as lighting or weather changes, remains challenging. In particular, most existing test-time adaptation methods rely on gradient-based fine-tuning and backpropagation, making them computationally expensive and unsuitable for real-time applications. To address this challenge, the Training-Free Dynamic Adapter (TDA) has recently been introduced as a lightweight alternative that uses a dynamic key–value cache and pseudo-label refinement for test-time adaptation without backpropagation. Building on this, we propose TDA-L, a new framework that integrates Low-Rank Adaptation (LoRA) to reduce the size of feature representations and related computational overhead at test time using pre-learned low-rank matrices. TDA-L applies LoRA transformations to both query and cached features during inference, cost-efficiently improving robustness to distribution shifts while maintaining the training-free nature of TDA. Experimental results on seven benchmarks show that TDA-L maintains accuracy but achieves lower latency, less memory consumption, and higher throughput, making it well-suited for AI-based real-time sensing.

In a rapidly flourishing country like Bangladesh, accidents in unmanned level crossings are increasing daily. This study presents a deep learning-based approach for automating level crossing junctions, ensuring maximum safety. Here, we develop a fully automated technique using computer vision on a microcontroller to reduce and eliminate level-crossing deaths and accidents. A Raspberry Pi microcontroller detects impending trains using computer vision on live video, and the intersection is closed until the incoming train passes unimpeded. Live video activity recognition and object detection algorithms scan the junction 24/7. Self-regulating microcontrollers control the entire process. When persistent unauthorized activity is identified, authorities, such as the police and fire brigade, are notified via automated messages and notifications. The microcontroller evaluates live rail-track data and arrival and departure times to anticipate ETAs, train position, velocity, and track problems to avoid head-on collisions. This proposed scheme reduces level crossing accidents and fatalities at a lower cost than current market solutions.

Non-intrusive load monitoring (NILM) is a methodology used to deduce the power consumption and operational status of electrical equipment by analyzing the aggregated power signal of a residential or commercial structure. Utilizing detailed electric load profiles obtained from smart meters has generated growing interest in employing this method for demand-side energy management within smart grids. NILM is a time series regression problem that involves appliance state classification, pattern recognition, and anomaly detection. Deep learning-based NILM techniques have shown promising results, but they are often applicable to a limited number of appliances due to their complex model architecture and huge computational cost. This paper proposes an improved method by using a shared convolutional encoder architecture for a multitask decoder network, adding multi-head attention blocks in between. The proposed approach outperforms state-of-the-art techniques in the publicly available REDD data set while requiring a reduced amount of training time.