88 research outputs found
Performance Analysis of Different Loss Function in Face Detection Architectures
No full textMasked face detection is a challenging task due to the occlusions created by the masks. Recent studies show that deep learning models can achieve effective performance for not only occluded faces but also for unconstrained environments, illuminations or various poses. In this study, we have addressed the problem of occlusion due to wearing masks in masked face detection technique in deep transfer learning method. We have also reviewed the recent deep learning models for face detection and considered VGG16, VGG19, MobileNet and DenseNet as our underlying masked face detection models. Moreover, we have prepared a dataset containing masked face and without mask from 120 individuals and enhanced the dataset using augmentation. After training the deep learning models with our own dataset, we have analysed the performance of the deep learning models for several types of loss functions. From the experiment, it is clear that all the deep learning models perform well in terms of classification losses like categorical cross entropy loss and KL divergence loss
Impact of Li-Fi in 6G : challenges, applications, and prospects
No full textWith wireless communication technologies still developing, integrating light –fidelity (Li-Fi) with current and future cellular networks offers a viable way to improve data transfer performance. This paper examines Li-Fi’s performance in conjunction with 5G and speculates its possible use in upcoming 6G networks. The advantages, disadvantages and capabilities of Li-Fi, such as its reliance on the visible light spectrum for data transmission and its possible benefits, which include increased security and high data transfer rates, are also discussed. Further, we look at how Li-Fi and 5G networks can function together, followed by the impact of 6G on light fidelity. Using simulation-based assessments and empirical research, we compare Li-Fi and Wi-Fi performance to prove the reliability and security of Li-Fi
A theoretical basis for brain waves with implications for a large scale integration required for cognitive processes
No full textA Technique to Minimize the Effect On Resonance Frequency Due to Fabrication Errors of MS Antenna by Operating Dielectric Constant
No full textEnergy-Efficient Straight Robotic Assembly Line Using Metaheuristic Algorithms
No full textThis paper focuses on the implementation of metaheuristic algorithms to solve straight robotic assembly line balancing problem with an objective of maximizing line efficiency by minimizing the energy consumption of the assembly line. Reduction in the energy consumption is of high importance these days due to the need of creating environmental friendly industries and also due to the increase in the cost of energy. Due to the availability of different types of robots in the market, there is a necessity of selecting efficient set of robots to perform the tasks in the assembly line and optimizing the efficiency of its usage in the line effectively. Two well-known metaheuristic algorithms: particle swarm optimization (PSO) and differential evolution (DE) are implemented to solve due to the NP-hard nature of the problem. Proposed algorithms are tested on the benchmark problems available in the literature and the detailed comparative results are presented in this paper. It can be seen that proposed DE algorithm could obtain better results when compared with PSO from the experimental study.</p
Classification of Functional Grasps Using Hybrid CNN/LSTM Network
No full textGestures made by a human can be classified using Electromyography (EMG) signals collected from the forearm; even with low-frequency devices. Numerous steps are required from data collection and pre-processing through to final classification. Traditionally, an important part of EMG signal classification is extracting features from the raw signal to reduce dimensionality. It is predominantly carried out manually before the signals are input into a neural network. In this research, we successfully used a CNN to extract the features automatically, and an LSTM layer was utilised to classify the gestures. This network architecture removes a step in the gesture classification process. Using the raw signals input into a CNN/LSTM hybrid increased classification when compared with an LSTM network that required features to be manually extracted from the raw signals.</p
Self-Balancing Mobile Robot with Bluetooth Control: Design, Implementation, and Performance Analysis
No full textThis paper presents a comprehensive study of an ESP32 microcontroller-based self-balancing mobile robot system designed in conjunction with an Android app for Bluetooth control. The robot employs an MPU6050 accelerometer/gyroscope to execute dynamic equilibrium control for robotic balance. This study explores the design of a system composed of an ESP32-based dual-platform architecture. The firmware for the ESP32 executes real-time motor control and sensor processing, while the Android application provides the user interface, data visualization, and command transmission. The system achieves stable operation with tilt angle variations of ±2.5° (σ=0.8°, n = 50 trials) during normal operation with a PID controller tuned to KP = 6.0, KI = 0.1, and KD = 1.5. In experimental tests, control latency was measured at 38–72 ms (mean = 55 ms, σ=12 ms) over distances of 1–10 m with a robust Bluetooth connection. Extended operational tests indicated the reliability of both autonomous obstacle avoidance mode and manual control exceeding 95%. Key contributions include gyro drift compensation using a progressive calibration scheme, intelligent battery management for operational efficiency, and a dual-mode control interface to facilitate seamless transition between manual and autonomous operation. Processing of real-time telemetry on the Android application allows visualization of important parameters like tilt angle, motor speeds, and sensor readings. This work contributes to a cost-effective mobile robotics platform (total cost: USD 127) through the provision of detailed design specifications, implementation strategies, and performance characteristics
Intelligent Terrain Mapping with a Quadruped Spider Robot: A Bluetooth-Enabled Mobile Platform for Environmental Reconnaissance
No full textThis paper introduces a new quadruped spider robot platform specializing in environmental reconnaissance and mapping. The robot measures 180 mm × 180 mm × 95 mm and weighs 385 g, including the battery, providing a compact yet capable platform for reconnaissance missions. The robot consists of an ESP32 microcontroller and eight servos that are disposed in a biomimetic layout to achieve the biological gait of an arachnid. One of the major design revolutions is in the power distribution network (PDN) of the robot, in which two DC-DC buck converters (LM2596M) are used to isolate the power domains of the computation and the mechanical subsystems, thereby enhancing reliability and the lifespan of the robot. The theoretical analysis demonstrates that this dual-domain architecture reduces computational-domain voltage fluctuations by 85.9% compared to single-converter designs, with a measured voltage stability improving from 0.87 V to 0.12 V under servo load spikes. Its proprietary Bluetooth protocol allows for both the sending and receiving of controls and environmental data with fewer than 120 ms of latency at up to 12 m of distance. The robot’s mapping system employs a novel motion-compensated probabilistic algorithm that integrates ultrasonic sensor data with IMU-based motion estimation using recursive Bayesian updates. The occupancy grid uses 5 cm × 5 cm cells with confidence tracking, where each cell’s probability is updated using recursive Bayesian inference with confidence weighting to guide data fusion. Experimental verification in different environments indicates that the mapping accuracy (92.7% to ground-truth measurements) and stable pattern of the sensor reading remain, even when measuring the complex gait transition. Long-range field tests conducted over 100 m traversals in challenging outdoor environments with slopes of up to 15° and obstacle densities of 0.3 objects/m2 demonstrate sustained performance, with 89.2% mapping accuracy. The energy saving of the robot was an 86.4% operating-time improvement over the single-regulator designs. This work contributes to the championing of low-cost, high-performance robotic platforms for reconnaissance tasks, especially in search and rescue, the exploration of hazardous environments, and educational robotics
- …
