Raghu Engineering College

In this research work, we present QL-ChASO-FOPID, which is a combination of Q-learning, chaotic atom search optimization (ChASO), and fractional-order proportional integral derivative (FOPID) control, which together find applications in regenerative braking of electric vehicle having multiple motors. This has three components: a Q-learning module with a 32-state space, which adjusts the FOPID gains in accordance with vehicle speed deviation and battery state; a torque-vectoring controller to maintain vehicle stability while performing split-μ braking; and a battery management module that also assists in defining maximum charging current depending on battery condition. The application of offline optimization alongside online learning leads to the development of a braking controller characterized by stable performance under different driving conditions. Through simulations of regenerative braking, it has been shown that the composite of QL-ChASO and the FOPID control provided additional advantages, including a decrease of 7.4% in the integral of the time-weighted absolute error (ITAE) and of 2.6% in the cost of energy handled by the DC bus, both relative to a conventional proportional–integral–derivative (PID) controller. The torque-vectoring system achieved a reduction of peak yaw-rate error of 70.2% in split-μ braking tests. These results show that combining online learning with torque vectoring improves both braking performance and vehicle stability.

Graphene oxide (GO) which is a derivative of graphene, has garnered considerable attention due to its remarkable mechanical, thermal and electrical properties. In this study, GO nanoparticles were synthesized using ball milling method with graphite powder as the precursor potassium permanganate as oxidizing agent for inducing structural modifications. The synthesized graphene oxide was characterized by scanning electron microscopy (SEM) and X-ray diffraction (XRD) to confirm the structure, morphology and its functional groups. Subsequently water-based graphene oxide nanofluids of distinct concentrations (1, 3 and 5

Human Action Recognition (HAR) is a significant research area in computer vision with applications in surveillance, healthcare, and human–computer interaction. This study proposes a hybrid deep learning framework that integrates the lightweight MobileNetV1 architecture with the scalable EfficientNetB3 network to achieve a balance between computational efficiency and recognition accuracy. The hybrid design leverages the complementary strengths of MobileNetV1 for efficient low-level feature extraction and EfficientNetB3 for enhanced high-level feature representation, enabling improved discriminative capability while maintaining reduced computational complexity. The model is evaluated on the UCF101 dataset, comprising 13,320 video clips across 101 action classes. A frame-based classification strategy with uniform sampling is adopted to reduce computational overhead, and data augmentation techniques such as rotation, shifting, shearing, and brightness adjustment are applied to improve generalization. Experimental results demonstrate that the proposed model achieves an accuracy of 89.18

This study examines the tribological and mechanical properties of Al7075 aluminium alloy reinforced with silicon carbide (SiC) and titanium diboride (TiB2) particles using the stir casting process. Reinforced composites showed improved density, hardness, and wear resistance, with TiB2 offering superior performance. Microhardness increased linearly with reinforcement content, and TiB2 composites achieved up to 6.67

Approximate computing offers a promising paradigm for energy-efficient hardware design by trading minimal precision for substantial gains in power and performance. This paper introduces the Modified Dadda Approximate Multiplier (MDAM), an innovative architecture that optimizes hardware economy, reduces power consumption, and delivers enhanced accuracy for applications including convolutional neural networks (CNNs) and image processing. Standard multipliers employed in CNNs are computationally intensive and energy-demanding, limiting their deployment in edge devices and real-time scenarios. The proposed MDAM leverages approximate computing techniques to balance accuracy with reduced power consumption, area, and processing time while maintaining acceptable performance for CNN tasks. Through its modified structure, MDAM achieves significant complexity reduction by minimizing partial products and streamlining carry propagation phases. The design incorporates error compensation mechanisms, truncation strategies, and a novel 4:2 Approximate Compressor (AC) to optimize the performance-error trade-off. Extensive experiments across multiple CNN models for image classification and object detection tasks validate the effectiveness of this approach. Both 8-bit and 16-bit versions of existing approximate multipliers (AMs) were implemented in Verilog and synthesized using MATLAB, Xilinx Vivado, and Cadence RTL Compiler for comprehensive performance evaluation. Comparative simulations demonstrate MDAM's superiority across key design metrics: up to 37.5