This paper presents a project for converting an electric public transport bus into a hybrid bus. It begins with an overview of design solutions for electric and hybrid drives. Next, traction calculations for the electric bus are presented. Based on the traction calculations, the parameters that the new hybrid drive system will have to meet will be determined. The drive system (engine and gearbox) for the hybrid bus was selected. A design solution was proposed, along with a theoretical presentation of the control system for the new hybrid bus drive system. Finally, a summary and conclusions are provided, along with possible future directions for solutions regarding the conversion.

Student behavior in classrooms is an important parameter to monitor academic success of a course. Various parameters like class attendance, participation in class activity and response could be used to measure the engagement of a class. The presented study uses various AI models like decision trees, support vectors and ensemble methods in order to identify trends in class engagement using various scores like accuracy, F1 score, recall and precision. The dataset was divided into a split ratio of 80:20 ratio of training and split. Random forest model was found to achieve highest performance with accuracy of 91.8%, precision of 0.91, recall of 0.92 and F1 score of 0.91. These results indicate that ML models are highly competent to capture relationships between faculty and students in classrooms in order to identify students that are at potential risk. The novelty of the proposed framework lies in integrating of prediction of behavior based on machine learning pedagogical interpretation. It contributes towards development of educational objectives aligned with Industry 4.0 and digital transformation. All the data presented in this work was anonymized and handled according to ethical guidelines laid out within institutional framework.

This review is a thorough review on the rapid convergence of Agentic Artificial Intelligence (AI) and Machine Learning (ML) with advanced manufacturing to enable a fully autonomous "Text-to-Print" workflow. We examine the key technological pillars that bridge design intent with physical fabrication. In this review, Section 1 introduces the motivation for automation to overcome critical bottlenecks in the design-to-manufacture pipeline, specifically in Computer-Aided Design (CAD) modeling, process parameter tuning, and quality control. Section 2 provides a conceptual foundation for Agentic AI, reviewing how many agentic AI, that include Large Language Model (LLM)-based agents, multi-agent systems, and planning frameworks like Reinforcement Learning (RL) and ReAct (Reasoning then Acting) create a new era of autonomous, goal-driven manufacturing. Section 3 reviews the state-of-the-art AI-driven geometry generation, comparing procedural, parametric, and the emerging "Text-to-CAD" models (e.g., MEDA, CAD-Llama) that leverage CAD APIs as tools to increase efficiency and accessibility in CAD designing. Section 4, the core of this review, analyzes the role of ML in the industrialization of Additive Manufacturing (AM). We conduct a deep review of ML applications in process optimization, in material science, and in defect prediction, while focusing on the critical roles of Physics-Informed Neural Networks (PINNs) and Digital Twins in achieving industrial-scale quality and repeatability. Finally, Section 5 compiles these components into a complete "closed-loop learning factory" concept, analyzing end-to-end case studies of self-correcting systems and concluding with a critical assessment of the foremost challengesdata inter-operability, model explainability (XAI), and the ethical governance of autonomous design. This review additionally surfaces three key findings: Multi- Agent Systems (MAS) consistently outperform single-agent designs for complex, multi-domain manufacturing tasks; intelligent AM systems carry significant but largely untapped potential for reducing material waste and supporting sustainable manufacturing; and LLM-based agentic frameworks exhibit a cross-process adaptabilitytransferring reasoning across different AM processes with minimal reconfigurationthat narrow, task-specific ML models cannot easily match.

Human tremors are involuntary oscillatory motions resulting from complex interactions between neuromuscular activity and the mechanical properties of the musculoskeletal system. This work parents a multi - degree - of - freedom (MDOF) dynamic model for tremor analysis. The study begins with the modelling of the upper limb modeled as an unconstrained MDOF system to establish baseline behaviour. The model is further extended to the lower limb, specifically the shank - ankle - foot system, to account for its role in postural stability and gait dynamics. A key contribution is the incorporation of support- contact mechanisms, where the ground is modeled as a spring- damper system to simulate real world conditions. The governing equations are formulated using classical vibration theory and implemented in a simulation environment for analysis under different conditions. Comparative analysis is performed between upper limb, lower limb without contact, and lower limb with ground contact. The results indicate that ground interaction significantly influences tremor amplitude and frequency. The proposed framework provides a more realistic and generalized approach for tremor analysis. Index Termsmulti-degree-of-freedom system, tremor anal- ysis, biomechanical modeling, lower limb dynamics, support- contact mechanics, vibration analysis.

Climate change poses increasing risks to rainfed crop production in semi-arid regions of sub- Saharan Africa, where rising temperatures and climate variability threaten smallholder livelihoods. This study assessed climate risk to cowpea (Vigna unguiculata L. Walp) production in northeastern Ethiopia and evaluated management-based adaptation options using the DSSAT CROPGRO-Cowpea model. The model was calibrated and validated using field data and applied to simulate crop responses under historical conditions (19942023) and future climate scenarios (SSP2-4.5 and SSP5-8.5) for the mid-century (2050s) and late-century (2080s). Results indicate that under the high-emission SSP5-8.5 scenario, cowpea production faces substantial climate risk due to accelerated phenological development, increased yield variability, and heightened vulnerability of rainfed systems. Although elevated atmospheric CO partially enhanced biomass and yield through improved photosynthesis and water-use efficiency, CO fertilization alone was insufficient to offset climate risk under extreme warming. In contrast, optimized sowing dates and supplemental irrigation during critical growth stages significantly reduced yield losses and variability across future scenarios, including SSP5-8.5. Early sowing minimized exposure to terminal heat and moisture stress, while supplemental irrigation stabilized reproductive development and improved yield reliability. The combined application of early sowing and targeted irrigation consistently produced the lowest climate risk, outperforming reliance on physiological CO effects alone. These findings demonstrate that climate risk to cowpea production under future warming is manageable through practical, low-cost adaptation strategies. Integrating adaptive sowing calendars and small-scale irrigation into extension services and climate-smart agriculture policies can substantially enhance the resilience of cowpea-based farming systems in Ethiopia and similar semi-arid regions. Electrical Engineering and Electromechanics | ISSN: 2074-272X | | https://ieeejournal.org/ | Page No : 2 | Volume 26 - Issue 7 - 2026 |

This study investigates the compaction behavior and interfacial mechanics of bilayer powder systems (cementbentonite, cementCaCO, and bentoniteCaCO) using Discrete Element Method (DEM) simulations. The approach integrates a HertzMindlin contact model and Janssen-type stress attenuation. Simulations were performed in a rigid die under compressive loads up to 8 tons, evaluating varied particle sizes and layer sequencing. The simulations revealed distinct axial and radial porosity gradients. Interface normal stresses ranged from 2.5 to 3.8 MPa, with shear stresses between 0.3 and 0.6 MPa. Interfacial porosity decreased from approximately 0.40 in coarser layers to 0.30 in finer layers. The data indicates that smoother transitions and lower local shear stress promote stronger load transfer and mitigate delamination risks. Additionally, image processing was employed to quantify peak-to-valley interfacial roughness at 180270 m. While minor particle size variations had little effect on global compaction, they significantly impacted local stress and densification at the interface. Validated against existing literature, this coupled DEM and image-based framework accurately predicts stress distribution, porosity evolution, and junction stability. Ultimately, these findings provide critical insights for optimizing multilayer powder compaction in ceramics, pharmaceuticals, and other particulate-based industries where interfacial integrity is essential.

The concept of a fermatean set as first proposed by senapati and yager as the extension of intuitionistic fuzzy set and Pythagorean fuzzy set. In this paper, we studied the notion of a fermatean Q-fuzzy point to study for the first time the notion of fermatean Q-fuzzy subgroups and its local properties. This new concept showed us to reformulate all the mathematical properties of fermatean Q-fuzzy subgroups structure.

The Hydro-geochemical and Pollution Index Characteristics of Industrial Wastewater in Jos- Plateau were investigated with implications of using biomaterials as adsorbents. A total of eight (8) wastewater samples were collected with two (2) samples each from four (4) industries within Jos Metropolis namely: Plateau State Water-board, Nasco Household limited, Grand Cereals and Stanel bakery. The biomaterials used as absorbents include groundnut shells, sweet potato peels and maize Cob. The ash of the biomaterials, wastewater and treated water samples were subjected to Atomic Absorption Spectrometry for heavy metals determination. The values of heavy metals in the wastewater and treated water were subjected to data analysis to determine the pollution index parameters and the values heavy metals were compared with WHO standards. The levels of heavy metals such as Ni, Cd, Cr and Pb in the wastewater exceeded the WHO permissible limits for drinking water. The treated water samples with biomaterials show significant reduction in heavy metals concentrations and the levels of heavy metals such as Cr, Pb, Cu and Mn fall below permissible limits. The nemerow pollution index (PN) value (0.386- 2.197) in the treated water with groundnut shell seems to be most effective adsorbents and classified as unpolluted and slightly polluted. The PN value (0.933-5.305) in the treated water of sweet potato peels was classified as slightly and moderately polluted. The treated water of maize cob ash shows the least adsorbent capacity with PN value (1.702-7.284) classified as slightly, moderately and heavily polluted. The water quality index (WQI) of wastewater (471.174- 8786.773) is unfit for consumption. The WQI of treated water using groundnut shell (35.009- 223.538) was categorized into excellent, good, poor, very poor and unfit for consumption. The WQI contents of treated water using maize cob (277.221-2,528.796) was categorized into very poor and unfit for consumption. Based on WQI, groundnut shell ash proves to be the most effective adsorbents having excellent water quality in sample VII and good water quality in samples V and VIII. The second most effective adsorbent is sweet potato peels with good water quality in sample V.

Facilities management, as a core component of physical asset management, plays a critical role in preserving building performance, mitigating hazards, and improving the quality of built environments. This study investigates the effect of facilities management on users safety and sense of security in buildings, and clarifies the mediating role of the performance of building physical systems. The research is applied in purpose and employs a quantitative descriptive-survey design. The population comprised residents, users, and managers of residential, office, commercial, and mixed-use buildings; 256 participants were selected via convenience sampling. Data were collected using a researcher-developed questionnaire. Exploratory factor analysis assessed construct validity, and partial least squares structural equation modeling (PLS- SEM) tested the conceptual model. Factor analysis yielded a KaiserMeyerOlkin (KMO) measure of 0.912 and a total explained variance of 58.85%, indicating an adequate factor structure for the instrument. Structural model results indicated that facilities management has a positive and significant effect on the performance of building physical systems ( = 0.809, p < 0.001). The performance of physical systems also had a positive and significant effect on users safety and sense of security ( = 0.446, p < 0.001). Additionally, facilities management exerted a direct and significant effect on users safety and sense of security ( = 0.479, p < 0.001). Findings show that the performance of physical systems partially mediates this relationship (indirect effect = 0.361, p < 0.001). These results suggest that implementing facilities management systems, expanding preventive maintenance, and continuously monitoring equipment can enhance safety, reduce operational risks, and improve users sense of security.

The sustainable management of public healthcare facilities is a critical challenge in todays 12 healthcare delivery systems, driven by increasing demands, limited resources, and the growing 13 need for environmentally and technologically responsible practices. This study seeks to explore 14 and formulate a strategic framework that ensures the long-term sustainability of public healthcare 15 facilities. The study focuses on optimizing resource efficiency, improving facility maintenance, 16 reducing environmental impacts, and integrating emerging technologies in healthcare facility 17 management. The study employs a mixed-methods approach, combining qualitative and 18 quantitative data collection. A sample frame for the study includes healthcare facility managers, 19 sustainability experts, healthcare professionals, and patients. An in-depth interview was 20 conducted with 8 facility managers to gather qualitative insights on current practices and 21 challenges in facility management. Additionally, focus group discussions were held with 22 healthcare workers to explore operational perspectives on sustainability. To complement this, a 23 quantitative survey was administered to 195 respondents, including healthcare staff and patients 24 from 4 public healthcare facilities in Akure Nigeria, to assess facility performance, user 25 satisfaction, and the integration of sustainability measures. Statistical analysis, including 26 descriptive statistics and regression analysis was used to explore the relationships between 27 sustainability practices and facility performance, resource efficiency, and user satisfaction. 28 Findings revealed that those facilities that have adopted energy-efficient technologies, efficient 29 water management, and effective waste management protocols are seeing positive outcomes, 30 including cost savings, improved patient satisfaction, and reduced environmental impact. 31 Similarly, the study revealed that many public healthcare facilities are aware of the importance of 32 sustainability, but financial and operational constraints limit their ability to implement 33 comprehensive strategies. The study therefore recommends increasing awareness, enhancing 34 policy frameworks, and exploring funding mechanisms such as public-private partnerships as 35 critical to overcoming the identified challenges and achieving sustainable management of public 36 healthcare facilities. 37

communication between human brain and external devices without muscular involvement. This work presents how a game can be controlled using electroencephalography (EEG) signals. In this work, EEG signals are collected with the use of a kit known as DIY Neuroscience kit consisting of an Ardunio Uno R4 Minima and BioAmp EXG pill with gel electrodes. The recorded signals are pre- processed using a band pass filter (1-45 Hz) and a notch filter of 50 Hz to remove unwanted noise and distortions. The signals after pre- processing are distributed into small overlapping segments. From each segments, several time-based and frequency-based features are obtained to record important patterns of brain activity. After completion of Pre-processing, Support vector Machine (SVM) and Linear Discrimination Analysis (LDA) is used to train the model, so it can distinguish between attentive and relaxed state. The result shows EEG Signals can be used to control video games. The Study also shows that low-cost EEG hardware combined with machine learning techniques can be used to make more brain-computer systems.