Al-Khwarizmi Engineering Journal

Early diagnosis and effective rehabilitation of cerebral palsy (CP) are essential for improving functional outcomes and reducing long-term complications. CP affects the ability to perform daily physical activities and is often associated with secondary health problems such as obesity and chronic pain. Owing to the complexity of the condition, rehabilitation typically requires continuous supervision by clinical specialists, which can be challenging in environments with limited resources. Recent advances in artificial intelligence (AI) have created new opportunities for enhancing CP healthcare. This review summarises the major applications of AI in diagnosis, clinical decision support, motion classification and rehabilitation systems. AI-based diagnostic tools—including medical-image analysis using MRI and CT—support the early detection of neural abnormalities. Motion-classification systems use physical activity data, functional motor scales and gait analysis features to detect deviations and evaluate treatment progress. In rehabilitation, AI is increasingly integrated into robotic systems, virtual reality environments, video game-based training and metaverse-based therapies, enabling adaptive and engaging therapeutic experiences. This study highlights the shift towards AI-enhanced interventions, discusses the importance of human–computer interfaces for improving interaction between patients and rehabilitation systems and outlines current limitations and challenges. Continued research is required to improve data quality, model generalisability and clinical integration of AI technologies in CP care.

Accurate detection and classification of plant diseases are central to sustainable food production and the reduction in crop losses. Conventional identification protocols are often dependent on experts. They are time-consuming and difficult to scale. This paper proposes a novel Advanced Hybrid Convolutional Neural Network (AHCNN) model combining the attention mechanism and simplified convolution mechanism. This model can obtain high accuracy and low computational complexity. It is trained using a high-resolution leaf dataset that includes 1532 leaf images, separated into 1322 training, 150 validation and 60 test samples, with the goal of a classification of specimens into three categories: healthy, powdery and rusty. Its architecture uses squeeze-and-excitation (SE) blocks and a spatial attention mechanism, which, together, can provide enhanced feature extraction and improve the interpretability of models. Results showed that the model achieved a validation accuracy of 98% and a test accuracy of 98.33%. With a parameter number of only 0.4 million, the proposed architecture provides a lightweight solution that performs much better than conventional deep learning frameworks in terms of computational efficiency. These attributes make AHCNN an interesting candidate for real time, drone-based detection of plant diseases as part of precision agriculture systems

Polylactic acid (PLA) is a widely used biomaterial in additive manufacturing (AM) because of its biodegradable and biocompatible nature and availability. Hence, this study focuses on the importance of PLA in biomedical applications such as personalised medicine, tissue engineering, drug delivery and customised implants and explores the possibility of PLA-based biomedical products. Different AM technologies, such as fused deposition modelling, stereolithography, and selective laser sintering, and their possible uses with PLA are presented. Apart from pure PLA, the use of composite materials based on PLA is important, as these materials can improve mechanical and biological properties, which are highly desirable for various clinical applications. Several critical issues, such as material properties, resolution, regulatory affairs, biodegradability and sustainability, are examined and discussed. In addition, the potential of PLA-based 3D printing technologies, including multimaterial 3D printing and bioprinting for biomedical applications, is explored

This research tackles major real-time detection obstacles such as changing lighting effects and object blocks and background elements and different object sizes because such problems occur frequently in autonomous driving and surveillance and smart infrastructure applications. A detection pipeline system with modular functionality enabled processing of images, videos and webcams in real-time and had specific optimizations for each input format. The YOLOv5s model was selected due to its accurate and fast performance characteristics so we deployed it in a cloud-based Google Colab system with GPU acceleration capabilities. Real-world data collection succeeded in quantitative analysis through assessment of inference duration together with frame speed and detection precision along with confidence values while qualitative methods measured box precision and label validity. The system produced exceptional results by processing images and video data within 28 to 35 milliseconds and webcam frames between 1.8 to 2.3 seconds while generating confidence scores between 0.70 and 0.93. Real-time applications benefit from this system because it presents stable detection while being environmentally flexible and practically applicable. YOLOv5 proves robust based on the discovered test results which indicate future potential deployments of intelligent visual monitoring systems across all dynamic environments

In this connection, this paper proposes a Deep Q-Network (DQN) approach to address the dynamic nature of the job-shop scheduling problem. Dynamic scheduling requires an efficient and reliable algorithm for handling disruptions such as machine breakdowns and job priority changes. When comparing DQN with traditional approaches such as genetic algorithm (GA) and PSO, the latter algorithms cannot cope with the problems. On the contrary, DQN gains knowledge from its experiences within the factory and develops a strategy for solving the scheduling problem through minimizing makespan and maximizing machine utilization. Moreover, using experience replay (ER) and target networks enables DQN to maintain stability and develop an optimal schedule. Empirically, it was found that DQN reduces makespan by 24.8% with machine utilization being 92%. From the results obtained, it can be noted that the learning parameters play a great role in determining the performance of the model. Thus, this study proves that DQN is an effective approach for addressing the issue under discussion and could also be used for developing other approaches such as multi-agent and double DQN

Enhanced traffic management and safety can be achieved using decentralised networks such as Vehicular Ad Hoc Networks (VANETs), which allow communication amongst vehicles. These networks support dynamic, real-time communications; however, they face difficulties in terms of scalability, security and network stability. Improved vision efficiency, cost reduction and enhanced connectivity in VANETs can be realised using multi-RSU and clustering methods, revealing real-world benefits. This research seeks to improve communication efficiency in VANETs by addressing issues such as network delay, power usage and communication cost. The goal is to improve network reliability in a dynamic vehicular setup with fixed Roadside Units (RSUs), ensuring constant connectivity and reduced overhead. Spectral Efficiency and Hybrid Multi-RSU Deployment with Cost Optimisation (SEHMR) approach introduces an innovative solution for improving vehicular communication. The system integrates clustering, scheduling and continuous cluster maintenance of vehicles. By analysing the initial population and probability values of vehicles, the Cost Optimisation (DEA) algorithm effectively optimises costs. RSUs are strategically positioned to improve coverage, communication and power efficiency. This approach is evaluated through extensive simulations using NS2 and the SUMO mobility suite, measuring performance through metrics such as efficiency, throughput, packet delivery ratio, data loss, end-to-end delay and overhead. The proposed SEHMR-VANET is compared with basic approaches such as HGAR-VANETS, ROOP-VANET and ISFF-VANETs. Results indicate that SEHMR-VANET achieves higher energy efficiency and lower delay and data loss, revealing its effectiveness in improving traffic communication

Floating photovoltaic (PV) systems enable solar energy generation on water bodies, alleviating land scarcity whilst providing potential cooling and reflection benefits that enhance panel performance. However, many floating installations employ flat PV modules without optical concentration or integrated tracking, which limits their energy yield in tropical environments such as rivers and lakes. This study designs a floating solar platform that incorporates a curved PV configuration mounted on a buoyant platform, a single-axis sun-tracking mechanism and an IoT-based weather and performance monitoring system designed for riverine deployment. This prototype employs an 18 V, 250 W monocrystalline flexible solar panel formed into a curved trough that is controlled by an Arduino-based tracking system and a NodeMCU-based monitoring subsystem to record voltage, temperature and humidity via a mobile application. Experimental testing on the Damansara River indicates that the floating curved system achieves a maximum power output of 231±5 W, which is higher compared with the output of a flat-land-based configuration (190±7), corresponding to an efficiency enhancement of approximately 21% at a solar radiation intensity of 1200 W/m². The floating configuration also exhibits reduced efficiency degradation with increasing ambient temperature due to the cooling effect of the water body and elevated humidity, which together mitigate thermally induced performance losses. These findings demonstrate that the proposed curved floating solar platform provides a feasible and environmentally friendly solution for increasing PV energy yield whilst conserving land resources in tropical regions

This research clarifies the most important principles under which wall-climbing robots operate. Two important aspects must be examined in the study of this type of robot. The first aspect is the adhesion method, which has been analysed extensively over the years with multiple approaches. The second aspect is the movement method, which may or may not be integrated with the adhesion method. Research published in the last two decades has divided these robots into five main types and another subtype, each of which provides high efficiency in a specific environment and application. However, the efficiency decreases in others. For example, a robot with suction cups needs a smooth surface, a robot with ducted fans requires a surface that is free of holes, and a robot with magnetic adhesion needs a metal surface. Two robots, one utilising bioinspired adhesion and the other relying on rope and/or rail grabs, function on different surfaces and perform separate applications. This study determines the most important and effective designs, along with their strengths and weaknesses. The suitable applications and environments of these robot are also presented.

Tin oxide is a promising anode for the oxidation of refractory organic compounds, with advantages of low cost and absence of secondary pollutants. In this study, a Cu/SnO2-Sb2O5 anode was fabricated by cathodic deposition in a rotating system. The effect of electrode rotation speed on the structural properties of the synthesised anode (Cu/SnO2-Sb2O5) was investigated in the range of 50–250 rpm. The structure and morphology of the prepared anode were examined via SEM and XRD. Results showed that the Cu/SnO2-Sb2O5 anode prepared at 250 rpm possessed a compact, multi-layer coating structure without cracks. The electrochemical performance of the Cu/SnO2-Sb2O5 anode was studied by examining its activity in degrading methylene blue (MB) at a concentration of 100 mg/L by anodic oxidation at 20 mA/cm2 for 4 h. The anode prepared at 250 rpm was able to degrade MB with an efficiency of 93.6%, which was higher than that obtained by the anode prepared at 50 rpm (85.18%). The pseudo-first-order rate constant at 250 rpm was 0.01258 min-1, which was 1.6 higher than that observed for the anode prepared at 50 rpm, confirming the importance of rotation in improving mass transfer and facilitating the deposition of tin oxide with high catalytic activity. The Cu/SnO2-Sb2O5 anode fabricated at 250 rpm exhibited an enhanced service life of 18 h (200 mA/cm2; 0.5 M NaOH), whereas that prepared at 50 rpm had a reduced service life of 12 h. Compared with the Cu/SnO2-Sb2O5 anode prepared by thermal decomposition, the present anode exhibited a service life of 1.5 times greater than that of the thermal method, confirming the efficacy of the electrodeposition approach in fabricating the Cu/SnO2-Sb2O5 anode

Mesoporous alumina is a promising drug carrier because of its extensive surface area, pore structure, biocompatibility and non-toxicity, facilitating high drug loading and release capacity. This study aimed to investigate the adsorption and release kinetics of levofloxacin (LF) on prepared γ-mesoporous alumina nanoparticles (γ-MNPs). Adsorption kinetic tests were carried out by a batch method, drug loading was investigated using an impregnation method, and the release kinetics were determined using a dialysis bag maintained at 310 K. Adsorption kinetics were analysed using three kinetics models: the pseudo-first-order, pseudo-second-order and intraparticle diffusion. Results showed that the adsorption of LF followed a pseudo-second-order reaction, with the adsorption rate and mechanisms regulated by two consecutive steps, suggesting that intraparticle diffusion was involved, but it was not the only rate-controlling step. Drug loading results showed that the amount of LF loaded into γ-MNP was 2.17 mg drug/g γ-MNP. The drug release behaviour was investigated at pH 7.4 and pH 5.4. The drug release profile for the LF/γ-MNP system showed that the drug release percentages were 18.00% at pH 5.4 after 48 h and 90.41% at pH 7.4 after 24 h. Drug release was characterised by the kinetic models, including zero-order and first-order kinetics. The data from the pH 5.4 system was best fitted by zero- and first-order kinetics. The drug release mechanisms from a matrix were interpreted using Korsmeyer–Peppas, Kopcha and Higuchi models. The results demonstrated that drug release at pH 7.4 exhibited non-Fickian diffusion, whereas that at pH 5.4 was Fickian diffusion. The Hugichi constants (kH), which were correlated with the release rate, were low in the two systems, confirming the results of zero- and first-order equations. Therefore, γ-MNP is a promising and effective delivery medium for LF treatment