Online learning platforms have gained immense popularity in recent years, offering learners a wide range of course options. However, the vast number of available courses can make it difficult for learners to find the most relevant and suitable courses for their needs and interests. To address this challenge, a deep learning-based approach for personalized course recommendations is proposed, leveraging sentence embeddings and user information to enhance the relevance and accuracy of recommendations. The method utilizes the Multilingual Universal Sentence Encoder (MUSE) to generate dense vector representations of course content, effectively capturing semantic relationships between courses. These embeddings are then combined with user-specific features, such as level of education, year of birth, and geographical location, to create comprehensive user profiles. Two state-of-the-art recommendation models, Convolutional Sequence Embedding (Caser) and Session-Based Recommendations with Recurrent Neural Networks (GRU4Rec), are employed and trained on the edX dataset to predict the most relevant courses for each user. Experimental results demonstrate that the Caser model, incorporating embeddings and user information, achieves a Precision@5 of 0.190, Recall@5 of 0.897, and a Mean Average Precision (MAP) of 0.635. The GRU4Rec model, with embeddings and user information, obtains a Precision@5 of 0.185, Recall@5 of 0.876, and a MAP of 0.611. These findings show the improved performance of the models, resulting in more accurate and personalized course recommendations for users.

This research is design and development an educational game for learners to gain knowledge in image processing topics while playing the game. The game is created to combine educational content with fun playing elements and enhance strategic thinking skills. This makes it an effective tool for both formal and informal learning. The user assumes the role of a robot named Enigmo in the secret town and has five missions to complete within that place and return safely. The missions consist of the sampling and quantization quest, the light and color adjustment quest, the depth of field quest, the hue color space quest, and the lens distortion and camera calibration quest. Each mission contains information and knowledge related to the mission stored in the user’s journey book, which the user can open and review lessons at any time while playing the game. The game is designed in first- and third-person perspective in a 3D environment and developed using Unity3D on Windows operating system. The research trial tested 10 users playing the game for an average of 60 min. Each user takes the quiz before and after playing the game. Out of a full score of 20, the average scores before and after learning from this game were 11.30 points and 16.70 points, respectively. The experimental results show that users gained knowledge of image processing from this game. The experiment also evaluates the usability in terms of usefulness, ease of use, ease of learning, and satisfaction with the average scores of 4.60, 4.15, 4.73, and 4.43, respectively.

Biological pest control has a strong advantage in its non-chemical effects on the environment. In this study, a fixed-time synergetic control scheme for the biological pest control problems represented by the n-dimensional Lotka-Volterra model was proposed. The proof of stability shows that the proposed controller can regulate the biological pest control systems with the characteristic of fixed-time convergence. The performance of the proposed scheme was demonstrated through simulation studies. The simulation results show that the pre-specified bound of the settling time can be satisfied regardless of the initial conditions, confirming a desired fixed-time convergence characteristic. Moreover, unlike the existing control policy based on the sliding mode control, the control inputs of the proposed policy are free from chattering.

Aim: To build wounds volume(3D) and area(2D) measuring system and device. Background: The measurement of the wound depth has been troublesome due to difficulty fo the procedures, physicians mostly avoid inspecting the wound depth as it could cause wound inflammation and infection. Objective: To build a contactless device for measuring wound volume and develop the system to support the wound treatment process which offers precise measurement and wound healing progression. Methods: Build a machine to control and stabilize 3D-scanner over the wound using a servo motor and apply the image processing technique to calculate the wound's area and volume. Comparing the machine accuracy by using Archimedes's principle testing with various wound model sizes, made from folding clay and pork rinds. Results: The device and system generate an error value of less than 15% which is within a satisfactory level. Conclusion: Knowing the wound depth is vital for the treatment, direct contact to the wound area can cause inflammation, infection, and increase time to heal. This device will help physicians to get more insight into the wound and improve the treatment plan for the patients. There are certain limitations to be considered for future work. Firstly, different software components used in the image processing and estimation process could be integrated to enhance user experience. Secondly, it is possible to apply Machine Learning techniques to identify the wounded area on the wound image file.

Various industrial structures or machines mostly consist of different shapes of ferromagnetic curvature surfaces. The magnetic wheel climbing robot is the suitable approach for achieving both adhesion and locomotion of the inspection robot. However, the adjustable magnetic force for robot adhesion is necessary, especially when the thickness of the surface is not uniform or the variation of the air gap between the magnetic adhesion units caused by the curvature of the surface. This can lead to the insufficient adhesive force. Furthermore, unnecessary driving torque of the motor to actuate the climbing robot from the over design of the magnetic adhesive force from the magnetic wheels can be avoided. Due to the level of the adaptive adhesive force is necessary to be considered, we designed the adaptive electromagnetic adhesive force mechanism for the curvature surface climbing robot with magnetic wheels. The PID controller was employed to control the electromagnetic force, and the adhesive force was measured by a load cell. This measurement signal was used as a feedback signal. In the paper, we investigated the capability of this adjustable magnetic force system. Five aspects of experimentation were implemented. It was clear that the light weight electromagnetic force adjustment mechanism could provide the flexibility to regulate the adhesive force for the magnetic robot while traveling on the ferromagnetic curvature surface.

This research proposes a system to inspect defective lenses with a polarization technique by using image processing and machine learning. Currently, a skilled operator checks the lens quality with the polarization method by eye and decides whether or not a lens is good (OK) or not good (NG). A 'not good' lens has a circle or a line appearing in the stress pattern of the lens. This research designs and develops a lens quality checking system with machine learning by simulating and prototyping a machine to experiment and collect persistent data, using the camera to capture and analyze images with image processing and machine learning techniques to decide on the lens quality in the computer. The experimental results show that the proposed system with a trained model with data augmentation and image preprocessing can achieve performance testing with 97.75% accuracy.

This paper presents the design and development of an automatic monomer filling system using machine vision in optics manufacturing. Currently, the production process is carried out by a trained worker. The monomer filling process, needs to be operated by hand and human vision is employed to control each step. The important procedure is to start-stop a valve to prevent liquid monomer overflow from glass molds. The proposed system consists of a Cartesian robot and a vision system. The vision system uses two cameras to capture four regions of interest (ROI) of four glass molds at the same time. The frame differencing technique is used to detect the difference between the last and the current video frames. The results show that each camera can be operated at real-time image processing at 52 frames per second, and the system can stop the valve for the filling monomer of each glass mold at the full level. As the result, productivity increased by 60% compared with the existing manual process making it possible to replace skilled workers.

An actuator or a support structure of a humanoid robot can be damaged in the same way that a human has musculoskeletal injuries.When an actuator is damaged or risked of being fatally damage, the robot may lack of some abilities to perform fundamental tasks such as walking.In this paper, we have applied an orthotic device which is a brace or a cast to support the damaged knee joint of a humanoid robot.By applying a brace to the knee joint, the walking parameters of the robot need to be adjusted to accommodate the motion constraint introduced by the brace.The limping gait was tested and compared with the normal walking gait.The experimental results showed that the limping gait resulted in slower walk and increased the energy consumption by 30% compared to the normal walking gait.However, a humanoid robot is able to walk despite of its damaged knee joint with the proposed orthotic device.

The Ebola virus disease (EVD) is the deadly epidemic diseases, which needs to be concerned about eradicating the spread of the disease. Studies about the policy or strategy to control the diseases have been conducted through different differential mathematical models as seen in several previous works. This paper focused on the study of applying feedback control scheme to determine the multiple control policies which are the combination of treatment and prevention actions for the EVD epidemic system. The synergetic control (SC) method can provide the chattering free characteristic for the control system. Moreover, terminal synergetic control (TSC) method can further improve the convergence rate of the control system. Thus, the TSC method was employed to define the multiple control polices for the EVD system in this study. Then, the stability of the control system was investigated. The control system was simulated to illustrate the performance of the designed terminal synergetic controller compared with the conventional synergetic controller. It was clear that using this control method, the EVD epidemic system could be regulated under the control policies without chattering, and the convergence rate of the control system was improved. Therefore, the determination the control policy for the Ebola epidemic system can be conducted alternatively and appropriately by the terminal synergetic controller design procedure.

This paper presents the design of a software prototype proposed for recording and making use of wound data and its treatment. The software has been developed as a part of a research project aiming to design an integrated hardware and software solution to support the wound treatment process. The hardware was designed to measure wound volume by using 3D-scanner. The software helps remove manual processes required for the collection, integration, analysis and presentation of wound diagnosis and treatment data. It also calculates operational parameters needed to support the diagnosis, for instance changes in wound dimension over time and volume of Lactated Ringer Solution required for patients. Our semiautomated solution offers higher accuracy of wound dimension measurement as compared to physician manual estimation and provide information needed for physicians to make a decision on wound treatment.

Traditional methods for measuring wound volume require manual work or the integration of an expensive 3D-scanner to accurately indicate the wound depth. In this work, we propose a novel affordable machine to facilitate the process of the wound measurement. We explained in detail the process of designing our prototype machine, which effectively remove general limitations of a 3D-scanner caused by unstable movement. The prototype is capable of measuring wound volume for wounds located on most parts of the body. Its accuracy might be varied by the wound shape; however, it is of satisfactory.

A novel obstacle avoidance algorithm for boat survey, "Follow the Gap Plus", is proposed by combining the "Follow the Gap" method and adapting the "Potential Field" method with an attractive field of obstacles. Then the simulation environment for testing algorithm are generated in various situations with real-world-like environment. The results show that the proposed algorithm can go to the desired destination with better performance than "Follow the Gap" method in the environment with obstacles about 4.47% of total distance, about 7.63% of total time used, and about 54.33% of average change of direction.

In this work, we proposed a method to accelerate learning by allowing a human to coach a robot behavior by inserting an intermediate target at the early phase of the reinforcement learning. By using an intermediate target, the different pair of policy and reward function was temporarily used to select an action that most likely to drive the robot toward the intermediate location, while the global reward function is still used for updating the state-action value. Q learning algorithm was used to test with the proposed method on three learning tasks: ball following, obstacle avoidance, and mountain car. The proposed technique resulted in better learning performance than the traditional RL.

SEIR model has been utilized to represent the behavior of various epidemic systems. Several kinds of diseases have been proposed and studied based on this model. The design of the vaccine law or policies to regulate the infection system is one of the important areas of studies. The vaccine policies can be determined based on the knowledge of the nonlinear control theory as seen in literature. The sliding mode control (SMC) method is one of the successful methods, which has been employed in various engineering and biological systems. Furthermore, the method has been developed as the fractional order sliding mode control (FOSMC) method, which has drawn interests of many researchers as proposed in literature. Thus, the study of the FOSMC method employing on the SEIR model was conducted to investigate the performance of the FOSMC method by simulation. Then, the controlled system by the FOSMC method was simulated and compared with that of the conventional SMC method. The simulation results showed that using the FOSMC method could manipulate all state variables to the desired reference signals with the faster convergence rate compared to the conventional SMC method. Therefore, the control of the epidemic system described by SEIR model with vaccination can be suitably obtained by using the fractional order sliding mode control method.

Vaccination is one of the approaches to control the epidemic system. The concept of the feedback control can be applied to define the epidemic vaccination policy as seen in the literature. Generally, the application can be achieved by using the sliding mode control (SMC). However, the control system can be interfered by chattering phenomena. Hence, it is beneficial to overcome the chattering in the control signal. Another desirable characteristic of the control system is the finite time convergence. To achieve these, this study examined the application of the finite time synergetic control in defining the vaccination policy for the SEIR epidemic system. Here, the vaccination policy was defined based on the FTSC method so that the recovered subpopulation could track the total population. The capability of the FTSC method was presented and evaluated through the simulation of the control epidemic system. Then, the simulation results of both FTSC and SC methods were compared. According to the simulation results, under the FTSC method, the state variable corresponding to recovered subpopulation could track the reference signal accurately as that of the SC method did. Additionally, the convergence rate of the tracking error could be improved by using the FTSC method with appropriate selection of controller parameters. Furthermore, vaccination policy was defined without chattering as a favorable characteristic of the SC method. Thus, it is evident that the finite time synergetic control method can be successfully implemented for the vaccination control in the SEIR epidemic system.

Introduction Virtual reality simulation is a promising alternative to training surgical residents outside the operating room. It is also a useful aide to anatomic study, residency training, surgical rehearsal, credentialing, and recertification. Discussion Surgical simulation is based on a virtual reality with varying degrees of immersion and realism. Simulators provide a no-risk environment for harmless and repeatable practice. Virtual reality has three main components of simulation: graphics/volume rendering, model behavior/tissue deformation, and haptic feedback. The challenge of accurately simulating the forces and tactile sensations experienced in neurosurgery limits the sophistication of a virtual simulator. The limited haptic feedback available in minimally invasive neurosurgery makes it a favorable subject for simulation. Conclusions Virtual simulators with realistic graphics and force feedback have been developed for ventriculostomy, intraventricular surgery, and transsphenoidal pituitary surgery, thus allowing preoperative study of the individual anatomy and increasing the safety of the procedure. The authors also present experiences with their own virtual simulation of endoscopic third ventriculostomy.

A virtual environment-based endoscopic third ventriculostomy simulator is being developed for training neurosurgeons as a standardized method for evaluating competency. Magnetic resonance (MR) images of a patient's brain are used to construct the geometry model, realistic behavior in the surgical area is simulated by using physical modeling and surgical instrument handling is replicate by a haptic interface. The completion of the proposed virtual training simulator will help the surgeon to practice the techniques repeatedly and effectively, serving as a powerful educational tool.