Robotics, embedded,
and machine learning.

Designing and controlling a tendon-driven continuum manipulator inspired by biological muscular hydrostats. The control stack combines a constant-curvature kinematic model, nonlinear MPC for real-time trajectory optimisation, and a Soft Actor-Critic residual policy that compensates unmodelled dynamics.

Contributed to the INDICATE-FH project focusing on novel diagnostic methods for food intolerances. My role involved developing computer-vision algorithms using OpenCV and TensorFlow for medical image analysis, implementing deep-learning models for automated detection of diagnostic markers, and collaborating with medical researchers on data collection, preprocessing, and validation.

The project brings together researchers from multiple disciplines to create innovative diagnostic and treatment approaches for food intolerance conditions.

View project details →

Developed autoencoder-based models for signal-to-noise ratio (SNR) analysis and denoising. The work involved benchmarking different bottleneck designs, loss functions, and noise models, then optimizing model architectures for measurable improvements in denoising accuracy.

This research contributes to broader applications of machine learning in signal processing and data analysis.

View on GitHub →

Developed indoor navigation systems for autonomous mobile robots using Bluetooth beacon and WiFi localization technologies. The system combined multi-sensor fusion for accurate positioning, implemented path planning algorithms with dynamic obstacle avoidance, and enabled real-time autonomous navigation in indoor environments.

This project contributed to autonomous robot navigation research and practical deployment of localization technologies in constrained indoor spaces.

Developed an autonomous lawn mower system using advanced image processing and waypoint-based outdoor navigation. The system implements computer-vision algorithms to detect and classify grass height (long vs. short grass), enabling dynamic blade adjustment for optimal cutting performance. Navigation operates without perimeter wire constraints, using GPS waypoint-based planning for autonomous outdoor operation.

This project demonstrates integration of perception, control, and autonomous navigation in a practical robotic application.

• Master Internship: Design of a Mechanical Control Interface for a Continuum Robotic Arm
• Focus: advanced control, bio-inspired robotics, parametric modeling, autonomous systems
• Coursework: Model Predictive Control, Deep Learning, Sequence Learning, Intelligent Agents, Computer Vision

• Thesis: Improvement of Human Face Detection and Recognition in Low Illumination and Occluded Conditions
• Focus: image processing, pattern recognition, embedded systems, signal processing

• Developed autoencoder-based models for signal-to-noise ratio (SNR) analysis and denoising
• Analyzed performance metrics and optimized model architectures for denoising accuracy
• Contributed to research on ML applications in signal processing and data analysis

• Contributed to the INDICATE-FH project on novel diagnostic methods for food intolerances
• Developed computer-vision algorithms using OpenCV and TensorFlow for medical image analysis
• Implemented deep-learning models for automated detection of diagnostic markers
• Collaborated with medical researchers on data collection, preprocessing, and validation

• Developed navigation algorithms for autonomous mobile robots in indoor environments
• Investigated sensor fusion and localization methods for indoor robot navigation
• Researched robotic lawn-mower systems — path planning and obstacle avoidance
• Supervisors: Prof. Nobuaki Nakazawa, Prof. Yoichi Shiraishi

• Demining robot development for humanitarian mine-clearance applications
• Embedded robotics and autonomous systems for real-world deployment
• Mechanical design, control systems, and field-ready reliability engineering

• Developing embedded software in C and Rust for laser-based intra-operative histology devices
• Implementing real-time control systems and hardware-software integration for medical equipment
• Software documentation and compliance with IEC 62304 (Medical Device Software Lifecycle Processes)

• Designed and implemented embedded systems for IoT devices and access-control systems
• Developed the core MQTT communication hub for distributed IoT device networks
• Improved system stability, reliability, and real-time performance of access loggers

Organized by Misr International University (MIU), Cairo, Egypt.

• Programming · C, Python, MATLAB, Rust
• Development & tools · Linux, Git, Docker, VS Code
• Robotics & control · Gazebo, Model Predictive Control, Simulink
• Simulation · COMSOL Multiphysics, MATLAB/Simulink, GNU Octave
• ML & computer vision · TensorFlow, PyTorch, Keras, OpenCV, scikit-learn
• Embedded hardware · Arduino, Raspberry Pi, ESP32, PIC microcontrollers
• CAD & mechanical design · SolidWorks, Rhino, Grasshopper
• PCB design · EasyEDA
• Circuit simulation · OrCAD, Proteus, LTSpice
• Data analysis & visualization · Matplotlib, Seaborn, Pandas

A. M. Arfi*, D. Bal*, M. A. Hasan, N. Islam, and Y. Arafat.

doi.org/10.1109/TENSYMP50017.2020.9230857

D. Bal, M. I. Tusher, M. Rahman, and S. R. Saymon.

doi.org/10.1109/STI50764.2020.9350480

D. Bal, M. I. Tusher, M. Rahman, and S. R. Saymon.

doi.org/10.1109/ICICT4SD50815.2021.9396924

D. Bal, A. M. Arfi, and S. Dey.

doi.org/10.1109/IEMTRONICS52119.2021.9422496