Muslim A. Alaran

PhD Student · Robot Learning · Montanuniversität Leoben, Austria

I am a PhD student and doctoral researcher at the Chair of Cyber-Physical Systems at Montanuniversität Leoben, advised by Prof. Elmar Rückert. I work on skill learning for humanoid robots (Unitree G1, R1), targeting industrial production and logistics tasks.

More broadly, my research lies at the intersection of robot learning, reinforcement learning, and embodied AI, with a focus on dexterous manipulation and generalist robot policies. Before this, I was a junior researcher at the AI for Robotics Lab (FER, University of Zagreb), and I completed my MS in Robotics at Nazarbayev University.

Portrait of Muslim Alaran

Research Interests

Robot Learning Humanoid Robots Reinforcement Learning Imitation Learning Embodied AI Dexterous Manipulation Foundation Models for Robotics Visual-Tactile Perception Motion Planning

News

Aug 2026 🎉 Started my PhD at the Chair of Cyber-Physical Systems, Montanuniversität Leoben, with Prof. Elmar Rückert, working on humanoid skill learning.
May 2026 Paper on time and object aware transformers for gesture recognition accepted in Robotics and Autonomous Systems.
May 2026 Paper on spatially aware value fusion for marine manipulation accepted to the 23rd IFAC World Congress (Busan).
Apr 2025 Systematic review of VEP-based brain-computer interfaces in gaming published in IEEE Access.
Mar 2025 Joined the AI for Robotics Lab, FER, University of Zagreb, as a junior researcher working on robot manipulation.

Publications

Hand grasping a ball under four video augmentations

A Time and Object Aware Transformer-Based Method for Structured Hierarchical Gesture Recognition

A. Papanikolaou, V. Sivtsov, M. Alaran, D. Shkolnik, E. Zereik, I. Markovic, I. Petrovic, F. Bonsignorio

Robotics and Autonomous Systems, Art. no. 105518, 2026

A hierarchical hand–object gesture recogniser (GCPR) that predicts the primary grasp gesture together with its power, contact type and release behaviour.

JournalDOI
Kinova Gen3 with a Robotiq gripper and its wrist-camera views of the sphere

Spatially Aware Value Fusion in Decomposed Reward Architectures for Marine Manipulation

V. Sivtsov, M. Alaran, D. Shkolnik, A. Papanikolaou, I. Markovic, E. Zereik, I. Petrovic, F. Bonsignorio

23rd IFAC World Congress, Busan, Republic of Korea, 2026

Adaptively re-weights the value heads of a decomposed-reward RL agent, so it can manipulate under wave-induced motion; validated in simulation and with hardware in the loop on a Kinova Gen3, outperforming existing methods across wave magnitudes.

Conference
Schematic of a VEP-based brain-computer interface for video games

Neurotechnology in Gaming: A Systematic Review of Visual Evoked Potential-Based Brain-Computer Interfaces

A. Keutayeva, C. J. Nwachukwu, M. Alaran, Z. Otarbay, B. Abibullaev

IEEE Access, vol. 13, pp. 74944–74966, 2025

A PRISMA-guided systematic review that screened 16,347 studies down to 46, covering visual-evoked-potential BCIs in gaming: signal acquisition, stimulation paradigms, applications and open challenges.

JournalDOI
Illustration of AI in healthcare

Challenges and Opportunities of Artificial Intelligence in African Health Space

M. A. Alaran, S. K. Lawal, M. H. Jiya, S. A. Egya, M. M. Ahmed, A. Abdulsalam, U. A. Haruna, M. K. Musa, D. E. Lucero-Prisno III

Digital Health, vol. 11, Art. no. 20552076241305915, 2025

A commentary on where AI can improve diagnosis, disease surveillance and resource allocation in African healthcare, and the privacy, equity, infrastructure and policy barriers it faces.

JournalDOI
Top view of a ship with its surge, sway, roll and yaw motions

APO-Optimized ARX-Based System Identification of a Ship Maneuvering System

F. Olanrewaju, M. Shafiullah, M. Al-Dhaifallah, M. Alaran

IEEE 22nd International Multi-Conference on Systems, Signals & Devices (SSD), Monastir, Tunisia, 2025

Tunes an ARX model of full 4-DoF ship dynamics with the Artificial Protozoa Optimizer, reaching a mean R² of 0.9437 and improving on existing identification methods.

ConferenceDOI
Perspective images transformed into fisheye views

Field of View Invariant Object Recognition Using Non-Linear Transformation Augmentation

M. Alaran, Z. Balgabekova, H. A. Varol

IECON 2024, 50th Annual Conference of the IEEE Industrial Electronics Society, Chicago, IL, USA

Builds RMFV365, a 5.1-million-image dataset of perspective images warped into many lens distortions, so that YOLOv7 recognises objects across camera types, from perspective to fisheye.

Planned path on the lab map, with the robot's 3D model

Cloud-Integrated Navigation System for Scalable Autonomous Ground Robots

U. Sharipov, S. Kasenov, M. Alaran, A. Askhatova, Y. Orynbay, P. Jamwal

30th International Conference on Mechatronics and Machine Vision in Practice (M2VIP), Leeds, United Kingdom, 2024

A ROS-based delivery robot that combines A* global planning and DWA obstacle avoidance with AWS cloud services for scalable indoor navigation.

ConferenceDOI
Annotated hemispherical camera image

A Data-Centric Approach for Object Recognition in Hemispherical Camera Images

Z. Balgabekova, M. Alaran, H. A. Varol

IECON 2023, 49th Annual Conference of the IEEE Industrial Electronics Society, Singapore

Introduces COHI, the first benchmark of real fisheye images (1,000 images, 39 classes); training YOLOv7 on MS COCO combined with its fisheye-transformed version improves accuracy by 2.5%.

ConferenceDOICode & data

Open Source

Kinova Gen3 grasping a bottle

finetune-octo-kinova-hybrid

Fine-tuning and deploying Octo, a generalist robot policy, on a Kinova Gen3 for goal-image-conditioned bottle pick-and-lift from ~110 demonstrations, with a hybrid action head.

Released pipeline (v1.0.0): demonstration collection, hand-eye calibration, TFDS/RLDS conversion, training, offline evaluation, and live deployment around a latency-aware receding-horizon controller.

UR5 arm tracking a Cartesian path

open-planning/roboplan · contributor

Pinocchio-based robot motion planning library.

Six merged pull requests adding dual end-effector support, a Cartesian trajectory type, acceleration and jerk limits, Jacobian base-frame support, and keyboard teleoperation.

Images transformed into fisheye views

wide-fov-dataset-toolkit

Fisheye dataset generation for object detection.

Converts perspective object-detection datasets (COCO, Objects365) into fisheye and wide-FoV training data via eight lens-distortion transforms with matching bounding-box remapping; generalizes the IECON 2024 pipeline into a pip-installable, tested package.

Research Experience

Doctoral Researcher — Chair of Cyber-Physical Systems

Montanuniversität Leoben · Leoben, Austria

  • Researching skill learning for humanoid robots (Unitree G1, R1) targeting industrial production and logistics tasks.
Aug 2026 – present

Junior Researcher — Artificial Intelligence for Robotics Lab

FER, University of Zagreb · Zagreb, Croatia

  • Developed and deployed imitation and reinforcement learning policies for manipulation on Kinova Gen3 arms.
  • Owned sim-to-real transfer for a decomposed-reward RL architecture with spatially aware value fusion for marine manipulation.
Mar 2025 – Jul 2026

Research Assistant — Robot Control and Learning Lab

Nazarbayev University · Astana, Kazakhstan

  • Built a safety-critical reinforcement learning framework for human–robot collaboration, enforcing Speed and Separation Monitoring constraints in PPO-based motion planning on a 7-DoF Kinova Gen3.
  • Prototyped obstacle-avoidance groundwork in MuJoCo, training the arm to reach goals among static and moving obstacles using reward shaping and curriculum learning.
Aug 2024 – Feb 2025

Research Assistant — Institute of Smart Systems and Artificial Intelligence

Nazarbayev University · Astana, Kazakhstan

  • Built a 5.1 million-image fisheye dataset using non-linear geometric mapping and trained YOLO models on it for object recognition.
  • Generated synthetic training data for object recognition using NVIDIA Omniverse.
  • Collected hemispherical-camera images into a 1,000-image fisheye benchmark for evaluation.
Jan 2023 – Jul 2024

Undergraduate Research Assistant

Dept. of Electrical & Electronics Engineering, University of Ilorin · Nigeria

  • Reviewed literature and contributed to the design of a multifactor student attendance system using fingerprint biometrics and RFID.
Jan 2020 – Mar 2021

Education

PhD in Robot Learning

Montanuniversität Leoben · Leoben, Austria

Advisor: Prof. Elmar Rückert · Chair of Cyber-Physical Systems

Aug 2026 – expected 2030

MS in Robotics

Nazarbayev University · Astana, Kazakhstan

GPA 3.7/4.0 · Dean's Honor List (Spring & Fall 2023) · Abay Kunanbayev Scholar
Thesis: Enhanced Object Recognition in Hemispherical Images using Data Augmentation and Synthetic Image Integration
Advisor: Prof. Huseyin Atakan Varol

Aug 2022 – Jun 2024

BEng in Electrical & Electronics Engineering

University of Ilorin · Ilorin, Nigeria

First Class Honors · Top 1% of graduating cohort
Thesis: Attendance Monitoring System using RFID and Fingerprint
Advisor: Dr. I.O.A. Omeiza

Oct 2015 – Aug 2021

Service & Activities

Awards & Honors