Research
Four areas, one body of work.
The areas overlap by design. Wearable sensing feeds the rehabilitation work, force modelling feeds the surgical arm, and the constraint that shapes all of it is the same: these systems have to work where compute, budget and specialist time are scarce.
01
Surgical robotics
Haptics, force sensing and needle steering for robot-assisted surgery.
Force sensing and haptic feedback for robot-assisted surgery, aiming to give surgeons a reliable sense of touch through a robotic interface, a capability still largely inaccessible in most African surgical settings. The work also covers tool to tissue interaction modelling, including beam on foundation modelling of percutaneous needle steering for biopsy and brachytherapy. Current experimental work uses the Hapkit platform from Stanford's haptics lab, feeding, alongside computer vision, into a dexterous surgical arm capable of gripping and suturing.
Papers
- Beam on Foundation Modelling and Closed Loop Control of Tool to Tissue Interaction for Percutaneous Needle Steering
IEEE NIGERCON 2026 · 2026
Accepted - Pain, Consent, and the Scalpel: Racialized Surgical Ethics from Sims's Operating Room to Contemporary African Caesarean and Laparoscopic Practice
10th Annual Lagos Studies Association Conference · 2026
Presented
02
Rehabilitation robotics
Motor function recovery, focused on partial stroke rehabilitation.
Technology to support motor function recovery, with particular focus on partial stroke rehabilitation, a need that is widespread and significantly underserved across Nigerian and African healthcare settings. The work extends the lab's wearable sensing into active rehabilitation support through a dedicated rehabilitation robotic arm, moving from passive monitoring toward assistance.
Projects
Papers
03
Wearable robotics
Inertial sensing with embedded TinyML for real-time health monitoring.
Wearable, sensor-based systems for real-time health monitoring, including inertial sensing paired with embedded machine learning models small enough to run on the device itself. This anchors the tremor detection and severity classification work and feeds directly into the rehabilitation robotics pipeline, where the same signal processing provides the control input.
Projects
04
Medical AI and diagnostics
Machine learning for screening where compute and labelled data are scarce.
Machine learning for medical screening and diagnosis in settings where compute, labelled data and specialist time are all limited. Current work covers multi label classification of thoracic pathologies from chest radiographs, with an emphasis on correcting severe class imbalance and on architectures small enough to run on modest hardware. Diagnostics sits alongside rehabilitation, mobility and surgical assistance in the lab's stated long term scope.
Papers
- Class Imbalance Corrected Multi Label Classification of Thoracic Pathologies in Resource Constrained Chest Radiograph Screening
IEEE NIGERCON 2026 · 2026
Accepted