Behnam Gholami
-
Email
bg24306@essex.ac.uk -
Location
Colchester Campus
Profile
- Markerless Motion Capture and Human Pose Estimation
- Computational Biomechanics
- Modeling and Optimization in Movement and Structural Systems
- Smart Structures, Vibration and Aeroelastic Control
Biography
Behnam Gholami is a PhD researcher at the University of Essex, where he is developing markerless motion capture systems using single RGB cameras for human movement analysis. His interdisciplinary research lies at the intersection of biomechanics, computer vision, and machine learning, with a focus on making human motion assessment accessible, scalable, and robust in real-world environments. His recent work integrates pose estimation algorithms with biomechanical modelling to estimate joint kinematics, particularly under conditions of occlusion or limited camera views. Before pursuing his doctoral studies, Behnam earned an MSc in Biomedical Engineering from Amirkabir University of Technology, where he focused on smart materials and optimal control strategies for vibration suppression in adaptive structures. His earlier BSc work also centered on active flutter control in aerospace structures using piezoelectric materials. He has authored multiple peer-reviewed publications in high-impact journals such as Smart Materials and Structures, European Journal of Mechanics A: Solids, and the International Journal of Structural Stability and Dynamics. Behnam has also presented his work internationally, contributed to patented technologies, and served as a research engineer and biomechanics analyst across several institutions. His research is driven by a passion for bridging mechanical systems design with human motion science to address real-world challenges in health, sport, and engineering.
Qualifications
-
MSc (Hon) Biomedical Engineering Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran (2017)
-
BSc (Hon) Mechanical Engineering Iran University of Science and Technology (IUST), Tehran, Iran (2013)
Research and professional activities
Thesis
Monocular Markerless Motion Capture Using a Single RGB Camera for Human Movement Analysis
Markerless motion capture using a monocular RGB camera offers a low-cost, accessible alternative to traditional multi-camera systems for human movement analysis. This study uses computer vision algorithms to estimate joint kinematics from single-view videos, even under occlusion. Validated against gold-standard systems, the method shows strong accuracy for lower-limb gait analysis, supporting applications in clinical assessment and sports outside lab settings.
Supervisor: Dr Bernard Liew , Dr Xiaojun Zhai
Research interests
Robust Single-Camera Markerless Knee-Flexion Estimation During Overground Gait Under Self-Occlusion
Various tools analyse human locomotion, with optical systems widely used for their accuracy. Marker-based motion capture (MBMC) is considered the gold standard but requires complex multi-camera setups. To address this, researchers have turned to single-camera, markerless methods using Human Pose Estimation (HPE), which estimates joint positions from 2D images. Despite advances, the use of HPE for lower-limb analysis in healthcare with single cameras remains limited.
Parameter Identification of the Lower Limb Spring-Damper Model using two-dimensional Kinematic Data
Several studies have used mass-spring-damper (MSD) models to analyze lower-limb biomechanics, but most are limited to single-degree-of-freedom systems. This study adopts a Standard Linear Solid (SLS) model, inspired by Hill’s Muscle Model, to estimate stiffness and damping of lower-extremity joints using 2D kinematics from a front-kick task. A three-link system was modeled, and parameters were estimated via forward dynamics. Further research is needed to optimize the model.
Optimal locations of magnetorheological fluid pockets embedded in an elastically supported honeycomb sandwich beams for supersonic flutter suppression
This study uses a Single Objective Genetic Algorithm (SOGA) to optimize the placement of magnetorheological fluid (MRF) pockets in a honeycomb sandwich beam to suppress supersonic flutter. The structure is modeled using beam theory, Winkler–Pasternak foundation, and quasi-steady piston theory, with MRF as a Kelvin–Voigt material. Simulations show that optimized MRF pocket placement improves damping and delays flutter onset, offering an effective mass–performance trade-off.
Optimal locations of piezoelectric patches for supersonic flutter control of honeycomb sandwich panels, using the NSGA-II method
This study investigates active flutter control of supersonic sandwich panels with honeycomb interlayers using optimally placed piezoelectric patches. NSGA-II is employed to determine optimal actuator/sensor locations. The structure is modeled using piston theory, Hamilton’s principle, and Galerkin method. Flutter bounds and aeroelastic response are analyzed, and proportional feedback is applied to suppress vibration. Control effects of two strategies are compared.
Active Flutter Control of a Supersonic Honeycomb Sandwich Beam Resting on Elastic Foundation with Piezoelectric Sensor/Actuator Pair
This study investigates active flutter control of supersonic sandwich panels with honeycomb cores on elastic foundations using piezoelectric materials. The system is modeled via beam theory, piston theory, and the Rayleigh–Ritz method, with flutter boundaries obtained using the p-method. Aeroelastic responses are computed using Runge–Kutta integration. Proportional feedback and LQR control are applied to suppress vibrations and raise the critical aerodynamic pressure.