<link>https://ehsansharafian.com/</link></image><item><title>Unplugged in Baxter State Park</h1> <article> <h1>Unplugged in Baxter State Park</h1> <p>Sat, 25 Jul 2026 00:00:00 +0000</p> <p>Some of the best reset buttons don’t have a power button at all.</p> <p>This summer, a group of us from UMaine packed up and disappeared into <strong>Baxter State Park</strong> for a weekend completely off the grid. No internet, no electricity, no notifications. Just good friends, big mountains, and the kind of quiet you only find deep in the Maine woods.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h3 id="trip-highlights">Trip highlights</h3> <ul> <li>🥾 <strong>Hiking Doubletop Mountain.</strong> Steep, rocky, and completely worth it for the view at the top.</li> <li>🛶 <strong>Canoeing</strong> on still, mirror flat water.</li> <li>🔥 <strong>Grilling saffron jooje kabob</strong> over the fire, a little taste of home in the middle of the wilderness.</li> <li>🌈 A surprise double rainbow over camp to send us off.</li> </ul> <p>The hike in followed a mossy stream through the trees, cool and green and completely silent.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>It opened up onto granite ledges and rushing water, the perfect place to stop, soak our feet, and catch our breath.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>And then, right as we started packing up to head home, the sky decided to put on one last show.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h3 id="watch-the-vlog">Watch the vlog</h3> <p>I put together a short video of the whole weekend. Take a look:</p> <div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;"> </div> <p>Highly recommend Baxter State Park for a summer adventure. Sometimes the best way to recharge is to fully unplug.</p> </article> <article> <h1>Real-Time Foot Height Estimation and Activity Classification Using a Foot-Mounted IMU Implemented on a Smartphone</h1> <p>Sat, 16 May 2026 00:00:00 +0000</p> <h2 id="why-it-matters">Why it matters</h2> <p>Tripping is a leading cause of falls, and it is closely tied to how high we lift our feet while walking. Measuring <strong>foot clearance</strong> continuously, out in daily life, could flag fall risk early — but doing it well with a wearable sensor is hard:</p> <ul> <li>A single inertial measurement unit (IMU) drifts, so integrated foot height accumulates error over time.</li> <li>The usual fix (zero-velocity updates) does not fully remove <strong>vertical</strong> drift.</li> <li>More accurate methods work only <strong>offline</strong> and do not tell you what activity the person is doing.</li> </ul> <h2 id="what-i-built">What I built</h2> <p>A <strong>real-time, single-IMU system that runs entirely on a smartphone</strong> — no cloud, no multi-sensor setup. It reconstructs the foot’s height trajectory stride by stride <em>and</em> classifies five locomotion activities (level walking, ramp ascent/descent, stair ascent/descent) at the same time.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Key contributions:</p> <ul> <li><strong>On-device Android app</strong> for stride-by-stride foot-clearance measurement and live activity classification.</li> <li><strong>Zero Height Change (ZHC) constraint</strong> — a biomechanically grounded correction that cancels cumulative vertical drift at each stride boundary.</li> <li><strong>Heel-strike velocity error</strong> used as a compact, physically meaningful feature to adaptively pick the right reconstruction model per activity.</li> <li><strong>Toe-height estimation</strong> from the IMU pose through a rigid-body transformation, with accuracy comparable to the foot itself.</li> </ul> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="validating-the-height-measurement">Validating the height measurement</h2> <p>The reconstructed foot trajectories separate cleanly by activity, and the continuous height signal tracks real-world ramps and stairs across a full walking circuit.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>To confirm the system measures real clearance, participants stepped over boxes of known heights and the estimated peak stride height was checked against each obstacle.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="results">Results</h2> <ul> <li><strong>96.08%</strong> overall activity-classification accuracy, with <strong>less than one gait cycle</strong> of latency.</li> <li>Level walking stayed at the ground reference (<strong>0.0 cm</strong>, 95% CI −1.8 to 1.8 cm).</li> <li>Cumulative height error stayed <strong>below 1.1 cm</strong> across ramp and stair negotiation (mean absolute error <strong>0.42%</strong>).</li> <li>Toe height was recovered with accuracy comparable to foot height.</li> </ul> <p>Together, this gives a practical foundation for <strong>real-time gait feedback and fall-prevention</strong> tools that could one day run on the phone already in someone’s pocket.</p> </article> <article> <h1>Human Movement Variability Conference 2026 (HMVC)</h1> <p>Fri, 01 May 2026 00:00:00 +0000</p> <p>I had the chance to attend the Human Movement Variability Conference 2026 (HMVC) in Omaha, NE, an incredible gathering of biomechanics researchers. HMVC brings together scientists and clinicians focused on understanding variability in human movement and its implications for health, rehabilitation, and performance. It was a great opportunity to present our work on <strong>wearable haptic feedback systems for gait training</strong> and to engage in discussions with leading experts about their latest advances in the field.</p> <p>Alongside the conference, we had the chance to tour the state-of-the-art facilities at the <strong>University of Nebraska at Omaha, Department of Biomechanics</strong>, which are among the best in the US. The lab features advanced equipment including a six-degree-of-freedom treadmill with an immersive virtual reality environment for movement analysis research.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> </article> <article> <h1>Swapping the Engine in My Subaru</h1> <p>Sat, 20 Sep 2025 00:00:00 +0000</p> <p>Some weekends you recharge in the mountains. Others you spend covered in oil in a parking lot, and honestly, those are just as good.</p> <p>My Subaru Outback had been fighting me for a while, so one weekend we set up shop in the parking lot and got to work.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>We started out hoping for a modest fix, replacing the valves and valve guides on the original engine to bring it back to life. I pulled the cylinder heads and we went through the valvetrain piece by piece.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>But the old engine had other plans. After a lot of work it became clear it was simply done. So we did what any stubborn mechanical engineer would do. We pulled the whole thing and swapped in a fresh engine.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Out with the old, up on the hoist and out of the car.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>I did this alongside my great friend and master wrench, and what could have been a frustrating repair turned into one of my favorite weekends of the semester. Engine hoist, hand tools, a cylinder head on the bench, and a lot of laughs. There is something deeply satisfying about understanding a machine well enough to take its heart out and give it a new one.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h3 id="a-clip-from-the-swap">A clip from the swap</h3> <div style="max-width:720px;margin:1.5rem auto;"> Your browser does not support the video tag. </video> </div> <h3 id="watch-the-full-vlog">Watch the full vlog</h3> <p>We filmed the whole weekend. Here is the story from start to finish:</p> <div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;"> </div> <p>New engine in, hood closed, and back on the road. Not a bad way to spend a weekend outside the lab.</p> </article> <article> <h1>Real-Time Activity Recognition Using Minimal Biomechanical Features: A Lightweight IMU-Based Classifier for Older Adults</h1> <p>Thu, 04 Sep 2025 00:00:00 +0000</p> <h2 id="why-it-matters">Why it matters</h2> <p>Knowing what activity a person is doing, and how well they are doing it, is the foundation of any real-world health-monitoring tool. But most activity-recognition systems fall short where it matters most:</p> <ul> <li>They lean on <strong>deep learning and large datasets</strong>, so they are too heavy to run in real time on a phone.</li> <li>They are trained almost entirely on <strong>young, healthy adults</strong>, and quietly assume everyone moves the same way.</li> <li>Older adults with mobility limits often negotiate stairs with a <strong>non-alternating leg (NAL)</strong> pattern, leading with the same leg every step, and generic models simply misread this as normal.</li> </ul> <h2 id="what-i-built">What I built</h2> <p>A <strong>real-time, smartphone-based classifier</strong> that uses only two interpretable inputs, the <strong>thigh angle and the foot angle</strong> from two small IMUs, to recognize five everyday activities: walking, stair ascent, stair descent, standing, and sitting.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Key contributions:</p> <ul> <li><strong>Minimal biomechanical features.</strong> Rather than a deep network, a simple and interpretable <strong>physics based KNN model</strong> classifies activity from just the thigh and foot angles, light enough to run live on a phone.</li> <li><strong>Adaptive toe-off segmentation.</strong> A segmentation step detects toe-off across different activities to cut each stride cleanly and reduce computation.</li> <li><strong>On-device Android app</strong> that collects labeled data and reports activity in real time.</li> <li><strong>Built for older adults.</strong> The model is explicitly trained to recognize the non-alternating leg stair pattern that generic classifiers miss.</li> </ul> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="see-it-in-action">See it in action</h2> <p>A short demo of the system running live: the phone screen shows the app classifying activities in real time while a participant walks and negotiates stairs.</p> <div style="max-width:360px;margin:1.5rem auto;"> <div style="position:relative;padding-bottom:177.78%;height:0;overflow:hidden;border-radius:12px;box-shadow:0 6px 24px rgba(0,0,0,.15);"> </div> </div> <h2 id="recognizing-how-older-adults-move">Recognizing how older adults move</h2> <p>The hard case is stairs. When someone leads with the same leg on every step, the leading and trailing legs trace very different angle patterns, so the system has to read each leg differently.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Separating the <strong>leading</strong> leg from the <strong>trailing</strong> leg in the foot and thigh angle traces is what lets the model distinguish this mobility-limited pattern from normal stair negotiation.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="results">Results</h2> <ul> <li><strong>99%</strong> activity-classification accuracy for young and older adults with normal movement patterns.</li> <li><strong>93%</strong> accuracy in distinguishing non-alternating leg stair negotiation from normal stair negotiation in older adults.</li> <li>Runs in <strong>real time on a smartphone</strong>, with no deep-learning model and no cloud.</li> <li>Validated on cohorts the model never trained on, including <strong>ten older adults</strong>.</li> </ul> <p>Together this shows that a small set of well-chosen biomechanical features can power an accessible, real-time health-monitoring system that respects how aging populations actually move.</p> </article> <article> <h1>American Society of Biomechanics 2025 (ASB)</h1> <p>Wed, 13 Aug 2025 00:00:00 +0000</p> <p>After missing the American Society of Biomechanics conference last year due to a thunderstorm and flight cancellations, it felt incredibly rewarding to finally attend in <strong>Pittsburgh, PA</strong>.</p> <p>I had the opportunity to present our work on a <strong>real-time and lightweight human activity recognition system</strong> using biomechanical features. It was an amazing experience to exchange ideas with experts in the biomechanics community and receive valuable feedback on our research.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>The ASB Annual Meeting is one of the premier conferences in the field, bringing together researchers, clinicians, and industry professionals to share the latest advances in biomechanics. This year’s meeting featured a wide range of topics from wearable sensing to clinical gait analysis.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Pittsburgh was a wonderful host city, known for its iconic bridges, vibrant riverfront, and thriving research community.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> </article> <article> <h1>The Effects of Real-Time Haptic Feedback on Gait and Cognitive Load in Older Adults</h1> <p>Wed, 11 Jun 2025 00:00:00 +0000</p> <h2 id="why-it-matters">Why it matters</h2> <p>Walking speed is one of the strongest predictors of health and independence in later life, and when it drops the cause is often <strong>neuromotor</strong>, the timing and coordination of movement, rather than weakness or poor fitness. A key culprit is reduced <strong>peak thigh extension (PTE)</strong>, how far the thigh swings back at the end of each step. If we could cue people to extend the thigh a little more, in real time and out in the world, we could lengthen their stride and speed them up, without a treadmill, a therapist, or a lab.</p> <h2 id="what-i-built">What I built</h2> <p>A <strong>wearable, smartphone-based haptic feedback system</strong>. Two IMUs track the thigh and foot angles, the phone computes PTE stride by stride, and small <strong>vibrotactile cells on the thighs</strong> deliver a gentle cue when a step needs adjusting.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>How the feedback works:</p> <ul> <li>Each stride’s <strong>PTE is compared to a personalized target</strong>.</li> <li>Two feedback strategies were tested: a cue when the step was <strong>incorrect</strong> (PTE below target, F_IR) and a cue when the step was <strong>correct</strong> (target met, F_CR).</li> <li>A third group received only <strong>verbal instructions</strong> (I_V), as a benchmark for the haptic system.</li> </ul> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="measuring-the-mental-cost">Measuring the mental cost</h2> <p>Any gait cue is only useful if people can follow it without concentrating so hard that they stop paying attention to their surroundings. To measure that <strong>cognitive load</strong>, participants did a probe reaction-time task while walking, responding to audio beeps, with reaction time measured precisely from the sent and received audio.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="results">Results</h2> <ul> <li><strong>+14% stride length</strong> and <strong>+18% gait speed</strong> from a single session with the haptic system.</li> <li>Gains were <strong>comparable to verbal coaching</strong>, but reached through a different mechanism (direct thigh feedback rather than instruction).</li> <li>Added cognitive load was <strong>small</strong>: reaction time rose only <strong>27 ms (F_IR)</strong> to <strong>74 ms (F_CR)</strong>.</li> <li>Thirty community-dwelling older adults, averaging about <strong>80 years old</strong>.</li> </ul> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Together this shows that real-time thigh haptics can meaningfully improve older-adult gait in a single session while keeping attentional demand low, a promising, low-cost path toward wearable gait training and fall prevention.</p> </article> <article> <h1>RehabWeek 2025 — Chicago</h1> <p>Mon, 12 May 2025 00:00:00 +0000</p> <p>What an incredible week in <strong>Chicago</strong> — stunning architecture along the river and an inspiring program at <strong>RehabWeek 2025</strong>, one of the largest international gatherings in rehabilitation technology.</p> <p>I was honored to present two posters from our group at the University of Maine <strong>Biorobotics and Biomechanics Lab</strong>:</p> <ul> <li><strong>Real-Time Human Activity Recognition Using Minimal Biomechanical Features and IMU-Based Classification</strong> — </li> <li><strong>Individualized Gait Training for Older Adults Using a Wearable Haptic Feedback System</strong> — </li> </ul> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>A highlight of the trip was visiting the <strong>Shirley Ryan AbilityLab</strong> — their state-of-the-art facilities, teams, and research made a strong impression.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>I also had the chance to try a self-balancing exoskeleton from <strong> </strong> — a great hands-on look at where wearable robotics for mobility and gait rehabilitation is heading. Sincere thanks to the RehabWeek 2025 organizers, session chairs, and the <strong>American Congress of Rehabilitation Medicine (ACRM)</strong> for such a valuable and inspiring event advancing rehabilitation science.</p> </video> </article> <article> <h1>Increasing Thigh Extension with Haptic Feedback Affects Leg Coordination in Young and Older Adult Walkers</h1> <p>Sat, 01 Mar 2025 00:00:00 +0000</p> </article> <article> <h1>IEEE EMBC 2024, Orlando</h1> <p>Mon, 15 Jul 2024 00:00:00 +0000</p> <p>I had the chance to present our research titled <strong>“Enhancing Gait in Older Adults with a Haptic Feedback System”</strong> at the 46th Annual International Conference of the (EMBC 2024) in the beautiful city of Orlando, Florida.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>EMBC is one of the largest international conferences in biomedical engineering, bringing together researchers, clinicians, and industry professionals from around the world. It was a great experience to exchange ideas with experts in the field and receive valuable feedback on our work in gait rehabilitation for older adults and individuals with walking impairments.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p>Orlando also offered some memorable moments outside the conference, including visits to Walt Disney World and Universal Studios. <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> </article> <article> <h1>Experience</h1> <p>Tue, 24 Oct 2023 00:00:00 +0000</p> </article> <article> <h1>Surgical Robotics — Control Systems & Inverse Kinematics</h1> <p>Sat, 30 Sep 2023 00:00:00 +0000</p> <p>During my time at , I worked as a Mechanical and Control Systems Engineer on an advanced robotic platform for minimally invasive laparoscopic surgery. My work focused on real-time control, inverse kinematics, and the surgeon-robot interface.</p> <h2 id="the-system">The System</h2> <p>The platform is a multi-degree-of-freedom surgical robot with dual manipulator arms, a surgeon console, and a training simulator for laparoscopic procedures. It was designed for telesurgery, enabling a surgeon to operate remotely over a network connection. During our testing, we successfully demonstrated this capability across two separate hospitals: the surgeon console was located in one hospital while the robot performed the procedure in another, carrying out a real surgical operation in real time.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="control--kinematics">Control & Kinematics</h2> <p>I developed real-time control systems for the robotic arms using TwinCAT 3 (Beckhoff) and implemented inverse kinematics algorithms enabling precise, coordinated motion for surgical tasks. I also built Python-based graphical interfaces for intuitive surgeon-robot interaction.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="surgeon-training">Surgeon Training</h2> <p>Beyond engineering, I trained surgeons and medical students on operating the robot, safety protocols, and clinical best practices using the simulation platform.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <h2 id="technical-stack">Technical Stack</h2> <p><strong>TwinCAT 3 · Python · MATLAB · Real-time Control Systems · Inverse Kinematics · Laparoscopic Robotics</strong></p> </article> <article> <h1>Trajectory Optimization of a Spot-Welding Robot in the Joint and Cartesian Spaces</h1> <p>Sun, 01 Jan 2023 00:00:00 +0000</p> </article> <article> <h1>Revisiting Screw Theory-Based Approaches in the Constraint Wrench Analysis of Robotic Systems</h1> <p>Fri, 01 Jan 2021 00:00:00 +0000</p> </article> <article> <h1>Trajectory Optimization for Spot-Welding Robots</h1> <p>Sat, 01 Feb 2020 00:00:00 +0000</p> <p>For my M.Sc. thesis, I collaborated with an automobile manufacturing company to optimize the motion of the robots that perform spot welding on the vehicle-body production line.</p> <h2 id="the-problem">The Problem</h2> <p>On a car-body assembly line, many serial robots perform hundreds of spot welds. Their trajectories were programmed manually by operators — reliable, but far from optimal in speed and energy use.</p> <h2 id="what-i-did">What I Did</h2> <p>I formulated the robot’s motion as an optimization problem and solved it two ways:</p> <ul> <li><strong>Minimum-time trajectory</strong> — a single-objective problem minimizing the welding gun’s travel time.</li> <li><strong>Time–energy trajectory</strong> — a multi-objective problem minimizing both travel time and energy consumption.</li> </ul> <p>The planning considered the <strong>position and orientation of the welding gun simultaneously</strong>, in both joint and Cartesian spaces, while respecting the robot’s joint limits and the required gun orientation at each weld point. I used <strong>genetic algorithm</strong> and <strong>particle swarm optimization</strong> to search for the best trajectories.</p> <h2 id="results">Results</h2> <p>Applied to a real robot on the production line, the optimized trajectories reduced the spot-welding operation’s travel time by:</p> <ul> <li><strong>up to 24%</strong> for the minimum-time trajectory</li> <li><strong>up to 11%</strong> for the multi-objective (time + energy) trajectory</li> </ul> <p>This work was published in the <em>International Journal of Robotics and Automation</em> (2023) — see the .</p> <p>compared with the manually programmed trajectories in use.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > </div> </div></figure> </p> </article> <article> <h1>Ehsan Sharafian M</h1> <p>Mon, 01 Jan 0001 00:00:00 +0000</p> <h2 id="about-me">About Me</h2> <p>I am a third-year PhD candidate in Mechanical Engineering at the University of Maine’s Biorobotics and Biomechanics Lab, supervised by Dr. Babak Hejrati. My research focuses on developing wearable inertial measurement unit (IMU)-based sensing systems for real-time gait analysis and fall prevention in older adults.</p> <h2 id="research-focus">Research Focus</h2> <p>My work integrates multiple sensing modalities to develop practical solutions for improving mobility and quality of life. I specialize in:</p> <ul> <li>Real-time gait analysis using distributed IMU sensor networks</li> <li>Foot clearance estimation for fall prevention</li> <li>Adaptive haptic feedback systems for gait correction</li> <li>Android mobile applications for sensor data acquisition</li> <li>Machine learning for terrain and obstacle classification</li> </ul> <h2 id="background">Background</h2> <p>I hold a strong foundation in mechanical engineering and biomedical research. My work spans hardware integration, software development, and biomechanical analysis, bridging the gap between laboratory research and practical applications.</p> </article> </main></body></html>