arXiv subjects

Duong Le

Publications and source records attributed to Duong Le.

3 recordsLinked to original sources

arXiv2026

A Replay-Constrained Simulation Framework for Personalization of Powered Knee--Ankle Prosthesis Controllers

Personalization of impedance controllers for powered prosthetic legs is critical to accommodating individual gait biomechanics but remains challenging. Existing methods rely on time-intensive human-in-the-loop exploration and/or constrain optimization to low-dimensional, single-joint parameter subspaces. Sim-to-real transfer has enabled high-dimensional locomotion control for legged robots, but in assistive device control the human partner remains un-modelable. We present a replay-constrained simulation framework: a MuJoCo-based simulator reproduces prosthetic knee-ankle dynamics while replaying recorded hip kinematics and feedback-based ground reaction forces from individual walking data, bypassing the need to model complex human neuromuscular control mechanisms. We demonstrate the framework with a deep reinforcement learning policy that personalizes phase-dependent stiffness, damping, and equilibrium angle at both joints simultaneously, maximizing a biomimicry-based reward computed solely from onboard prosthesis measurements. Experiments with three participants with transfemoral amputation during level-ground walking at 0.8~m/s demonstrate strong simulation-to-hardware predictive validity (Pearson $r=0.96$--$0.997$). The best-performing policy on hardware was consistently predicted within the top five simulation policies for all participants. The learned controllers improved overall biomimicry rewards by 42--59\% relative to the unpersonalized baseline. The framework supports scalable high-dimensional personalization of powered prosthetic legs and is amenable to extension to higher-dimensional controller parameterizations such as neural-network controllers.

cs.RO↗

arXiv2025

Magnetically Guided Endothelial BioBots: A Next-Generation Strategy for Treating Complex Cerebral Aneurysms

Cerebral aneurysms affect three to five percent of the population, and rupture remains a major cause of stroke-related death and disability. Current therapies, surgical clipping, endovascular coiling, and flow diversion, have improved outcomes but each carries limitations. Clipping is invasive and often unsuitable for deep or posterior lesions. Coiling is prone to recurrence from compaction or incomplete occlusion, particularly in wide-neck or fusiform aneurysms. Flow diverters offer improved durability but rely on rigid metallic scaffolds that may malappose in tortuous vessels, compromise branch arteries, delay endothelialization, and necessitate long-term dual antiplatelet therapy. These shortcomings highlight a gap in current management: devices primarily provide mechanical occlusion but fail to conform to complex geometries or reliably promote rapid, complete endothelialization. As a result, aneurysm necks may remain exposed to persistent flow, delayed healing, and thrombosis. To address this, we propose magnetically guided endothelial BioBots as a next-generation therapeutic strategy. BioBots are biodegradable hydrogel carriers embedded with magnetic nanoparticles and coated with primed endothelial progenitor cells. Delivered through microcatheters and guided by external electromagnetic fields, they can assemble across aneurysm defects. Once localized, they form a conformal, geometry-adaptive endothelial patch that provides immediate antithrombotic protection and, as the hydrogel degrades, leaves behind a stable, functional endothelial lining. By integrating microrobotic navigation with regenerative vascular biology, BioBots may overcome the central limitations of current devices and enable safer, more durable treatment for complex aneurysms.

q-bio.TO↗

arXiv2022

Deep Convolutional Neural Network and Transfer Learning for Locomotion Intent Prediction

Powered prosthetic legs must anticipate the user's intent when switching between different locomotion modes (e.g., level walking, stair ascent/descent, ramp ascent/descent). Numerous data-driven classification techniques have demonstrated promising results for predicting user intent, but the performance of these intent prediction models on novel subjects remains undesirable. In other domains (e.g., image classification), transfer learning has improved classification accuracy by using previously learned features from a large dataset (i.e., pre-trained models) and then transferring this learned model to a new task where a smaller dataset is available. In this paper, we develop a deep convolutional neural network with intra-subject (subject-dependent) and inter-subject (subject-independent) validations based on a human locomotion dataset. We then apply transfer learning for the subject-independent model using a small portion (10%) of the data from the left-out subject. We compare the performance of these three models. Our results indicate that the transfer learning (TL) model outperforms the subject-independent (IND) model and is comparable to the subject-dependent (DEP) model (DEP Error: 0.74 $\pm$ 0.002%, IND Error: 11.59 $\pm$ 0.076%, TL Error: 3.57 $\pm$ 0.02% with 10% data). Moreover, as expected, transfer learning accuracy increases with the availability of more data from the left-out subject. We also evaluate the performance of the intent prediction system in various sensor configurations that may be available in a prosthetic leg application. Our results suggest that a thigh IMU on the the prosthesis is sufficient to predict locomotion intent in practice.

cs.RO↗