Manuele Capece
Reinforcement Learning Engineer @ Digimatic Robotics srl
Italy
Picerno
Computer Software
Reinforcement Learning, IsaacLab, Neural Network, Simulation, Data Analysis, Data Visualization, Supervised Learning, NVIDIA Isaac Sim, Isaac Lab, ROS 2, URDF, RViz, Reti neurali, PyTorch, Gazebo, Linux, Biometria, Visual Studio, Monitoraggio ambientale, OpenGL
Experience
Reinforcement Learning Engineer
Potenza, Basilicata, Italia
Physical Modeling of a 16 DoF quadruped robot -Conversion of CAD models into URDF format and configuration of the digital twin in USD format, including the modeling of closed kinematic chains. -Development of a high-fidelity simulation environment in NVIDIA Isaac Sim/Lab, involving parameter tuning and optimization to ensure simulation accuracy and validation of sim-to-real consistency. Training of Locomotion Policy -Design and implementation of a reinforcement learning training architecture in Isaac Lab, leveraging the DirectManager environment built on the RSL-RL framework with the Proximal Policy Optimization (PPO) algorithm. -Observation modeling for the actor–critic network, using a latent representation of observation history to remove privileged information (e.g., robot linear velocity) from the observation space. -Reward engineering with PyTorch, involving the design of custom reward functions to achieve natural, stable, and energy-efficient locomotion behaviors. Sim-to-Real Policy Evaluation -Validation of the trained policy through both sim-to-sim and sim-to-real transfer experiments. -Deployment of the learned policy onto the real quadruped robot hardware, using ROS 2 control tasks and the LibTorch C++ library, implementing command velocity control via joystick interface.
Studente tirocinante
Potenza, Basilicata, Italia
Tirocinio formativo con finalità di tesi dal titolo "Modellazione del digital twin di un robot quadrupede in ambiente di simulazione NVIDIA Isaac Sim". Attività svolte: realizzazione del modello digitale URDF, creazione del mondo virtuale in Isaac Sim, verifica della corrispondenza tra realtà e simulazione, miglioramento del modello digitale, implementazione del modello finale in Isaac Lab Obiettivi raggiunti: Il digital twin realizzato è in grado di riprodurre il complesso comportamento del robot reale. Il processo di addestramento di un modello di attuatore basato su rete neurale MLP è riuscito a migliorare ulteriormente la corrispondenza ottenendo un errore inferiore del 37% rispetto al modello ideale fornito dal simulatore.
Manuele Capece's Contact Information
Phone
Find the Right Leads
Find Verified Contact Data
What LeadContact does well
Find verified emails, phone numbers, and decision-makers with 98% accuracy.
Find Leads
Find the right people by company, role, industry, location, and more.
925M+ professional profiles

Find Emails
Access verified email addresses for your target contacts.
657M+ emails

Find Phone Numbers
Get cross-validated phone data from multiple top sources.
239M+ phone numbers

More Accurate. Lower Cost.
Find contact data in 1 tool with 98% accuracy
LeadContact integrates leading enrichment tools to deliver more accurate contact data—without paying for each one.
Great conversations start with the right contact.
It’s time to find yours.


