Project STEP (GY) (Launching Fall '24)
-
Designing and building humanoid robots from ground up to automate daily tasks.
Humanoid robots are uniquely equipped to walk over, climb, and navigate obstacles relative to other types of robots. Despite the intense research interest in these systems, the cost and maintenance of useful humanoid designs remains high, limiting their adoption as platforms for basic research and experimentation. Our goal is to leverage recent advances in additive manufacturing, actuators, and computing for the development of open-source, widely-available biped robots.
The overarching goal is to develop an accessible (open-source, affordable, and easily manufactured) biped robot as a robust platform for perception, gait, manipulation, and locomanipulation research. This includes: 1. Reduction of cost and manufacturing complexity with 3-D printed designs 2. Implementation of contact-aware motion planning algorithms for control 3. Release of control algorithms and other software libraries that are well-maintained and easily accessible. 4. Release of system models for integration with common simulation environments for testing.
Methods & Technologies
- 3D CAD software
- Designing and building prototypes
- Simulation and analysis methodology (FEA, MATLAB etc.)
- Actuators and wire harness design
- Additive and subtractive manufacturing (3D printing, CNC milling, laser cutting, etc.)
- CUDA and GPU programming
- 3D reconstruction, depth estimation, object detection, tracking & state estimation, instance and semantic segmentation
- Physical simulators (e.g. Mujoco, IsaacSim, Drake, Pybullet)
- Multi-DOF control algorithms (trajectory-optimization, model-predictive control, whole-body control)
- Spatial kinematics, dynamics and related mathematical concepts
Areas of Interest
- Mechanical Engineering
- Electrical and Computer Engineering
- Computer Science
- Business and Marketing
- Any major with technical expertise, a passion for teamwork, learning and building robots
Subteams
- Mechanical Design
- Robotics Software and Controls
- Electronics
- Operations
Partners
- NYU Tandon School of Engineering
- Applied Dynamics and Optimization Laboratory
Faculty Advisor
- William Peng
- Email: william.peng@nyu.edu