Embedded Files
We have updated the research project offers for current NYU students. Please e-mail the listed contacts directly.
Please tell us your motivation to the specific project, your skillsets and past experiences + something that shows that (CV, website, photos, videos, github, etc), and your transcript.
Students from all research backgrounds are welcome – for example, mechanical engineering, bioengineering, computer science, materials etc.
Updated July 2026
Slime Mold Environmental Co-Evolution
Description — Slime molds are eukaryotic single-celled organisms that have no brain or central nervous system. Despite their simplicity, they have been shown to be capable of surprisingly complex tasks, such as solving mazes and anticipating periodic changes in their environment. Due to these unique characteristics, they are an interesting organism to study. In this project, we will investigate how a slime mold can affect its environment, and how the environment can, in turn, affect the slime mold. This may include adjusting the light source, temperature, moisture level, food distribution, or other environmental factors based on the mold's behavior, including the direction and speed of growth, or electrical signals measured from the organism. This is an exploratory project that will focus on studying the co-evolution of the mold and the environment through an embodied intelligence perspective.
Requirements — Experience working with live organisms, such as plants, fungi, or cells. Basic knowledge of signal processing, sensor integration, and electrical circuits. Computer vision experience preferred.
Workload — 20% cultivating organism and literature review, 30% experiment and analysis, 20% signal processing (electrical and/or vision), 30% system integration.
Contact — Suzanne Oliver / sho8511@nyu.edu
Object Detection and Haptic Feedback for Handheld Local Navigation Device for Visually Impaired Users
Description — White canes are widely used as mobility aids for people with visual impairments. However, they can be uncomfortable for long-term use and carry a risk of repetitive use injuries. In this project, we propose to create a sensorized cane handle that detects obstacles with vision and depth sensors, but is lighter and more portable than a traditional cane. Haptic feedback will be provided to the user via soft pneumatic chambers that inflate when an object is detected to mimic the sensation of hitting an obstacle with a cane. This project will involve improving an existing prototype of the device, integrating sensors, processing camera and time-of-flight data in real time, and conducting user testing to assess the comfort and usability of the end product.
Requirements — Strong signal processing and electronics background, experience with mechanical design and programming. Experience with computer vision, soft robotics, and/or assistive technology is preferred.
Workload — 50% sensor integration and signal processing, 30% mechanical prototyping, 20% user testing
Contact — Suzanne Oliver / sho8511@nyu.edu
Coordinated Controllable Stiffness Feather Star Inspired Robot for Multimodal Swimming
Description — This project seeks to understand how multiple feathers can be coordinated to achieve multiple modes of swimming in 3D space. Modes of swimming may include, but are not limited to, paddling, lift based/carangiform, gliding, and crawling. This project will involve building off of a prototype feathered robot with a tendon-driven stiffening mechanism. By parameterizing the morphology of the feathers and robot controller and embedding the robot with proprioceptive sensors, we aim to discover robust morphologies and effective control schemes to achieve underwater locomotion that adapts to a variety of environments.
Requirements — Experience with mechanical design, programming, and electronics. Experience with ML is preferred but not required.
Workload — 40% sensor integration, 30% mechanical design, 30% testing and optimization
Contact — Magan Lee / ml10362@nyu.edu
Soft Robotics for Performing Arts
Description — This project explores soft robotics as a creative and expressive medium within live performance. Working at the intersection of engineering and performing arts, the student will design and prototype a soft robotic wearable or body-worn element intended to interact with a live performer — the specific performance form (dance, theater, music, or another discipline) is open. The project is exploratory: rather than optimizing for a fixed technical benchmark, the student will iterate on actuation, form, and timing based on hands-on testing, investigating how a soft robotic system can function as an expressive partner rather than a functional device. The ideal candidate is comfortable working without a fully defined endpoint, and enjoys shaping a research direction through experimentation as much as executing a specification.
Requirements — Fabrication skills for soft robotics (casting, textile/garment integration), Electronics skills (microcontroller, motor/pump control), Programming skills (Python or Arduino/C++), Genuine interest or background in a performing art, CAD/modelling experience (Fusion 360, SolidWorks) a plus, not required.
Workload — 30% prototype design and fabrication, 30% control system development, 40% iterative testing and creating performance choreographing
Contact — Nana Obayashi / no2358@nyu.edu
Does Your Body Change How You See the World?
Embodiment-Conditioned Navigation with Vision-Language Models
Description — Large vision-language models (VLMs) can suggest navigation paths from images, but they are trained on general data with no knowledge of a robot's body. This project fine-tunes a VLM on image-text descriptions written from the physical perspective of 2-3 different robot morphologies, then compares its navigation plans against the base model on the same scenes. The goal is to test whether embodiment-specific language shifts planning toward physically feasible choices for each body and what that reveals about how much physical reasoning is already latent in large-scale pretraining.
Requirements — Programming (Python), Familiarity with reinforcement learning and a deep learning framework (PyTorch), Experience with a physics simulator (MuJoCo, Isaac or ROS) is a bonus, but not required. Interest in working with real robot hardware, not only in simulation.
Workload — 30% data collection and model fine-tuning, 40% simulation and policy training, 30% deployment and testing on hardware
Contact — Wendi Liang / wl2693@nyu.edu
Task-Driven Design Optimization of Tendon-Driven Continuum Robots
Description — An octopus arm grasps well not only because of how it moves, but also because of how its body is built. Tendon-driven continuum robots raise the same question: how long should each segment be, how stiff should the backbone be, and where should the tendons go? Today these choices are mostly made by hand. This project automates the search. The student will write a generator that turns a list of design parameters into a complete physics simulation, so hundreds of candidate designs can be created and tested in minutes. Each design competes on grasping tasks, wrapping its whole body around objects of different sizes and weights, while an optimization algorithm evolves better bodies generation by generation. The winning designs are then built with a flexible backbone and 3D printed disks and tested on the real robot, to see whether the champion in simulation is also the champion in the real world.
Requirements — Programming (Python), Experience with a physics simulator (MuJoCo, PyBullet, or similar) is a bonus, but not required, Basic CAD or 3D printing experience is a bonus, but not required, Interest in working with real robot hardware, not only in simulation.
Workload — 30% simulation setup, 40% design optimization and evaluation, 30% hardware building and testing
Contact — Wendi Liang / wl2693@nyu.edu
Wi-Fi Backscattering for Wireless and Batteryless Shape Estimation of a Soft Silicone Tissue
Description — Wi-Fi backscattering is a method of transmitting data wirelessly and free of battery. Instead of signal generation, these devices harvest Wi-Fi waves and communicate with reflection and slight alternation of those waves. In this project we aim to develop a signal reflective device embedded into a silicone mold and an ML model capable of recognizing shape based on reflected signal and study its potential applications in soft robotics.
Requirements — Circuit design, signal processing, and electromagnetic (EM) physics, strong programming skills (Python, C++). Machine learning (ML) experience is a bonus, but not required.
Workload — 60% simulation development, 20% ML model development, 20% testing and optimization
Contact — Daniil Filimonov / df2789@nyu.edu
Closed-Loop Control System for Undulatory Amphibious Robots with Fluid Center of Mass [filled]
Description — Continuing our previous work on the use of mass transfer in peristaltic robots for locomotion optimization in unmodelled terrains, this project seeks to integrate a robust closed-loop control system into the project and study the usefulness of mass shifting in autonomous navigation in unmodelled terrains. By embedding a variety of sensors (IMU etc.) in existing robotic platforms and establishing a control system that recognises terrains and adapts to them, the research will create a framework for autonomous navigation of amphibious robots with a fluid center of mass.
Requirements — Control systems experience, sensor fusion, embedded systems knowledge, strong programming skills (Python, C++). Machine learning (ML) experience is a bonus, but not required.
Workload — 20% embedded system development, 60% programming and control, 20% testing and optimization
Contact — Daniil Filimonov / df2789@nyu.edu
Project archive
Project archive
Coupled Jetting and Tentacle Swimming in a Squid-Inspired Robot [filled]
Description — Cephalopods are known for their unique use of tentacles and jetting to achieve a wide variety of maneuvers. Many robotic platforms have explored jetting and tentacle swimming for locomotion individually, but the interplay between the two modes remains underexplored. In this project, we propose integrating the two locomotion methods to create a robot that combines jetting with tentacle swimming to explore the fluid dynamics of a soft tentacle in a controllably turbulent environment. This project will involve prototyping a robot capable of jetting and tentacle swimming, using PIV to investigate the coupled fluid dynamics, and using simulation/optimization techniques to design a controller for the robot.
Workload — 45% design and fabrication, 45% experimental testing, 10% control/optimization
Requirements — strong hardware and electronics skills, programming experience (MATLAB, python, experience working with microcontrollers preferred), interest in fluid dynamics and soft robots.
Contact — Magan Lee / ml10362@nyu.edu
Sensor-Integrated Soft Handle for Ergonomic Navigation [filled]
Description — White canes are widely used as mobility aids for people with visual impairments. However, they can uncomfortable for long-term use and provide limited information to the user. In this project, we propose to create a soft, sensorized cane handle, that prioritizes ergonomics and provides the user with haptic feedback about their surroundings (for example, detecting wetness of the ground or obstacles outside the canes reach). Signals will be analyzed in real time to provide useful haptic feedback (through vibration or squeezes) to the user. User testing will be conducted to assess the comfort and usability of the system compared to a standard cane.
Workload — 40% sensor integration, 40% soft wearable device design and fabrication, 20% haptic control
Requirements — strong hardware and electronics skills, CAD, programming basics (MATLAB, python), interest in wearable robots and assistive devices. Biomechanics knowledge a plus but not required.
Contact — Suzanne Oliver / sho8511@nyu.edu
Cane Without a Cane: Obstacle Detection with Soft, Sensorized Cane Handle [filled]
Description — White canes are widely used as mobility aids for people with visual impairments. However, they can be uncomfortable for long-term use and carry a risk of repetitive use injuries. In this project, we propose to create a sensorized cane handle that detects obstacles with vision and depth sensors, but is lighter and more portable than a traditional cane. Haptic feedback will be provided to the user via soft pneumatic chambers that inflate when an object is detected to mimic the sensation of hitting an obstacle with a cane. This project will involve prototyping the cane handle, actuating the soft pneumatic chambers, integrating sensors, processing camera and time-of-flight data in real time, and doing user testing to assess the comfort and usability of the end product.
Workload — 40% sensor integration and signal processing, 50% design and fabrication, 10% user testing
Requirements — strong hardware and electronics skills, CAD knowledge, programming experience (MATLAB, python, experience working with microcontrollers preferred), interest in wearable robots and assistive devices. Biomechanics knowledge a plus but not required.
Contact — Suzanne Oliver / sho8511@nyu.edu
Page updated
Google Sites
Report abuse