Robotics Studio

In partnership with NASA JPL, our senior design team tackled the issue of orbital space debris.

Our goal was to design a mechanism that could safely and effectively capture a piece of space debris that is 10-30 cm in diameter.

Our design consists of an inner and outer aperture. The outer aperture is mailbox-like and is responsible for the initial grab of the space debris. The inner aperture functions like a garage door between the initial capture area and the storage container (not pictured).

Together, the inner and outer apertures create a grabbing mechanism that can grab and store many pieces of orbital space debris.

We went through several design iterations before choosing our final design based on manufacturability and material constraints.

We performed motion studies and impact analysis to guide our manufacturing decisions.

With mentorship from NASA engineers and Columbia faculty, we were able to begin machining a full-scale prototype!

The goal of this project was to create an autonomous walking bipedal robot completely from scratch. We decided to make a design inspired by Baymax from Disney's Big Hero 6.

We were responsible for every phase of design—initial sketching, 3D modeling, printing and prototyping, assembly, electronics, and programming.

Every design iteration was a cycle of modeling, simulation, printing, and testing.

Simulations were conducted in pybullet. A hill-climber algorithm were used to determine the best constants for motor actuation.

Our final robot was able to walk at nearly 1 cm/s. In the future, we want to improve this and add cosmetic changes, such as an outer shell.

Check back later for details on this project!

With a few user-inputted parameters, this Python script generates a 3D model for a lattice-structured lampshade. The model is rendered within Blender.

We first hand-sketched the shape of our lampshade and determined the best parameters and equations to define it.

Then, we wrote code to generate rings of a specified thickness and radius—arranged layer by layer—into our lampshade shape. The resultant model is then rendered and exported as an STL.

The model is then scaled, sliced, and printed!

The full code is available on our GitHub repository.

We were challenged to design a desk and chair that was topologically optimized to support the greatest load with minimal mass.

I began by creating two basic initial models in SOLIDWORKS.

Then, I imported the models into Altair Inspire and designated the design spaces for optimization.

The desk is intended to hold a load of 200 kg on its surface. The chair is designed to hold at least 150 kg, with an additional fifth of that load being placed on the back of the chair.

After optimizing for 30% of their respective initial design space volumes:

After applying a PolyNURBS fit:

Finally, a usable model! I exported the optimized fits as STEP files for rendering and STL files for printing.

Using Altair Inspire, I can optimize the geometries of any future parts to save material.

We were challenged to design an actuated four-bar linkage that could press buttons on game board as fast—and as accurately—as possible.

We began by first calculating the optimal length of each link based on a prototype coupler design. The coupler holds a solenoid that is used to press the buttons.

Once the parts were manufactured, it was time to put together the electronics hub. We used a 12V motor, encoder, H-bridge, limit switches, and Arduino to control our linkage.

Additional constraints, such as total build volume and manufacturing quality, were factored into our performance evaluation.

We achieved a final score of 192 button presses per minute! You can view the full video here.