In a Ground Zero type disaster, one of the most critical components of a recovery is rescuing survivors trapped undernearth rubble in a timely and efficient manner.

What it does

WERM is a Soft muscle robots carrying sensor payloads that can slip underneath and through most debris in order to allow first responders to have a more efficient and accurate determination of where human search efforts should be focused.

How we built it

Power and control

A WERM runs off a 12V/1A power supply, making a standard car battery sufficient as a reliable voltage supply in response areas. A web UI allows for WERM control on a laptop; the operator can approximately control the direction of the WERM via tether.


A single WERM is constructed from 6 to 12 repeating actuator segments. Each segment expands to 150-200% of its original volume. By adding a frictive surface to each segment, we can replicate an earthworm's inching movement by repeatedly expanding and contracting the entire WERM.

A single segment (muscle) is composed of a silicone matrix with ethanol bubbles. Resistive coils are wound through the segment. When a current is sent through the muscle, the entire matrix expands; when the current stops, the matrix contracts. As seen in the attached video, a 1kg load can be lifted with this setup.


Sensor payloads must be light and cheap. We've experimented with an audio & RF setup: using a simple mic and RF receiver, we detect human breathing when the WERM is near a person.

Challenges we ran into

Initially, our team was going to work on a system that used physical RFID tags, but after seeing the work being done in Texas, we decided that it would be best to work on a more original idea. This lead us to utilizing most of our time on in the beginning towards validating two ideas, a predictive model for small scale WMD explosions and a microphone analysis system that could potentially find victims under rubble using RF and sound waves. In our validation process, we spoke to over 7 beneficiaries, all of whom believed that being able to support first responders first contact during SAR operations would be extremely useful at a global scale.

While we were fleshing out a prototype, we quickly learned that the biggest issue with survivor detection using RF and sound waves had to do with how difficult it is for both to penetrate through heavy debris without being extremely expensive and noisy. We came to the conclusion that our final prototype should be able to bring sensors as close to potential victims as possible in order to lower the costs of the detection technology.

Using these points, we refined our ideation process and leveraged some of the research that our group members conduct at Columbia's Creative Machines Lab. Given the potential pain points first responders face, we believe our solution provides the biggest gains for during the most critical initial moments.

Accomplishments that we're proud of

Given all of the ideas we originally started off with, collectively we are proud that we were able to pinpoint an idea that both was viable and had many dual use applications.

What we learned

As software engineers, we often feel several degrees removed from the products we are building. The accomplishment our team is most proud of was making sure our solution had viable applications via user interviews. In total we interviewed over 12 subject matter experts, ranging from FEMA directors to military operators.

What's next for WERM

Next for WERM is developing more resilient membranes so that it can navigate through more arduous terrains. We will be doing more tests on the extent of the capabilities of the WERM in addition to the most effective detection methods.

In terms of WERM application, we see it going beyond just a rubble detection system. It has applications that extend to maintenance and militaristic operations.

Built With

Share this project: