Power Puff — Power Puff is a project done in collaboration with students from MIT Sloan, MIT School of Engineering and SharkNinja. This product is a robotic vacuum machine that cleans all types of floor surfaces, without requiring user efforts.


PowerPuff is a robotic vacuum machine that cleans all types of floor surfaces, without requiring your efforts. With an ultrasonic sensor, the vacuum can detect the floor type that it is approaching. After detecting a surface type, PowerPuff will then "switch" to the most appropriate cleaning mode. It does this by lifting up and down the cleaning tool that is compatible to each surface. For example, if PowerPuff detects hardwood floor, it will change to the wet mop by dropping down the wet mop pads.

UX Lead/Product Designer

Will Sripakdeevong
Adriene McCance
Priyanka Chaturvedi
Lee Evangelakos

Armandeep Sidhu
Anna Fountain
Esther Kim


How can cleaning be an effortless activity, rather than a burdensome task?

Along with the Design Problem, we had a few set of goals:
Create a product using "smart" technology.
Make trial-and-error iterations to find the best practice. 
Incorporate and balance out each discipline throughout project development.

Team Direction

As soon as the team of 8 designers, entrepreneurs, and engineers first met, it was extremely important to get everyone on the same page and have the team dynamics on the right track. With so many potential directions, it was important to have everyone on-board to be participating in every initial process of this project. Designers initially took a responsibility to facilitate ideation workshops and this process really reflected the outcome of this product and was overall good experience for everyone to get to know each others and learn skills from different backgrounds.


Supporting our Decision 

Following SharkNinja’s advice we see our product as having broad appeal, primarily because of its ease of use and time-saving qualities.  We estimate the potential market for our product as current home cleaners of all intensity levels as well as the ‘wish I was more of a cleaner’ segment.  This effectively means our product would be desirable to almost all households as we see it a supplement to relieve some (but not all!) of current cleaning chores.  This product is not a deep cleaner so can always be supplemented for the more serious cleaner.  

In America alone there are 125M households who might theoretically desire such a cleaning device at the right effectiveness level and price-point.  More realistically though, our short-term addressable market would be Roomba purchasers since our product would offer a superior experience along the same, supplemental, non-intrusive cleaning goals that the Roomba offers.  The Roomba currently holds 18% market share in the $2.3 billion high end US vacuum market ($414 million), so there is certainly market potential for this product by surpassing the Roomba.  

In order to capture additional share beyond the Roomba’s customer base we would have two options.  On the high end side, we could create a product that is cleaning much more effectively than the Roomba does to the point where this might serve as a standalone (non-supplemental) cleaning solution and attract more customers by offering a much more desirable product.  

Alternatively, if we could reduce a Roomba equivalent at a lower price point we might attract new customers who desired a Roomba but previously did not find its functionality to be worth the price point.  We will be more informed about whether to pursue one of these strategies further on in our product development cycle.  


To understand the users' feelings towards household cleaning and their pain points and tendencies, each of us observed number of people completing different tasks. 


Rapid Prototyping 

For our first testing, we patched up existing cleaning tools to a Roomba. Observing its movements and the effectiveness of each tool, we improvised new ideas along the way; for example, how would the mop module be attached? Perhaps inserting a roll of wet pad underneath the bot. What if the cleaning bot had a camera on it? So on..


Second Testing and Issues

While we believed that the multi-surface roll was a good idea, we ran into some usability problems. Because of the fact that we had wet pad, dry pad, and a brush sharing the space on the roll, each tool had a very small surface area. This would make cleaning ineffective. Also, it was technologically difficult to have the roll come to a complete stop (from continuously rolling) and turn exactly so that one cleaning tool would align parallel to the ground. 


Final Engineering Concept

Comming up with an engineering concept was a challenge as we faced limitations. We were able to prove failure with rapid prototyping and moved on to our final decision. Through testings we were able to prove that we could make cheaper roomba with better effective cleaning service 



With the timeframe we had, we faced limitations to build or hack already excisting roomba. As a solution we engineered the bot to simply raise one cleaning tool up and down to proove the efficiency of the cleaning. This was a great solution for our team, because the up-and-down mechanism is much more doable than having a cylinder roll up to a specific point. 

User Persona

With the research data that we collected, to best illustrate our targeted users, we came up with a persona named Roxy. 

Floor Plan

Based on the poluation research, this is how typical American houses would look like for younger professionals. There would be two to three types of floor types ranging from wood, carpet, to tile. 

Business Model

Product Rendering

Look-Like Model

Power Puff Mobile App

IOT Smartphone integration makes powerpuff possible to be controlled whenever users are on the go. For example, users can remotely run powerpuff while at work to be able to invite friends home after work. Also, they can set up times and modes for powerpuff to run individually without having to worry about floor conditions. 

Quick Start

On the main screen, powerpuff can start its job simply by one touch of a button. There are 4 pre-sets which they can choose; Asap, Standard, Full, and Manual. 

Modes & Schedule

Users can manually set up modes in detail depending on cleaning status. Also, they can set up schedules and frequency.