The personal desk fan just got a whole lot smarter!  The SmartFan is a low-power desk fan which learns when and where to focus its attention; using facial recognition to direct cooled air toward regions of the body which most affect comfort.

The SmartFan is part of a holistic comfort control strategy which encourages the customization of your indoor environment in an intelligent way.  It’s a concept centered around the idea that you (and not the facilities manager or your boss) know what makes you the happiest.  When you’re happy, you’re more productive; which makes your boss happy.  You could say its a win - win.  Plus, its good for the planet.  So maybe its a win - win - win. 

It has been suggested that increasing the ambient room temperature and cooling the body from increased air velocity (such as from a desk fan) has the potential to lead to substantial energy savings.  One recent study reported savings in the range of 17-48%.  Yet, most desk fans only have the ability to perform one of two actions - they can sit still or oscillate back and forth.  Because most people don’t sit still nor do they rock back and forth all day, desk fans often spend an inordinate amount of time cooling the area around the actual person.  Let’s just say they often miss the point (I know, bad pun).   

Results show occupant comfort can be improved (while using less energy) if they only focus on cooling the face and neck regions of the body.  The SmartFan uses two small DC fans (drawing approximately ⅓ of the power of a typical desk fan) and a built-in video camera to track the position of the user’s face and adjusts the fans accordingly.  It strongly believes in the motto, ‘cooler heads will prevail’!

The SmartFan also joins the Internet of Things revolution by communicating with everyday devices and the larger central building control system.  And if that wasn’t enough, it also has the ability to learn over time.  That is to say, it uses state of the art machine learning algorithms to learn when and how to maximize your comfort.  Welcome to the future of personalized comfort control.  


The facial recognition algorithm measures the distance (in pixels) from the centroid of the region of interest to the center of the image.  These values are then converted into angular positions for each of the motors on the Smart Fan.

Raise your hand if you’ve ever been completely and utterly uncomfortable at work (and we’re not just talking about the awkward comment your boss made this morning)?  Several recent studies underscore the fact that thermal comfort is one of the primary variables affecting occupant satisfaction.  Ok, you can put your hand down now.

As an architect, one of my primary objectives is creating an indoor environment that fulfills the dual goals of user satisfaction and energy performance.  Yet, most buildings often fail to meet these two needs. Why, you ask?  A quote by Stewart Brand, author of the book How Buildings Learn: What Happens After They’re Built helps explain our situation:

"Almost no buildings adapt well. They’re designed not to adapt; also budgeted and financed not to, constructed not to, administered not to, maintained not to, regulated and taxed not to, even remodeled not to.  But all buildings ... adapt anyway, however poorly, because of the usage in and around them are changing constantly."  

Conceptual sketches of the Smart Fan.

Four working prototypes have been built (Black, Blue, Yellow, and Red).

An exploded axon drawing of the Smart Fan.

The concept progressed from initial sketches to detailed CAD models over the course of several months during the summer of 2012.  As the parts of the enclosure were being 3D printed, I developed and built the PCB (printed circuit board) which would facilitate the wireless communication between the host computer and the desk fan. Once assembled, I wrote some simple test programs to evaluate the Open CV facial recognition algorithms and other various comfort parameters.

At present, there are four working prototypes (black, blue, yellow, and red) all of which ‘talk’ with an online interface as well as other local hardware devices (personal heaters, task lights, air conditioners, etc).  I am in the final stages of the software development where I am implementing the advanced machine learning algorithms; giving the SmartFan the ability to learn your preferences and design an optimized schedule for comfort and energy performance.  

The online interface used to control the Smart Fan and various other devices.