Category Archives: Ubiquitous Computing

Great paper: Unsupervised Indoor Location

I haven’t blogged in a long time, so I’m looking for any excuse. Coming across a really cool paper seems a good one.

The paper is from Romit Roy Choudhury’s group at Duke, and appears to be a collaboration with some researchers at EJUST in Egypt. The title is No Need to War-Drive: Unsupervised Indoor Localization, and was published in MobiSys (one of that groups’ 4 papers), which is taking place right now.

The main idea is very clever. Basically, dead reckoning on a mobile phone sucks, but there are lots of ‘landmarks’ that can be established from various sensors (e.g., magnetometer, sound, ambient light, etc.), that can be used to ground the estimate. These landmarks are discovered automatically using clustering. If the goal is to place a user on a map, then they can also use certain preexisting landmarks. For instance, they are able to accurately detect when a user moves up or down an elevator, escalator, or staircase, and so they can use the known locations of these structures to anchor landmarks to a map. They did some thorough testing in two buildings on campus as well as in a mall, and the accuracy is exceptional. The paper is written somewhat informally, but tells a great story of how the system was developed. There isn’t a strong novel contribution, but a fantastic combination of some clever tricks to build an indoor localization system that requires no calibration.

Kudos to the authors on a great paper. For anyone thinking about indoor localization, I recommend this work.

Are the sensors coming?

A recent blog post on the New York Times blog, titled The Sensors are Coming! describes a whole list of sensors that are expected to be included in future smartphones and mobile devices. Most smartphones contain accelerometers, which, by detecting the constant downward acceleration due to gravity, can determine the orientation of the device.  This allows the screen to orient itself when you rotate the device.

Accelerometers are also used for gaming, allowing you to steer a car by tilting the phone, for example. Recent phones such as the Nexus S now are shipping with gyroscopes, which are much more accurate at detecting changes in the orientation of the phone than accelerometers alone, allowing for more precision in the gameplay.

The NYT blog mentions other sensors, such as heart-rate, skin temperatures and perspiration sensors for determining physiological conditions, and environmental sensors such as additional microphones, thermometers, altimeters, and humidity sensors.

I would guess that the altimeters would, in fact, be barometric pressure sensors, since in my experience with the awesome Mobile Sensing Platform devices that Intel was kind enough to donate to our research group, they can fairly accurately discern between different floors of a building, for example, if properly calibrated. Calibration is no easy task, and it is more likely that these sensors will be used to detect relative changes in altitude, but this can be useful, for example, to improve the ability of phones for estimating calorie expenditure.

I believe that more environmental sensors may be included in future smartphones, mostly because they’re cheap and work well. I’m less certain about physiological sensors, for a number of reasons.

First, and most obvious in my mind, is that measurement apparatus for heart rate, galvanic skin response, and the like, need to be carefully placed on the skin in order to produce any meaningful data. I attended the AAAI Spring Symposium meeting on Computational Physiology in March 2011, and heard numerous stories from sensor vendors such as BodyMedia about how hard it is to build these devices. They work pretty well, but they are fickle and even at best the data is very noisy. Good enough to roughly approximate heart rate and caloric expenditure, but not precise enough for much else. The notion that your heart rate could be monitored by a sensor in a phone carried in your pocket (not even touching your bare skin let alone being held firmly against it), is incredibly unlikely. Moreover, I can’t imagine that there is any reliable way to measure heart rate in a way that is robust with respect to how people hold and carry their phones. More likely is that devices like the ones produced by BodyMedia will be more prevalent.

The second reason is that if these physiological sensors were capable of being used in a medical diagnostic-like setting, then they would have to get approval from organizations like the FDA. Since the idea is that they would be included in phones on which various applications can be installed, it is hard to argue that they cannot be used for this purpose, since anyone can write an app that would perform this function (maybe not well, but that’s not what is important). This is certainly not my area of expertise, but I would imagine that this would mean that any phone that includes these types of sensors would have to be approved by the FDA, or equivalent organization, and this isn’t something that phone vendors are going to put themselves through.

So in conclusion, I don’t think that the fancy physiological sensors are going to be included in smartphones, I think that they will still be relegated to special-purpose devices. These devices will (and already do) interface with your phone to facilitate interaction. It’s certainly fun to imagine a future where Angry Birds levels get easier or harder depending on the level of frustration the player is experiencing, measure by GSR.