The Applications and Dangers of Accelerometers By Michael Murphy Supervisor: Adrian Perrig and Cy. Lab at CMU Background Information Why is this important? Accelerometers in the World Today Our Project Continued Accelerometers: What are they? Accelerometers are tiny sensors which allow devices to sense acceleration. They’re made of tiny piezoelectric sensors, three of which are arranged in right angles to one another. Each carries a tiny charge when compressed by a force like motion or gravity, giving it a sense of force in all three dimensions. The most commonly used accelerometer within our phones is the LIS 311 DLH, at 3 x 3 x 1 millimeters, it is a tiny, low power, high performance linear accelerometer. It senses the forces of acceleration in the X Y and Z planes to a precision of six decimal places. With the advent of smart phones in the world today, accessibility for accelerometers has never been higher. This allows for a more variety of uses, such as in our case specifically, the interplay between cyber and physical security with these smart phones. In our case, we want to find out how android phones could be used as threats to someone's physical security. Accelerometers have been used for a variety of uses throughout the world today, from medical to research, from car performance to robotics, as well as being used in the detection of earthquakes. Or used in studies with baseball pitchers to prevent arm injuries. However, this is when accelerometers were generally used only by the scientific community. Nowadays the applications are much more limitless. The typing detection program currently has limited success in very specific situations, however, promising results have been shown. Currently with only the accelerometer, it is not quite accurate enough to detect what the user is typing, only when they do press the screen. However, with the addition of the orientation sensor, it seems possible to detect where on the screen the user pressed. While we were unable to go into testing this because of time purposes, future results seem promising. An accelerometer, and a size comparison. For example, one could use the accelerometer in someone's phone to identify what they’re currently doing. Sitting, walking, running, riding an elevator, going up or down stairs, etc. Or even find out where they are while driving. By using the accelerometer and knowing their start point, one could theoretically find out the persons path, speed at every point during the drive, and end destination. Conclusion So as you can see, being able to track such information as what a user is doing, or even typing , is quite possible with such a sensor as an accelerometer. Such a device that is commonplace within our society shows us that this field of research has quite the potential for security threats and requires much further investigation. Activity Recognition Activity recognition is the attempt to recognize actions and motions of a user from a observations from their actions using various sensors. This has many different potential applications and the possible connection to many different fields of study has made it a lucrative endeavor. To understand activity recognition better, consider a scenario like this: a grandmother wakes up her small house where she currently lives alone. She turns on the stove and starts cooking her breakfast, After going through her morning routine, tracked via a device, a computer-generated voice gently reminds her to turn off everything. The activities she's done is stored to a server so her loved ones can make sure that she is doing alright, even An example of visual activity miles from her. recognition, in this case, tracking specific movements within martial arts. TEMPLATE DESIGN © 2008 www. Poster. Presentations. com Our Project We had many ideas for what we wanted to do, and ended up selecting two very interesting projects. The first half of our project is a Activity recognition program, a program which can identify what the user is doing throughout their day. And the second half is a typing interpretation program, which consists of using the accelerometer to identify what the person is typing on the Android itself. The activity recognition program ended up being able to sense general actions no matter the orientation of the phone using the accelerometer. We were able to detect within 95% accuracy if someone was walking, running, jumping, going up or down stairs, and could even. You can the program teach sense much more specific actions for each user, and have it than just the phone turning with recognize those afterwards. accelerometers. Soon this guy will know much more about you than you think. Closing and Contact Info Michael Murphy Olin College of Engineering This research made possible thanks to: TRUST-REU Carnegie Mellon University Cy. Lab Kirstin Gates Adrian Perrig
Скачать презентацию The Applications and Dangers of Accelerometers By Michael
8dfe537b60b105a41acf6fc241a5992a.ppt