The other day on the MSDN forums, someone asked about how to detect a shaking motion on Windows Phone 7. I've been playing with the accelerometer lately so I took great joy in answering this along with providing a working implementation. The question was asking about shaking motion in a left-right direction. I made a class that detects left-right and up-down motion (totally ignoring the Z-axis all together for now). Though extending it to consider the Z-axis wouldn't be hard.
The code for detecting the motion has been abstracted in a class called ShakeDetector. The algorithm used has a few variables/constants defined that can be modified to tune the behaviour of the class. The classes constructor accepts an [optional] parameter of how many times the phone should be shaken before the motion is considered acceptable. <codeminimumaccelerationmagnitude> can be raised or lowered to control how hard the device needs to be shaken to be considered acceptable. And MinimumShakeTime takes a time span that defines the maximum length of time over which a shake sequence must occur to be considered acceptable. Once the user moves the phone in a way that meets the requirements for the type of shake we wanted to detect a ShakeDetected event is raised.
I've reduced the direction in which the device is moving to one of 8 directions (North, East, South, West, and the directions in between those). I could have kept the direction as an angle and just ensured that there was atleast a minimum difference between the angles but I thought using the directions on a map would make it easier for someone else to understand.
void</span /> _accelerometer_ReadingChanged(object</span /> sender, AccelerometerReadingEventArgs e)
{
//</span />Does the current acceleration vector meet the minimum magnitude that we
</span /> //</span />care about?
</span /> if</span /> ((e.X*e.X + e.Y*e.Y) > MinimumAccelerationMagnitudeSquared)
{
//</span />I prefer to work in radians. For the sake of those reading this code
</span /> //</span />I will work in degrees. In the following direction will contain the direction
</span /> //</span /> in which the device was accelerating in degrees.
</span /> double</span /> degrees = 180</span />.0*Math.Atan2(e.Y, e.X)/Math.PI;
Direction direction = DegreesToDirection(degrees);
//</span />If the shake detected is in the same direction as the last one then ignore it
</span /> if</span /> ((direction & _shakeRecordList[_shakeRecordIndex].ShakeDirection)
!= Direction.None)
return</span />;
//</span />This is a shake we care about. save in in our list
</span /> ShakeRecord record = new</span /> ShakeRecord();
record.EventTime = DateTime.Now;
record.ShakeDirection = direction;
_shakeRecordIndex = (_shakeRecordIndex + 1</span />)%_minimumShakes;
_shakeRecordList[_shakeRecordIndex] = record;
CheckForShakes();
}
}
void</span /> CheckForShakes()
{
int</span /> startIndex = (_shakeRecordIndex - 1</span />);
if</span /> (startIndex < 0</span />) startIndex = _minimumShakes - 1</span />;
int</span /> endIndex = _shakeRecordIndex;
if</span /> ((_shakeRecordList[endIndex].EventTime.Subtract
(_shakeRecordList[startIndex].EventTime)) <= MinimumShakeTime)
{
OnShakeEvent();
}
}
The example code can be found in my SkyDrive account here. If you want to see the program in action, there is a video on YouTube.
CodeProject
I attended Southern Polytechnic State University and earned a Bachelors of Science in Computer Science and later returned to earn a Masters of Science in Software Engineering.
For the past few years I've been providing solutions to clients using Microsoft technologies for web and Windows applications.
While most of my CodeProject.com articles are centered around Windows Phone it is only one of the areas in which I work and one of my interests. I also have interest in mobile development on Android and iPhone. Professionally I work with several Microsoft technologies including SQL Server technologies, Silverlight/WPF, ASP.Net and others. My recreational development interest are centered around Artificial Inteligence especially in the area of machine vision.