top of page

How does Air Mouse Work in a Remote Control

With the popularity of smart TVs on the market , Currently, there is a smart remote control on the market, and also including the Smart Air Mouse remote control.


Smart Air Fly remote control which takes advantage of the air flying squirrel movement and gesture recognition technology.


Many of Dusun's production like DSR-0479DSR-0480DSR-0427 smart air mouse remote control use this technology, the hardware is equipped with a 3D orientation axis gyroscope and a gravity sensor , this remote control with six axes ( gyroscope + gravity sensor ), the hardware basis, through 2.4GHz radio frequency technology to communicate with the device, so as to achieve the purpose of remote control devices. 


These products has been able to completely replace the mouse , no desk operations fullest remote enthusiasm.

Use of these technology products, responsive cursor can do , there will be no delays , and some products also joined the movement compensation function , thereby preventing shake , move the cursor more stable.


In some products, once a firmware upgrade , you can also do gesture recognition and correction function can even recognize the wrist up and down motion, flipped around, push the button to play, pause, volume , and switch the screen size , etc.

2.4GHz RF communication technology eliminates cursor delay and distance limitations , over up to 10m , and will not be disturbed.

Now let's learn the basic principles of air mouse technologies below.

The air mouse remote control is a device that uses the data output from the gyroscope to control the cursor on the screen. 


The principle seems relatively simple, but in the using process, there are still many problems that need to be solved. 


This article is a summary of the Dusun's previous development. 


Basic Principle Map the air mouse's Pitch angular velocity and the Z axis (Yaw) to the mouse's movement speed.


The X-axis (Pitch) angular velocity and the Z-axis (Yaw) of the air mouse are mapped to the moving speed of the mouse. 


Therefore, the way to grasp the device is determined in advance, such as which side faces forward and which side faces up.


Sensitivity also needs to be determined experimentally. 


If you can get the resolution of the screen or the screen size for dynamic adjustment, the effect is better.

Anti-static offset

When the mouse is not moving, the cursor naturally cannot move. 


But have to point out that cheap gyroscopes themselves have static output. Some can even reach 20 degrees / s. 


Fortunately, the offset usually does not change due to the environment. 


Therefore, it is necessary to make corrections before leaving the factory. 


This kind of problem is easier to solve.

Smoothing algorithm

Consumer-grade gyroscopes are limited by cost, the accuracy is certainly not very high and the data will be jittery. 


By increasing the sampling rate, the so-called oversampling technique, multiple data are filtered and smoothed at the same time to feedback the current state of motion as accurately as possible.

Motion plane does not match

As shown in the figure above, if you rotate in the Y direction with Z as the axis, the mouse will draw a horizontal track.


But when the device itself has a certain tilt, such as a rotation angle on the X axis, then like the previous movement, a diagonal line will be drawn. This can be confusing for the user.


If you have an accelerometer, measure the tilt of the device, and through the sensor fusion algorithm, you can relatively avoid such problems. 


But the algorithm requirements are higher.

Rotation, not displacement

Ordinary mice are based on displacement, how much is moved, and the cursor will move the same distance.


But the air mouse is not, it is based on rotation. 


If you move the air mouse on a non-rotating basis and move it a certain distance, you will find that the cursor has not changed at all (may cause shaking due to some interference). This will make the user confused. 


There may be algorithms that attempt to take horizontal or vertical motion by taking displacements by so-called acceleration integration. 


But this kind of algorithm is definitely not good, because the accuracy limit of the sensor will be very inaccurate, and there is basically impossible to solve it.

Anti-key jitter

The buttons of a normal mouse are basically made at the top of the mouse, so the buttons do not cause mouse displacement. 


If the button is imagined to be on the side of the mouse and the button is very hard, jitter may occur. 


Therefore, the air mouse has high requirements on the buttons, needs to be as soft as possible, and has clear tactile feedback. 


When pressed, you can try to keep the mouse from shifting. The most troublesome thing is the double-click operation.


When you double-click, the power of a person clicking a mouse is usually much larger than clicking. And there will be two consecutive jitters. 


If the displacement is too far, the operating system will consider it a drag operation rather than a double click, which requires additional processing.

Gesture Recognition

This involves a more advanced problem.


The air mouse is obviously off the keyboard, some shortcut operations cannot be realized, and certain actions can be detected by certain gestures, thereby triggering the previously defined operations. 


But this involves the topic of machine learning and requires high user learning costs. The same operation, the features made by different users may be completely different.


In battery-powered systems, power consumption must be considered.


If motion is not detected within a certain period of time, it will automatically enter a low-power mode and send data intermittently.

In conclusion

It is not difficult to achieve the simplest air mouse through the basic principle, but the user experience is not good. By adding other methods, the accuracy can be improved. 


However, it must be pointed out that although the method of solving the motion plane mismatch can solve the matching problem, it will reduce the real-time performance and produce a feeling of similar time lag, so the trade-off still needs specific analysis.


Different sensor and hardware design, need to adjust the parameters, through repeated experimental comparisons to determine a reasonable solution.


bottom of page