1.1 Motivation and Objective
In the last decades, many commercial mouse devices were developed with different technologies. The first generation of a commercial mouse device is so called mechanical mouse, which uses a single ball that can rotate in any direction. This ball is connected against to two rollers. One roller detects the forward-backward motion and other the left-right motion. The movement of these two rollers is detected by an encoder and an electrical signal is send to the computer. In the computer, a driver software in the operation system converts the signal into motion of the mouse cursor along X and Y axes on the screen [15]. The disadvantage of this kind of mouse is that it often requires maintenance, due to the moving parts that can easily accumulate dust and lint. Besides that, it does not perform well in slippery surfaces, requiring in most of the cases a mouse pad for better performance.
The second generation of mouse devices is so called optical mouse, which uses an optoelectronic sensor that takes successive pictures of the surface on which the mouse operates. The surface is illuminated by an LED or a laser diode. Changes between one frame and next are processed by the image processing part of the chip and translated into movement on the two axes using a block matching algorithm.
Comparing this generation of mouse device with the previous one, it presents higher sensitivity and practically do not require any maintenance. However, most of the optical mouse do not work well in glossy and transparent surfaces and demand a higher average of power.
The next generation of mouse devices that is starting to appear in the market is
called inertial mouse devices, which can use gyroscope or accelerometer sensors to detect movement for each axis supported. These two kinds of sensors consume quite less power than the optical sensors, also has a huge potential to cost less than optical sensors after mass production. Besides that, a new way to interact with computer systems will be allowed, since they do not require a surface to operate. Moreover, when using a wireless battery-powered mouse device, it will increase easy-of-use and due to the small consume of energy, it can be used during long period of time without recharging. Another benefit of using inertial sensors is their size. They are extremely small ICs that can easily be embedded in unusual objects, like a ring, a watch or glasses. Such objects that may be used as mouse devices, especially for handicap people.
In the market, it is already possible to find hybrid devices that use optical sensor and inertial sensor. The first one is only used in a 2D surface and the second one is used on fly. However, adding an optical sensor would increase the final product cost, the power consumption and the product size.
The objective of this thesis is to propose and evaluate an inertial device mouse based on a three-axis accelerometer that can substitute the optical sensor in a hybrid device, in this way, only inertial sensors will be embedded in the system, reducing the total power consumption, the production cost and size.
1.2 Survey of Previous Work
A patent [1] claiming an inertial mouse system based on accelerometers was filed in 1988, describing that such mouse would consume less power than optically based mouse, and offer increased sensitivity, reduced weight and increased easy-of-use.
Since then, many accelerometer sensors were designed to be used for mouse applications Error! Reference source not found.. However, because the nature of
estimating the displacement based on acceleration signals is extremely difficult, there is not such a mouse device yet in the market competing with the optically based mouse.
The biggest challenge is to apply accurate signal processing methods to integrate the acceleration signal. This integration can be as simple as the one proposed in [7], or as complicated as in [2], which uses Kalman Filter. Another way to estimate the displacement is to use pattern recognition algorithms, as the one proposed in Error!
Reference source not found., which uses Fuzzy-Neural Networks.
Most of the research papers used as survey for this thesis propose signal processing techniques to be applied with accelerometers for different applications than mouse devices, such as robot positioning [3], gesture recognition Error! Reference
source not found., static balancing control of humanoid robots [20], detection of
small displacement for portable devices [18] and detection for human actions [19].Accelerometers also are common designed as a sensor for inertial navigation systems [17], especially in mobile robot applications, as in [21] and [22].
The other few papers that use accelerometers for mouse device systems are based on tilt angle [23], which instead of performing translation movements, as used in optically based mouse, the user must rotate the device to move the cursor on screen.
This method can be used in hand gesture recognition devices [24], in handicap assistant devices [25] and also in gaming devices [26].
1.3 Thesis Subject and Contribution
The subject of this thesis includes the design and construction of a prototype that contains an optical sensor and a three-axis accelerometer embedded in the same circuit board. The reason to have both sensors in the same board is to guarantee that they are under the same conditions and suffer the same displacement when moving the
prototype in a 2D surface. In this way, it is possible to compare the output of both sensors by applying the same input (the user’s interaction with the prototype).
In the computer side, the mouse driver software only requests the relative displacement in X and Y directions, which physically means the velocity in both coordinates. The extraction of the velocity from the optical mouse is straight-forward, since the sensor already returns the relative motion based on the successive images captured by its optical sensor. For inertial systems based on accelerometers, it is necessary to integrate the acceleration measured by the sensor. This integration can be performed in many ways. In this thesis, three digital signal processing methods are used to integrate the acceleration:
Fuzzy-Neural Network Estimator;
Kalman-Filter;
State-Machine based Filter;
The performance of each technique is determined by comparing their resultant velocity curve of X and Y directions with the optical sensor resultant velocity curve. A multiple comparison is possible by collecting data from the accelerometers and optical sensor during some seconds and processing it off-line using a software application running in the computer.
After evaluating the performance of each integrator techniques, only the best ones are ported to run in the prototype, which has a limited microprocessor.
1.4 Outlines of Thesis
The content of this thesis is organized as follows.
Chapter 2: details about the design and the construction of the prototype used in this project are described. The description includes information about the main components used on the circuit board, as the microcontroller, optical and
accelerometer sensors. It also includes the specification of the software running on the host and on the prototype.
Chapter 3: the integration techniques of the acceleration coming from the accelerometer sensor are described. For each technique, the mathematical model and details of the algorithm are presented.
Chapter 4: the experiment results are presented according to the developing steps of algorithms in chapter 3. Graphics containing the resultant velocity curve of each technique are presented, and the experimental results are discussed.
Chapter 5: the conclusion of this thesis and the possible improvement in the future is presented in this chapter.