Here is my gizmo for CW2. Just to point out I moved away from the initial voice recognition for calls for the simple reason that it is something standard in most phones and does not require an 800 word explanation. I talked about Bob's computer usage (because of course... he needs to use his pc!) and very briefly touched on how the arm can be used to move very small objects. Also, i must point out (because i was confused since yesterday about this) the 2D arm that i talk about for computer navigation is a virtual arm i.e. there is no physical arm but just a software.
I suggest we keep a little section at the end for rounding up other assisting technologies like the voice dialing and motorized shelves that come down for hard to reach areas and maybe label this section as "other features" that we mention briefly in a small paragraph rather than tackling them as large issues as we are doing with these big gizmos.
Let me know what you think and if its explain well because it was actually really confusing for me to understand at first :P
Cheers,
Maria :)
Using the Computer and Moving Small Objects
The Problem:
Having lost a lot of the movement in his arms Bob is found at a difficulty when managing day to day computer activities such as using the computer mouse - an activity that is so simple yet so hard for Bob. Several assistive technologies exist that provide an alternative to mouse control; with the most common being the physical joystick. Unfortunately, this may not be the best solution for Bob as his arms do not have the full control required to hold and move the physical joystick.
The Solution:
The proposed solution is to equip Bob with a vocal joystick to cater for the daily computer activities he needs to perform. House (2009) et al present a voice controlled computer-simulated arm, which is a software program that simulates the computer mouse. Although the most natural solution to control such a problem would be to have a natural speech recognition system integrated, such a system will be limited in commands and thus in smoothness of operation.
In order to provide more elaborate control, the vocal joystick works using the human voice in a non verbal manner, removing constraints caused by the natural language. House et al (2009) explain that it works with continuous vocal sounds with variations in pitch, amplitude and vowel quality for precise control of computer applications and electro-mechanical devices.
Using the 2D simulated arm, Bob will be able to navigate a WIMP (windows, icons, menus and pointers) interface by simply using his voice. House et al (2009) take it even further with a program called Voice Draw, which allows the vocal joystick to be combined with speech recognition for the drawing of objects on a computer. Such a functionality make it possible for Bob to explorer his artistic side regardless of his impairments.
Requirements:
To implement this non-invasive system, few hardware is needed. Mainly:
a standard microphone
a sound card
a computer
The main working force behind this system is a library with all the parameters used to control the joystick. The portability of such a library make it possible to use the vocal joystick with an unlimited number of applications providing Bob with more flexibility in the programs he uses.
How it works:
The vocal joystick extracts loudness, pitch and vowel quality and maps these parameters to controls that result in the desired action or motion. In order to simulate mouse control, House et al (2009) explains that each vowel corresponds to a specific direction, while the volume of the voice projected controls the velocity of the arrow movement on screen. Such control is achieved through 2D mapping and can be seen in figure 1.0 below. Mouse clicks are also created through the use of “ck” and “ch” sounds.
Figure 1.0: Vowel mapping to cursor movement for mouse control
Robotic Arm
House et al (2009) have also expanded this vocal joystick to presente the idea of a real 3D robotic arm which allows the user to move small objects as desired in real life. The 3D arm called the VoiceBot; works in the same manner, using the same parameters and allowing the user to manipulate and control each joint of the arm through an interface between the vocal joystick and robotic arm itself.
Hardware Overview:
The arm itself has five degrees of freedom which allows it to move. These include shoulder rotation, shoulder bend, elbow bend, wrist rotate and wrist band as can be seen below. At one end there is a 2-prong gripper for grabbing small objects. Communication between the computer and the arm is done through a serial port and a microprocessor controls servos at each joint.
Figure 1.2: Joint angles of Robotic Arm
How it works:
Bob will be able to control the 3D arm in two modes:
Position mode: this moves the entire arm using pitch and vowels for the direction, while loudness is used for speed.
Orientation mode: the wrist motion will allow for fine tune of the grippers’ positioning.
House et al (2009) explain that the vowels [ae] and [uw] are used to alter the wrist bend, while [iy] and [aw] to rotate the wrist.
Figure 1.3: Vowel sounds associated with arm movement
Although there is no feedback mechanism or sensors for the 3D robotic arm, the current solution can be an excellent prototype for future improvement. The current robotic arm will be sufficient for aiding Bob when his arms fail him at moving small objects, such as a glass across the table. The flexibility of the vocal joystick also ensure that Bob can use the computer without any problem, ensuring accuracy in movement with some practice, all allowing Bob to live a normal life through assistive living.
Reference: http://dub.washington.edu/djangosite/media/papers/vb_chi09.pdf
Videos:
http://www.youtube.com/watch?v=PXmccs4GIqI&feature=player_embedded
No comments:
Post a Comment