21-08-2014, 11:17 AM
VOICE OPERATED INTELLIGENT ELEVATOR
VOICE OPERATED.doc (Size: 700.5 KB / Downloads: 28)
Abstract:
Now-a-days we can find huge number of apartments crowded in nearby area with multi storage building capacity .The apartment consists of several floors. This project better suits for all, mainly for physically challenged people. To move from one floor to another we generally use staircase or Elevator which is normally operated by using switches. The main purpose of designing this project is to operate the Elevator by using voice commands.
Elevator has to be moved vertically by recognizing our voice commands by using speech Recognition module. By giving different commands lift can be moved from one floor to another automatically.
If voice command is received successfully by Elevator an acknowledge will be coming in the form of beep sound from buzzer. In this way activation of buzzer depends on the voice command received by Elevator.
The project is built around AT89S52 microcontroller Voice Recognition module is
interfaced with microcontroller to drive the Elevator. A 16X2 LCD display is used for the display of Elevator where it is currently placed and its status can be obtained.
This project uses regulated 5V, 1A power supply. 7805 three terminal voltage regulator is used for voltage regulation. Bridge type full wave rectifier is used to rectify the ac output of secondary of 230/12V step down transformer.
INTRODUCTION
Rational and accurate movement control of separate mechanisms, devices and other parts is very important in lots of technological equipment.
Most frequently, executive devices are electric drives and collection, processing and forming of control algorithm is conducted by electronic-digital means in such systems. Automation specialists of the broad qualification are required for creating and exploiting such mechatronic systems.
A lot of up-to-date control elements and software is used for these specialists’ training. Therefore the model of the lift, which is suitable for training purposes because it contains all parts of a control system used in any industrial lift is presented in this article.
A lift is a characteristic electromechanical device controlled in electronic digital way containing controlled electric drives, differently realized positioning systems, quite a complex logic control system and having high requirements for reliability and safety.
In order to have a model, which is as realistic as it’s possible, it has an industrial design except the lift cage and its shaft model. It has no relay contact and up-to-date control elements are chosen for the model: cage drive is with the asynchronous motor and frequency converter, control logic is realized through easily programmed controller, sensors of the cage position are non-contact, control commands are given in a voice mode (reserve version is sensor programmed control panel).
Commanding/controlling of the computer and applications, using speech for handling the environment (smart house) are very promising fields, especially for disabled people. For those who have difficulties in entering data with other input tools like keyboard, mouse, etc., speech recognition is an effective alternative to alter or to combine input methods.
People with motor disabilities can control various devices via speech input. One of smart house elements could be the control of a lift by voice commands. There are four main factors related with speech recognition problem:
THE PROJECTION BASED SPEECH
RECOGNITION ALGORITHM
Model of a lift can be controlled by a small number of voice commands (floor number, stop and go), but the high recognition accuracy of voice commands is necessary.
Speaker independent recognition would be more convenient, because it would not be necessary to train the recognition program for a new user. But speaker independent recognition is more complicated than speaker dependent one and ensures less recognition accuracy.
Therefore speaker dependent projection based recognition algorithm was selected . It is based on very small number of non-uniformly spaced feature vectors in opposition to 5-10 msec stated speech preprocessing.
The projection based speech recognition algorithm was developed as a highly phonetically motivated alternative to DTW (Dynamic Time Warping) or HMM (Hidden Markov Models) speech recognition models.
The projection method is based on phonetically segmented speech features. The segmentation function is smoothed using linear and nonlinear filtering.
Each utterance then is divided to stationary and transient phoneme like segments. The average number of the stationary - transient segment pairs is approximately equal to number of phonemes per utterance.
CONCLUSIONS
1. The model of a voice-controlled lift is universal and it enables to realize regimes of real working lift by using up-to-date means.
2. A voice recognition program and its connection with the controller can supply a sufficient amount of commands necessary for the lift control.
3. The model of a lift is a useful tool for training students in specialization of automation, voice signal recognition and control technologies as well as for specialists’ qualification improvement in similar specialization.
4. Voice controlled systems are especially useful for disabled people. Speaker dependent projection based recognition algorithm ensures a sufficiently good recognition accuracy of voice commands.
It can be improved by increasing the amount of references and by selecting acoustically different voice commands. References can be collected from many speakers and averaged. The presented recognition algorithm in such way can be transformed into the “multi-speaker independent” one.