Possibility of Developing a Voice Recognition System in an Aircraft - Research Paper Example

? Possibility of developing a voice recognition system in an aircraft This paper seeks to research whether is it’s possible to come up with a voice recognition technology for use in aircrafts that are to be flown by disabled people. Since time immemorial, communication between air traffic controllers and pilots was done through voice radio which is prone to various errors and more so, a disabled person could not become a pilot as this is a profession that requires use of almost all parts of the body. But with an increase in air traffic, thanks to the continued technological innovation in the world today, new methods of communication in aircrafts have been devised in order to curb increased traffic along with the errors that occurred with voice radio and the need to incorporate even the disabled into the profession of piloting (Adams, 2009). Able Flight is a non-profit organization, which offers scholarships to people who have desire to training in the aviation industry. This organization works in conjunction with institutions that offers pilot training. Since it is non-profit, Able flight thrives on the generous donations of well wishers, corporate sponsorships and funding from the foundation. The IRS principles and policies are the ones that govern this organization. Its objective is, to help the handicapped and disabled to get the attitude and knowledge and ability to fly a plane. This is because the disabled need distinct solutions in aircrafts that will help them become pilots just like others. An instance would be the use of voice recognition systems in aircraft. This would ensure that the hands of the pilot are not need in order to communicate with air traffic controllers. Important to note is that, the Able flight program does not only assist the disabled but also looks to help those who are normal and have a passion to fly planes along with the interest but then lack the ability. The main aim is to look into how able flight tries to help the disabled fly planes along with offering solutions to the difficulties that come by during flight training. In the end, this paper will look at the process of implementing voice recognition technology in a pilot's communication with air traffic controllers as argued by (Chen, 2006). Able flight helps disabled people to fly airplanes in several ways. This organization however does not offer flight training classes but teaches life lessons to the disabled, which helps them become good pilots and flight attendants. Able flight works with different schools of flying such as the Purdue University. This organization has been working with the said university for the past two years and two students have been taken to the university and taught to fly. Each of the two students has a disability which to date has been a hindrance to them become pilots or looking into the aviation industry. Regardless of this, with help from the scholarships from Able flight and the special aircrafts, the two have gotten light sport pilot certificates after a 5-week training period. Jacky (2008) argues that the most interesting part of the whole program is that, scholars can cover the whole package in a month while the flight scholars at Purdue take about six months to finish the course. Due to the density of the schedule at Purdue, the scholars and flight instructors take most of their time at the airport, in the classroom and in their specially modeled aircraft (Jacky, 2008). A partial day engross an early morning arrival, 90 minutes of flying, debate, more flying and landings, lunchtime meals, even intense flying and ground institute lessons. Note that, all of these things take place at the Purdue University campus flight school, which does not mean that, Able Flight has no role that it plays. As far as those disabled scholars are concerned, Able Flight must participate or take part in ensuring that they obtain exactly what they went to acquire. The organizational program is developed in a way that, it has to follow up the students’ progress, know their strong points and motivate them to keep it up; identify their weak points and help them both find permanent solutions to them in order to make sure they progress. Generally, Able Flight makes follow up on all the disabled students it gives scholarships and makes sure they are safe and healthy. The school included time for social time, where some of the Able Flight staff goes there to spend sometimes with them as well as join them with the other flight students so that they can feel accepted and embraced in the community. The school’s training personnel also help the disabled students to participate in school’s activities and to mingle with others so that they do not feel lonely or rejected. Besides, the school’s staff and Able Flight made it possible for the students to live together in First Street Towers, which is a university’s owned dwelling hall. Whereas the Able Flight scholars continue learning to fly, the flight instructors also learn new techniques of teaching and concerning conquering adversity. Derek Stewart, a Purdue flight scholar and verified flight instructor stated that, “I wanted to get out of my comfort zone. I thought it would be a problem for me and a challenge to my teaching skills.” He added on that “Even within the first two weeks, I was able to see them develop as people and as pilots.” Therefore, other than just sponsoring and giving the disabled students scholarships, Able Flight give they full support in terms of both financial and mental support (Victoria, 2012). Voice recognition is basically defined as the ability of a device, mostly an electrical one to recognize an individual's voice, mostly used for security purposes. It is important to note that, a voice is unique to only one person, just like a fingerprint. Understanding fully the challenges that come with an efficient voice recognition that is not linked with the speaker, there are several fields that have to be explored: digital signal processing, linguistics and phonetics. In voice recognition, a method of matching templates is used which is founded on the general principles of basic computer programming and digital electronics. This program will be especially helpful as at times pilots and air traffic controllers may be unable to communicate effectively. Communication between these two sometimes becomes difficult as the pilot cannot multi task, turning knobs and flying the plane while at the same time communicating. Voice recognition as opposed to speech recognition looks to know the speaker is, rather than know what the speaker is saying (Chen, 2006). One of the major problems with voice radio communications used in this manner is that, all pilots being handled by a particular controller are tuned to the same frequency. As the number of flights air traffic controllers must handle is steadily increasing (for instance, Shanwick handled 391,273 flights in 2006, an increase of 5.4% - or 20,000 flights - from 2005, the number of pilots tuned to a particular station also increases. This increases the chances that one pilot will accidentally override another, thus requiring the transmission to be repeated. In addition, each exchange between a controller and pilot requires a certain amount of time to complete; eventually, as the number of flights being controlled reaches a saturation point, the controller will not be able to handle any further aircraft. England (2004) worked with Swedish pilots who were flying the JAS-39 Gripen and found that, recognition tended to deteriorate with an increase in G-loads. He more so noted that adaptation needs to improve the results in all the instances and introduction of models for breathing in the cockpit improve the levels of speech and voice recognition significantly. Spontaneous speech also seemed to be a cause of problems for the recognizer and there have so far been no effects of speakers who speak broken English. Something that could however be used to improve the voice recognition would be use of restricted vocabulary and a proper syntax as these could greatly improve accuracy (Victoria, 2012).However, researchers are looking to come up with speaker-independent designs of voice recognition programs. As we speak, there are some that are being tested for two fighter aircrafts: the F35 lightning II and the Alenia Aermacchi M-346 Master. These speaker independent systems have come up with word accuracies of more than 98%. High voice recognition accuracy is strained environment in terms of noise and stress, is completely possible but could be workable in a usual aircraft environment. Acoustic noise in an aircraft is the greatest challenging in achieving perfection in voice recognition. In a helicopter for instance, the levels of voice recognition will be very poor since the pilot does not wear a mask so acoustic noise in the microphone is very high. Various countries including, Canada, UK, US, France, Sweden and others, have carried out a lot of research in the voice and speech recognition field in a bid to attain perfection, the results have been good with voice recognition being used in various places such as: controlling communication radios, setting navigation systems, and controlling automated target hand over systems. Just as was seen in fighter aircrafts, overriding of voice was also an issue in helicopters. Usually, voice and speech recognition gets more difficult as the vocabulary used by the speaker and the more similar, phonetically that the words spoken are. When it comes to word sequences, language models or artificial grammar, these are used as a means to limit the number of word combinations. A finite network can be defined as the simplest language model where words that can be used for a specific word are explicitly defined.  Voice recognition takes two main approaches the use of templates for matching or analysis of the speaker's features. Template matching is the simpler of these two techniques and it more so has the highest accuracy if and when it is used properly. Regardless of this, it has the greatest limitations. The first step of the voice recognition approach involves the user speaker uttering words into a microphone. The signal from the microphone is then digitized by a converter from analog to digital and stored into memory (Victoria, 2012). This is done in order to determine the meaning of the word after which the computer tries to match the input with a voice sample or template that has been digitized. The digitized sample of voice or template has a standard and known meaning. The fact that each person's voice is unique to him/her, the program for voice recognition cannot have a template for each and every user, means that the program has to be trained every time there is a new user's voice so that the program can be recognized by the program. In a session of training, the program shows a word or phrase that has been printed and the user utters the word a few times into the microphone (Cater, 2009). The program computes a statistical average of the multiple samples of the same word and stores the averaged sample as a template in a program data structure. With this approach to voice recognition, the program has a "vocabulary" that is limited to the words or phrases used in the training session, and its user base is also limited to those users who have trained the program. This type of system is known as "speaker dependent." It can have vocabularies on the order of a few hundred words and short phrases, and recognition accuracy can be about 98 percent. A more general form of voice recognition is available through feature analysis and this technique usually leads to "speaker-independent" voice recognition. These similarities will be present for a wide range of speakers, and so the system need not be trained by each new user (Chen, 2006). There are some types of speech differences that the speaker-independent method can deal with, but which pattern matching would fail to handle, include accents, and varying speed of delivery, pitch, volume, and inflection. The speaker-independent speech recognition system is quite difficult as it has quite a number of hurdles having a variety of accents from speakers of various nationalities. The accuracy of voice recognition in speaker-dependent systems is less than that of speaker-independent systems by about 90 to 95 percent (Jacky, 2008). It is furthermore of great importance to note that there are various voice recognition systems that are available. The one that has been implemented so far is the one that can handle discrete conversations. This differentiation is proven by determining whether they can handle either of the three: connected word, discrete words or continuous speech. Majority of the voice recognition systems that are in use today adopt the discrete word systems, which are usually the easiest to implement. In this system there must be pauses between the speaker's words. This system is perfect for conversations whereby the speakers is generally expected to give single word responses but can't work where multiple responses are (Fourcin, 2009). On the other hand, in a connected word voice recognition system, the speaker can speak in multiple words or phrases though he must be careful not to slur towards the end of one word into the beginning of the next word. Implementing naturally continuous speech is very troublesome, almost impossible since it needs a lot of co-articulation of words which basically means that words adjacent to each other run on together without pauses or sign of division (Martin, 2006). Conclusion In conclusion, it is of importance to note that much needs to be done in the voice and speech recognition field before it can be said to be effectively used by air traffic controllers in communication with pilots and more so achieve improved performance in operational settings. Since most voice recognition systems cannot handle continuous speech or prose, it is important to look into further research on how these systems can be improved so as to be able to handle prose as well as one response commands. Moreover, acoustic noise is another great challenge of voice recognition systems, thus it is important to look for ways that this noise can be reduced, especially in fighter aircrafts and helicopters where the pilots do not wear masks. The use of voice recognition in aircrafts will not only be beneficial to the normal pilots but more so to the disabled pilots. This is because they will be able to work in a career that was pre-dominantly for those who were normal. Able Flight is doing a great job in making the disabled an active part of the society and making better people. Moreover, voice recognition would help Able flight to go even further in helping the disabled get to do what their hearts desire and more so pursue their passions in the aviation industry (Chen, 2006). References Adams, R. (2009) Sourcebook of Automatic Identification and Data Collection. New York: Van Nostrand Reinhold. Chen, K. (2006). How Voice Regontion Works. chicago: Chicago universty press. Cater, P. (2004). Electronically Hearing: Computer Speech Recognition. Indianapolis: Howard Sams & Co press. Fourcin, A. Harland, G.(2009) Speech Input and Output Assessment.UK: Ellis Horwood Limited. Jacky, B. (2008). Speech Synthesis in Response to Human Command. london: Adventure Works Press. Martin, K. (2006). How Voice Regontion Works. Nairobi: UoN . Victoria, W. (2012). Testing Speech in Mechanical Auto Mobiles. Thika: MKU press. Yannakoudakis, E. Speech Synthesis and Recognition Systems. UK: Ellis Horwood Limited. Read More