The Positive Impact of Voice Recognition

This article originally appeared in the ATC Network Special Bulletin 'Training and Simulation - September 2011. To read the complete publication go to:
http://www.atc-network.com/training

Greg Thibeault, Assistant Professor, Daniel Webster College, School of Aviation Sciences
United Two Sixteen Heavy turn left heading three six zero, descend and maintain six thousand verify information Mike. This is an example of a clearance issued by an Air Traffic Control (ATC) student running a simulation exercise. Typically, the clearance is read back and executed by a pseudo-pilot. However, at Daniel Webster College (DWC), Ray, one of 16 embedded voice files, reads back the clearance.

Execution of the clearance occurs internally, without the assistance of a pseudo-pilot, and the student moves on to the next task. The benefit of Voice Recognition and Response (VRR) has been realized. As we enter the second decade of the 21st century, voice recognition is firmly entrenched and universally accepted in many facets of society. Recall the last time you phoned an airline to check on your flight status or reservation? How about your drive to work today? Did you use voice recognition to change your music selection or to interact with your navigation system? In order to improve and expand the services offered to an increasingly technical and savvy customer base, while simultaneously remaining competitive in the marketplace, both academia and industry leaders have embraced VRR. Why is VRR rapidly becoming a staple of ATC simulation? Simply, VRR has matured, and is being used regularly in the education and training of the next generation of controllers.
According to the Federal Aviation Administration (FAA), Revenue Passenger Miles (RPM) or one paying passenger travelling one mile, are projected to more than double over the next two decades, from 787 billion to 1.7 trillion in 2031. U.S. airlines will reach the one billion passengersper- year mark by 2021, two years earlier than last years prediction of 2023.1 For more than 25 years, DWC has been a major player in the field of educating tomorrows aviation workforce; especially in the field of Air Traffic Management (ATM). Like any career field that is experiencing a shortfall of qualified employees, an ATC-based education is in great demand. (This is not unlike what the field of nursing is currently experiencing due to the retirement of the baby-boomer generation.)2 In the face of this demand, DWC-ATC has positioned itself to meet the influx of students while ensuring the programs high standards are being adhered to. Simulation, and in particular VRR, has been a critical ATM program component.
For generations ATC simulation has been conducted by having an individual work as an air traffic controller transmitting clearances to a pseudo-pilot, or as the United States Navy references them, bug runners. The pseudo-pilot, at his own work station, listens to the clearance, responds as areal pilot would, and then makes keyboard entries to ensure the simulated target performs as cleared. In a VRR simulation environment, instead of using a pseudo-pilot, tasks are accomplished by the software installed in the simulation system. The software recognizes the grammar spoken by the air traffic controller, converts the voice clearance into a text message using the same format that a pseudo-pilot would apply via keystrokes and voila! The clearance is executed.

Within the last 5 years, VRR use and available functionality has increased exponentially. An original and current member of the FAA Controller Training Initiative (CTI) program, DWC prepares students for careers as Air Traffic Control professionals. VRR has been an extremely effective tool faculty and students interact with. Working alongside UFA Inc., a worldwide ATC simulation provider, VRR innovator and leader, DWC students have reaped tremendous benefits from the integration of VRR and ATCOACH-Global Edition (GE). Two prominent Air Navigation Service Providers (ANSP) NAV CANADA and Deutsche Flugsicherung GmbH (DFS) have taken notice.
The simulated radar environment at DWC is a four corner post TRACON, with similar airspace that is being used daily at Level 11 and 12 TRACON facilities in the United States.
Besides having a number of secondary airports in the airspace, the primary airport, Manchester, has four runways; 18R/35L, 18L/35R, 08/26, and 07/25. This set-up allows for a wide range of airspace and air traffic concepts to be taught, with a heavy hand from VRR. For instance, from a simple single runway instrument or visual approach configuration, to a complex simultaneous independent parallel runway and/or converging runway, the simulation lab hums along, mimicking a working facility.

The DWC radar laboratory has four Radar Control positions, four Radar Associate positions, and four Pseudo-Pilot positions. During the current 2010-11 school year, lab demand has been at an all-time high. Specifically, demand for AT310 and AT315 ( Junior year ATC courses) and AT410 and AT415 (Senior year ATC courses) (see table).
article5 ufa image 1 (2)

From the numbers presented, it is evident that simulation demand and expense could be overwhelming. If DWC employed pseudo-pilots in lieu of VRR, and demand remained constant, the current program could not be offered because it would be too costly. Specifically, if four pseudo pilots were employed during each scheduled lab session, and each earned $45 for each two hour and fifty minute per lab (a $15.90 hourly rate), annual salary costs alone would be $50,400 for 280 scheduled sessions. The expense for pseudo-pilot support for the practice time each student is required to invest would be $100,170. Thus, excluding annual volunteer and tutoring lab use, which is considerable, VRR is saving DWC over $150,000 annually. Imagine the financial benefits a large ANSP might experience?
Financial benefits are worthless if the quality, quantity, and variety of ATC concepts offered are diminutive. In other words, it is fair to say that deploying VRR has to equal and/or exceed what pseudo-pilots bring to the table. DWC has worked extensively with UFA software engineers to ensure their VRR product ATVoice contains extensive functionality and is easily updatable.

Consider what occurs thousands of times a day in the United States when an aircraft checks on frequency with a departure controller. The aircraft calls departure, usually stating what altitude it is leaving and the altitude it is climbing to on a specific heading and is cleared on a published Standard Instrument Departure (SID). The controller needs to be able to apply the various approved radar identification methods ,ensure mode C accuracy, confirm altitude assignment, vector the aircraft to join its cleared routing, perhaps issue a speed restriction, issue a traffic call applying merging target procedures, and ultimately issue a frequency change as the aircraft is handed off to the next sector. Employing ATVoice with ATRadio, DWC students can not only do all of these tasks, but they can do them in any order desired, while using standard phraseology as stated in FAAO 7110.65T, completely free of live pseudo-pilot support.

Another typical scenario is a simulated arrival rush. Envision simulating an approach wall at a busy TRACON. The wall might have two to three Feeder and Final positions open in order to provide service to an arrival rush that equals an Airport Acceptance Rate (AAR) of 72 -80 aircraft to parallel runways, 36L and 36R. The final controller for runway 36L is conducting ILS approaches while the final controller for runway 36R is conducting visual approaches. The following information and clearances might need to be issued: ATIS verification, altitude and speed assignments, wake turbulence advisories, instructions to intercept the localizer, PTAC ILS clearances, clearance to maintain visual separation from the parallel runway traffic with instructions to follow the preceding traffic that has been reported in sight, and runway assignment changes. These clearances and many more are practiced over and over again by students at DWC solely using VRR.

According to Thibeault [first person?], Having worked most of my ATC career at a busy TRACON, one of the most important aspects I needed in order to both achieve CPC status and remain an accomplished controller was repetition. He continues that, Working similar situations over and over again allowed me to gain the experience necessary to excel. Not only does VRR allow this repetition to occur at an extremely low cost, but most importantly, it allows an individual to learn how to correctly apply classroom concepts and theory in a greatly reduced period of time. For example, a student might have difficulty learning how to
run an efficient ILS final and applying the required PTAC phraseology. In a pseudo-pilot driven environment, the student may get the opportunity to practice on the simulator a few hours a day or week. With VRR, the student can practice as many hours as the simulator is available. If the student is also using ATCoach, the student can review recorded simulation runs performed by instructors-to see how it should be done and/or they can review their own recorded simulation runs so that they can see and listen to how they have performed.

The student quickly sees what improvement may still be necessary. Students are often asked to review a recorded simulation run which has purposely 10 to 15 errors. Errors might include incorrect phraseology or proceduresin the recording. Students are then assigned the task of viewing and listening to the scenario. Upon completion they must document all of the errors, provide the proper references for each, and state how the operation should have been conducted. This type of assignment has proven to be highly effective in shortening the time required to learn required tasks.

On paper VRR appears to be an ideal tool for the next controller generation. In 2006, the FAA requested that NASA-AMES Research Center assess the effectiveness of VRR being used on a STARS simulation system. The research was conducted in the training room of the Boston Consolidated TRACON, located in Merrimack, New Hampshire. Twelve scenarios actually used in training varying in complexity and type were run in two phases. Boston TRACON Certified Professional Controllers (CPCs) were used to run these scenarios, Both VRR and pseudo-pilots were used, and data points were collected. VRR evaluation measures included scenario development time, VRR errors, and the impact of errors on training. Results of the study indicated that VRR is acceptable for TRACON instructional purposes and it performed well as compared to the existing Pseudo pilot based system.3 the NASA-Ames team recommended that VRR be tested using developmental controllers exclusively. NAV CANADA has not only done testing, but has begun VRR installation. Deutsche Flugsicherung GmbH (DFS) is also in the early stages of deployment.

NAV CANADA has integrated VRR into simulated Air Traffic Control Tower (ATCT), Terminal, and En-Route environments. The company is well underway with deployment plans in each of its seven Area Control Centers (ACCs). Developmental controllers, who are selected for either a Terminal or En-Route option, receive generic IFR training followed by specific IFR training for their area of responsibility. Working directly with UFA, NAV CANADA has been able to redesign the grammar the VRR engine must recognize so that it can be used by its developmental workforce while also meeting ICAO requirements. There are many reasons why NAV CANADA pursued VRR.

Professor Thibeault recently interviewed Garry Brown, NAV CANADA Manager Operational Training Initiatives, to discuss VRR.

Thibeault: Why did NAV CANADA pursue VRR for its training platform?
Brown: To reduce the bottleneck restriction a centralized training facility placed on student throughput, and to enable more student practice without a support resource dependency.
Thibeault: Where are you now concerning implementation of VRR at your facilities?
Brown: Generic IFR training is the process and has been deployed at five of seven sites. Generic Terminal Training has been delivered at one site and is the pre-requisite to the Specialty VRR.
Specialty IFR VRR has been installed at Vancouver. Moncton and Winnipeg are scheduled for Spring 2011, Gander and Toronto will be Fall 2011, and Montreal requires a bilingual solution.
Thibeault: Have you seen that your phased implementation plan of doing Generic IFR first and then Specialty IFR is the proper course to follow?
Brown: NAV CANADA developed a Generic Terminal Course that leveraged an ATCoach Generic Terminal Voice module that has proven to reduce the Specialty training requirement. The
Generic portion also ensures the students capability using VRR is improved and they do not have to re-learn another process.
Led by the dual efforts of Karl-Heinz Steffens, Ph.D., head of the DFS ATM Simulator Center, and Michael Slotty,ATM Simulator Center Project Manager, DFS has begun the process of integrating VRR into its training process. In the summer of 2011, training will begin for over one hundred developmental controllers who will be working high altitude sectors, specifically Flight Level Three One Five and above. In preparation of using VRR exclusively, field trials are currently underway at Munich Center.

According to Uwe Gaudlitz, Munich Proficiency and Simulation Manager, trial results are exceeding expectations. From the first day this project began, I have been amazed at how many diverse clearances can be issued and how efficiently VRR accepts these clearances using English with a German accent, he says. The flexibility we will gain in our training program will be enormous. No longer will training be negatively impacted by a lack of pseudo-pilots, and lab availability. Developmental controllers will have much more access to simulation equipment, which will result in a reduction of the time necessary to be in training. The future of VRR is exciting and promising. Tremendous strides have been made in the last few years which have positively impacted the training and education of tomorrows controller workforce. It should definitely be a component that is available to meet the increased demands aviation will bring in the decades ahead.
1 Adminstration, Federal Aviation. (2011, February 15). FAA Forecast Predicts Air Travel to Double in Two Decades. Retrieved February 17, 2011, from FAA Press Release: http://www.faa.gov/news/press_releases/news_story.cfm?newsId=12439
2 American Association of Colleges of Nursing. (2010, September). Nursing Shortage Fact Sheet. Retrieved January 29, 2011, from American Association of Colleges of Nursing: http://www.aacn.nche.edu/media/FactSheets/NursingShortage.htm
3 UFA Inc. (2007, April). NASA-AMES Voice Recognition and Response Evaluation Report. Retrieved February 12, 2011, from ufainc.com: http://www.ufainc.com/news/57/

About the Author
Greg Thibeault is an Assistant Professor with Daniel Webster College (DWC) based in Nashua, New Hampshire. Professor Thibeault has over 29 years of experience within military and civil Air Traffic Control and Academia. Currently, Thibeault manages and teaches in the DWC School of Aviation Sciences, Undergraduate Air Traffic Management Program and the Graduate MBA for Aviation Professionals Program. Currently, Thibeault helps manage and teach in the DWC School of Aviation Sciences, Undergraduate Air Traffic Management Program and the Graduate MBA for Aviation Professionals Program. He also serves as an Air Traffic Control Subject Matter Expert for UFA and other corporate clients.
Contact
From
ATC Network
Website
www.atc-network.com
Date