日本財団 図書館


Design of User-friendly Voice Control and Guidance System for Safe Navigation
Nam-Sun Son(Korea Research Institute of Ships and Ocean Engineering, KORDI)
Sun-Young Kim (Korea Research Institute of Ships and Ocean Engineering, KORDI)
 
 Abstract: As the technology of Speech Recognition (SR) and Text-To-Speech (TTS) develops rapidly, voice control and guidance system is thought to be very helpful for safe navigation. However, Voice Control and Guidance System (VCGS) is not yet so popularly included in Navigation Supporting System (NSS). The main reason of this is that VCGS is so complicated and user-unfriendly that navigation officers hesitate to use VCGS. Frequent errors in operating VCGS due to low rate of SR are another reason. To make VCGS more practicable for safe navigation, we design the user-friendly VCGS. Firstly, by using interviews we survey functions and procedures that navigation officers want to be included in VCGS. Secondly, to raise the rate of SR, we tune the environmental noise in bridge and to reduce the errors due to low rate of SR in operating VCGS, we design the functions of self-correction. In addition, we apply a user-independent SR engine so that procedures of learning of speakers are not necessary. Using simulator experiments the functions and procedures of the user-friendly VCGS for safe navigation are evaluated and the results of evaluation are fed back to the design. Therefore, we can design the VCGS more helpful for safe navigation. In this paper, we describe the features of the user-friendly VCGS for safe navigation and discuss the results of simulator experiments.
 
1. INTRODUCTION
 As the technology of Speech Recognition (SR) and Text-To-Speech (TTS) develops rapidly, voice control and guidance system is thought to be very helpful for safe navigation. Car Navigation System tells drivers current position using GPS, reduces the fatigues of drivers, and supports safe navigation [2,3].
 Voice Control and Guidance System (VCGS) of Ship have been developed for One Person Bridge Operation (OPBO) [4]. The purpose of VCGS in OPBO is to furnish the user-friendly interface of voice instead of manual interface, and reduce the load of simply repeated work in bridge. Therefore, navigation officers can acquire navigation information from VCGS in parallel with a lookout (i.e. multitasking) and cope with emergency effectively.
 From the middle of 1990s, research consortium of VCGS for coastwise vessels has organized in Japan. lBS on a basis of the research contained the voice reporting in emergency [4,5].
 However, Voice Control and Guidance System (VCGS) is not yet so popularly adopted in Navigation Supporting System (NSS). The main reason of this is that VCGS is so complicated and user-unfriendly that navigation officers hesitate to use VCGS. Frequent errors in operating VCGS due to low rate of SR are another reason.
 To make VCGS more practicable for safe navigation, we design the user-friendly VCGS. Firstly, by using interviews we surveyed functions and procedures that navigation officers want to be included in VCGS. Secondly, to raise the rate of SR, we tune the environmental noise in bridge. In addition, to reduce errors due to low rate of SR in operating VCGS, we added the several functions of self-correction such as user-friendly functions; automation, stand-by mode, exit mode, confirmation mode, and so on. In addition, we used a user-independent SR engine so that procedures of learning of speakers are not necessary. Using simulator experiments the functions and procedures of the user-friendly VCGS for safe navigation are evaluated and the results of evaluation are fed back to the design. Therefore, we can design the VCGS more helpful for safe navigation.
 In this paper, we describe the features of the user-friendly VCGS for safe navigation and discuss the results of simulator experiments.
 
2. ANALYSIS OF FUNCTIONAL REQUIREMENT OF VCGS
 There are five main navigational duties in the bridge; reporting the status of ship, controlling the speed and direction of the ship, acquiring the position information of other ships, controlling emergency and planning passages [1,8].
 
 In the manual procedure, if the master asks the navigation information to navigating officer, the officer answers to the master from looking out the navigational equipment. However, using the technology of Speech Recognition (SR) and Text-To-Speech (TTS), the master asks the computer through microphone, and the computer automatically reports to the master by voice with monitor displaying [6,7]. Fig.1 shows this conceptual diagram of voice input and voice output in VCGS. Here, we design the flow of voice and navigational information. In addition, we apply a user-independent SR engine so that a procedure of learning of speakers is not necessary. In addition, we use the database in forming the sentence of report that will be spoken out through the TTS engine.
 
Fig.1 Conceptual Diagram of Voice Input and Voice Output in VCGS
 
 We survey the functions of VCGS through referring to "Bridge Procedures Guide " [10] and interviewing with navigational officers and determine the realizable functions out of them with current technology. Table 1 shows the functional requirements of VCGS.
 
3. DESIGN OF USER FRIENDLY PROCEDURES OF VCGS
 In the former paragraph, we discussed the function requirement of VCGS. Here, we discuss the procedure and interface of VCGS. In the design of the procedure of VCGS, we must take care of determining the report type. Too frequent report of voice can make the bridge very noisy [1,8].
 
 We determine three main types of reports on a basis of references and interviews; Scheduled Report (SR), Query Report (QR), and Emergency Report (ER). Here, SR and ER are automatic report that does not need any question. The last column in Table 1 shows the type of report for each navigational function of VCGS.
 
 In the design of procedures, to reduce the errors due to low rate of SR in operating VCGS we focus on the user-friendly concept, i.e. automation instead of step by step, manual confirmation. There are three or four automatic steps in each report.
 
1. Recognizing questions using the SR engine.
 
2. Searching navigational information from database (ECDIS) or data packets.
 
3. Forming the sentence to report from the searched information.
 
4. Speaking report using TTS engine.
 
 Through these automations, users feel familiar in VCGS as they have a talk with a man.
 
 The general procedure of each report is shown in Fig.2. Procedures of Automatic Reports (AR), i.e. SR and ER, need only TTS engine because the procedures do not need any question of navigational officer that must be recognized in the computer using SR engine.
 
Table 1 Functional Requirement of VCGS
Navigational Function Data Base Contents Type of Report
Reporting Of Ship Status Passage ECDIS database Distance or time to(from) arrival(departure) Scheduled Report(SR)
Query-Report(QR)
Position GPS Longitude,latitude SR,QR
Direction GYRO Heading angle SR,QR
Speed Speed log Speed, Turning rate SR,QR
Controlling of Ship Steering gear, Auto-pilot & E/G Rudder, E/G
Thruster
Rudder Angle
E/G(thruster)RPM
SR,QR
Emergency Control Collision/Grounding Alarm ECDIS database, C/G Avoidance System Distance or time
To obstacle(other ship, buoy, land, grounding position)
Emergency Report (ER)
Track-off Alarm ECDIS database Distance or time
Off the planned route
ER
Disorder of Equipment Rudder, E/G , and etc Overshoot value ER
Information of Harbor or Environment Harbor ECOIS database
(Buoy, Breakwater, etc)
Name, distance and direction to the object
(Buoy, Breakwater, etc)
SR
Wind Anemoscope Anemometer Speed and Direction SR,QR
Current Tachometer Speed and Direction SR,QR
Depth Echo sounder Depth SR,QR
 
 In this procedure of AR, VCGS (computer) automatically forms the sentence of report using the database (terminology and navigational information) in scheduled time or emergency, and the sentence will be spoken out through the TTS engine.
 
 In the procedures of Query Reports (QR), if navigational officer ask a question to VCGS(computer), the computer recognizes the navigational terminology for the question, using the SR engine. Therefore, VCGS acquires the navigational information for the recognized terminology in the database (data packet). Finally, VCGS forms the sentence of report using the found navigational information and speaks it through the TTS engine.
 
Fig.2 The Procedure of report in VCGS
 
 Here, we apply some other user-friendly functions, i.e. self-correction to reduce the error, due to low SR rate and acquire the information quickly, in the procedure of QR.
 
1. We design three additional modes, for automatic changing between QR and AR; Stand-by mode, Exit mode I, and Exit mode II [6,7]. After the procedure of QR ends, VCGS automatically changes to the Stand-by mode so that VCGS can wait for the next QR or prepare for AR. In addition, we can use the Stand-by mode to stop the procedure on the way of QR. In the Exit mode I, we use just AR without QR. In the Exit mode II, we use only manual procedure without VCGS. Here, we must take manual action like the proper terminology or mouse button for changing to Exit mode.
 
2. We design the confirmation mode in the procedure of QR; repeating an order (question). As soon as the navigational officer (user) asks some question to VCGS, VCGS automatically repeats the question. Some seconds later, the next recognition step automatically continues. If the recognized terminology does not exist in the database VCGS asks the user to ask a question again.
 
3. The navigational officer (user) can consecutively ask additional question to VCGS as soon as the procedure of one QR ends.
 
 Through these user-friendly functions in the procedure of QR, the navigational officer (user) can quickly acquire the navigational information that he wants.
 
 Consequently, we apply the user-friendly automation and functional modes for self-correction in procedures of report so that the navigational officers (users) can reduce the errors due to low SR rate and use VCGS familiarly. Therefore, we can rapidly grasp the navigational condition of the ship.
 
 In addition, the navigational officers (users) reduce the fatigue of work in bridge using VCGS more largely than in manual procedure, so the users can carry out other work (multitasking) simultaneously with a lookout.
 
4. DEVELOPMENT OF USER FRIENDLY INTERFACE IN VCGS
 
Fig.3 The arrangement of VCGS in the virtual IBS
 
 We integrated VCGS with Conning Information Display (CID) for Integrated Bridge System (IBS). Therefore, the navigational officer (user) can quickly grasp the navigational condition of the ship and simultaneously control the steering gear and other navigational equipment comfortably. Fig.3 shows the arrangement of VCGS installed with navigational equipment in the virtual IBS. In addition, we connect VCGS with sound-card, microphone for voice input, and speaker for voice output. Data packet in the network offers every navigational information to VCGS in real time.
 
 Fig.4 shows the software interface of VCGS developed from this research. We integrate this interface with CID, but both its input and output are all voice types. However, we display the necessary status of VCGS.
 
Fig.4 The software interface of VCGS which has graphic display for collision risk
 
 The user-friendly concept also applied to the design of the interface of VCGS, that is, automation in changing between AR and QR, so that the navigational officers conveniently use the VCGS.
 
 Fig.4 shows only the minimum graphic information for voice interface, that is, voice input and voice output like "ALARMS". For navigational officers (users) to use VCGS with ease as they have a talk with a man, we applied the automation of voice interface which is included in the graphic interface. In addition, we made administrator of VCGS set the configuration of voice interface using the other utility.
 
 This interface has three types of reports as discussed in the previous paragraph; SR, ER and QR. The former two reports are done automatically. In the procedure of each report, users ask a question through the microphone of VCGS and get the report of voice from the speaker of VCGS like the real conversation in bridge. In addition, This interface has four functions of report in VCGS. The four functions from survey, are as follows; reporting the status of ship, controlling the speed and direction of the ship, acquiring the position information of other ships, controlling emergency and planning passages. Data packet of ECDIS and other navigational equipment in the network offers every navigational information to VCGS in real time.
 
 In the report of, specially, collision alarm in emergency, VCGS analyzes the real-time collision risk with obstacles [9]. Therefore, VCGS display the 3-color radial graph according to direction and level of collision risk; green for safe status, yellow for caution, and red for dangerous emergency. In addition, VCGS offer the voice alarm that contains the position and time to collide with obstacle.
 
 If the ship has a chance to collide with some obstacle, the alarm control of VCGS will flicker and the radial graph will show the risk and direction of collision.
 
 Simultaneously, VCGS will offer voice alarm to the user. If the user recognizes the emergency and takes proper actions, the procedure of emergency controlling ends.
 
 Consequently, VCGS that connects with CID offers report of voice for navigational information, so that the navigational officer (user) can quickly grasp the navigational condition and effectively cope with the emergency.







日本財団図書館は、日本財団が運営しています。

  • 日本財団 THE NIPPON FOUNDATION