The new aircraft noise monitoring system for 3 airports in the Osaka area. Kazuyuki Shibuya 1 Kenji Matsubara Yoshio Nishino Junshi Izumi

Kansai Airports Co., Ltd. 1-banchi, Senshu-Kuko Kita, Izumisano-shi, Osaka 549-8501, Japan Etsushi Fujita Taichi Higashioka

Nihon Onkyo Engineering Co., Ltd. Data Science Div. 1-21-10 Midori, Sumida-Ku, Tokyo 130-0021, Japan

ABSTRACT There are three airports in the Osaka area and each of them has been analyzing and evaluating aircraft noise and publishing its results. However, since these three airports are adjacent to each other, the airspace is complicated, and the monitoring range is wide, which caused so much manual work for data review before obtaining an accurate noise impact assessment. This paper introduces a new aircraft noise monitoring system that has been constructed to reduce the human workload and accurately and integrally monitor aircraft noise at the three airports with flight-track measurement. The new system reduces the time for noise analysis and evaluation using artificial intelligence, and it has become possible to publish accurate noise conditions more quickly and at a lower cost. In addition, since it is difficult for an airport operator to obtain air traffic control data, we constructed a system that can monitor the position coordinates of aircraft as recent as 15 seconds ago. By oper- ating this system integrally with the noise system, we can now visually confirm the noise level at the measurement point and the position of aircraft on a map, and publish the real-time noise data online for residents living near the airports.

1. INTRODUCTION

The Kansai region in Japan is home to several tourist cities, including Kyoto, Nara, Osaka, and Hyogo, and has the world's 10th largest hinterland with a population of 19.28 million 1 . Kansai Air- ports operates three airports in the Kansai region, Kansai International Airport (KIX), Osaka Interna- tional Airport (ITAMI), and Kobe Airport (KOBE), to support tourism and business demand in the region. A location map of the 3 airports is shown in Figure 1.

1 kazuyuki.shibuya@kansai-airports.co.jp

worm 2022

worm 2022

ITAMI

Central

Osaka

KOBE

KIX

Figure 1: Location map of the three airports (Photo from https://maps.gsi.go.jp/)

ITAMI is a highly convenient urban airport located 11 km from central Osaka. ITAMI used to be the only airport in the Kansai region, and during the period of rapid economic growth in the 1970s, ITAMI grew dramatically, but at the same time, serious noise pollution became a social problem. Residents living near the airport filed lawsuits demanding compensation, and also petitioned for pol- lution mediation, demanding the abolition of the airport. As a result, the government voluntarily de- cided to change the operation hours to be limited from 7:00 a.m. to 9:00 p.m. This operation has continued to the present day. In addition to restricting operating hours, ITAMI is currently imple- menting source control measures such as limiting aircraft movements, setting landing charge to en- courage the introduction of low-noise aircraft, introducing an operation system to reduce aircraft noise, setting priority flight paths, and reducing the use of reverse thrust during nighttime landings. In addition, in areas around airports where noise impact is significant, efforts are being made to com- pensate for relocation, develop buffer green zone, and subsidize soundproofing work.

In order to solve the aircraft noise problem around ITAMI, an airport with no noise was required in the Kansai region, and KIX was built in 1994 on the sea 5 km off Osaka Bay. KIX was built on the sea based on strict environmental assessment such as noise, and the company is operating KIX with an environmentally friendly manner by using routes that have as little impact on the land as possible. On the other hand, ITAMI was considered for closure due to its noise problem, but the local government and local residents' groups agreed to keep ITAMI in operation on the condition that it would be maintained for the time being, with restrictions on operating hours and the number of air- craft movements, since ITAMI plays a significant role in stimulating the local economy and in provid- ing convenience as an urban airport.

As a result of these backgrounds, KIX has become an international gateway airport mainly for western Japan with a concentration of international flights, ITAMI has become a core airport for domestic flights in harmony with the environment, and KOBE, which opened as an offshore airport in 2006, has become a regional airport serving the aviation needs of the Kobe area.

Therefore, even today, ITAMI is required to properly monitor the impact of aircraft noise, while KIX and KOBE place importance on monitoring the impact of aircraft noise on the land area. There- fore, KIX, in particular, has established an environmental monitoring plan in order to achieve an airport with no noise pollution, checks the aircraft noise situation on a daily basis, and publishes detailed noise values.

The three airports adjacent to each other in the Kansai region create a complicated airspace, and the large number of measuring stations (32 in total), especially some of stations for KIX, which are

located far from the airport, and the small difference between background noise and aircraft noise, have made it necessary for a large amount of human work for data review for accurate noise impact assessment. To reduce this human work, a new aircraft noise monitoring system was established to evaluate the aircraft noise at the three airports accurately and in an integrated manner by introducing new technologies such as wake (position coordinates) measurement. An overview of the system is presented below.

worm 2022

2. OVERVIEW

The new aircraft noise monitoring system (hereafter referred to as the "new system") consists of 32 measuring stations and a central station (Figure 2). Seven of the stations, which are located at a distance from the maintenance base, are equipped with IP cameras so that the status of the micro- phones and sensors can be checked in real time. Besides, 4K resolution camera is installed at the station located inside ITAMI to check the runway operation 2 . In addition, the new system is con- nected to a flight log data receiving line and the Internet, and has functions for automatic acquisition of flight log data, automatic publication of noise data, and remote access to the system (Figure 3).

× 11 × 2

Measuring Stations

Operation device

× 9

PCs (KIX, ITM, UKB)

ITAMI

× 10 *

× 4 * × 1

× 5

Servers and storages

KIX

KOBE

* Including 1 station which shared by KIX and KOBE.

Figure 2: System configuration chart

The new aircraft noise monitoring system Related external system

Central station

Flight log data souse

Internet VPN

Werther data

Report publishing server

Real-time publishing server

Operation devices * Measuring stations

* Can connect to central station both VPN and Internet which convenient.

Figure 3: Network configuration chart

In addition to noise measurements, transponder response signals emitted by aircraft and the signals of Ground Proximity Warning System (GPWS) are measured at each station. These measurements enable the followings.

 Automatic determination of aircraft noise  Determination of the time of the aircraft's closest approach to the station  Automatic matching of noise data with flight log data.  Determination of aircraft flight position. Ex.) Flightrader24 (https://www.flightradar24.com/)

The data measured at the station is sent to the central station via IP network for aggregation pro- cessing. The functions of the central station are shown in Table 1. The GUI for the various functions can be accessed from anywhere in the world via the Internet from an authenticated Web browser, ensuring business continuity. It also has a function to publish measured quasi-real-time noise values on the website.

Table 1: Functions of the Central Station A. Aggregation of noise

1. Noise data registration 2. Flight operation-related information registration * 3. AI judgment of noise type 4. Matching of noise and flight log data B. Takeoff/ landing runway determination * C. Noise scrutiny and measurement result output

 Judgment of noise type, correction of noise/ operation  information linkage, and output of Daily/ Arbitrary period/

Monthly/ Yearly reports. D. Wake calculation E. Real-time processing of noise and wake F. Automatic data publication G. Provide Web GUI of the system * Flight log data provided by airport operation and the results of this sys-

tem's takeoff/landing runway determination 3 including two-point time difference method 4 . From the next section, following sections state AI decisions of the A. and B. of Table 1, D. inte- gration of the wake system, and E. real-time processing functions.

3. AI ANALYSIS SYSTEM

The new system incorporates two AIs: noise type determination and takeoff/landing runway de- termination.

3.1. Aircraft Noise Comprehensive Evaluation

When evaluating the impact of aircraft noise, it is necessary to discriminate automatically meas- ured noise into “aircraft sound” or “non-aircraft sound”. This process is required because there are some cases which non-aircraft noise triggers the noise measurement, or non-aircraft noise is occurred over the aircraft noise. This work has been done manually to organize the data with high accuracy, but the time required for this work has been a problem.

The AI calculates the probability that the measured noise is an aircraft noise, and expresses the certainty of the noise as a numerical value ranging from 0 to 100%. The results derived by the AI are also displayed on the system's GUI, and the judges only check the noise for which the AI indicates around 50%, which significantly reduces the time required for the discrimination process.

worm 2022

Non-aircraft Aircraft

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

Relative number of data

0 10 20 30 40 50 60 70 80 90 100

The percentage derived by AI [%]

Figure 4: Result of AI analysis of data measured at one of the measuring stations

located in ITAMI area

Table 2: Range of percentages reconfirmed by humans and their effect on the evaluated value.

(Unit: dB) Lower threshold [%]

Upper threshold [%] 100 95 90 85 80 75 70 65 60 55 50 0 0.00 0.69 0.72 0.73 0.73 0.74 0.74 0.75 0.75 0.75 0.75 5 0.13 0.82 0.85 0.85 0.86 0.86 0.87 0.87 0.88 0.88 0.88 10 0.19 0.88 0.90 0.91 0.92 0.92 0.93 0.93 0.93 0.94 0.94 15 0.22 0.91 0.93 0.94 0.95 0.95 0.96 0.96 0.96 0.97 0.97 20 0.23 0.92 0.95 0.96 0.96 0.97 0.97 0.97 0.98 0.98 0.98 25 0.23 0.92 0.95 0.96 0.97 0.97 0.97 0.98 0.98 0.98 0.99 30 0.24 0.93 0.96 0.96 0.97 0.97 0.98 0.98 0.99 0.99 0.99 35 0.24 0.93 0.96 0.97 0.97 0.98 0.98 0.99 0.99 0.99 0.99 40 0.24 0.93 0.96 0.97 0.98 0.98 0.98 0.99 0.99 0.99 1.00 45 0.25 0.94 0.96 0.97 0.98 0.98 0.99 0.99 0.99 1.00 1.00 50 0.25 0.94 0.97 0.97 0.98 0.98 0.99 0.99 1.00 1.00 1.00

Table 3: Percentage ranges that judges reconfirm and labor savings ratios.

(Minimum:0.00, Maximum:1.00) Lower threshold [%]

Upper threshold [%] 100 95 90 85 80 75 70 65 60 55 50 0 0.00 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.04 0.04 0.04 5 0.00 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.03 0.04 0.04 10 0.00 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.03 0.04 0.04 15 0.00 0.01 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.04 0.04 20 0.00 0.01 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.03 0.04 25 -0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.03 30 -0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 0.03 0.03 35 -0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.03 0.03 0.03 40 -0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 0.03 45 -0.01 0.00 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 50 -0.02 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.02 0.02 0.02

worm 2022

Figure 4 shows the results of AI analysis of data measured at one of the measuring stations located around ITAMI in September 2021. The horizontal axis of the graph shows the percentage derived by AI, and the vertical axis shows the relative number of data “actually aircraft noise” in blue and “ac- tually non-aircraft noise” in red. Since there is a difference in the number of measured noise events, with 3921 aircrafts and 1322 non-aircrafts, the vertical axis of the graph shows the relative frequency to the total number of events for aircraft and other, respectively.

Table 2 shows the impact on the evaluated L den when the percentage of human reconfirmation is changed at this station, and Table 3 shows the labor-saving rate in that case. For example, when the judge reconfirms only 40% - 60% of the events, the impact on L den is +0.02 dB, or 99% labor savings compared to the case where a human checks all the events. Since the accuracy of AI varies depending on the environmental conditions in which the stations are installed, we consider the range of percent- ages to be reconfirmed for each station. This way, it is possible to reduce human labor while mini- mizing the impact on the evaluation value.

3.2. Takeoff/landing runway determination

We have been using takeoff/landing runway determination system (DL-TLS) 3 from the conven- tional aircraft noise monitoring system. This enables us to link flight log data provided by airport operator and noise measuring data at each station and analyze detailed noise by type of aircraft or runway used. The conventional method logically analyzes the change pattern of radio wave infor- mation at the end of runways when aircrafts take off and land, however, if it’s not the expected pat- tern, misjudgments or ghost records occur. In the new system, in addition to improving judgment accuracy by introducing new technology and improving judgment processing, the introduction of AI has made it possible to calculate the accuracy level of takeoff/landing runway judgment results. Since judges scrutinize only data which result of the system is not sure, the new system contributed to labor savings in terms of operation.

4. INTEGRATED WAKE SYSTEM

Aircraft flight path is an important factor in aircraft noise assessment and noise control. However, in Japan, it is difficult for airport operators to obtain Air Traffic Control (ATC) data, so in this renewal of the system, wake acquiring system was established. By integrating the noise system with the wake system, it is now possible to grasp the actual status of aircraft noise in an integrated manner.

The new system uses three methods: Automatic Dependent Surveillance-Broadcast (ADS-B), Pas- sive SSR (PSSR) and Multilateration (MLAT) to figure out flight path, and the wake can be confirmed by utilizing advantages of each. ADS-B and PSSR are used to determine wake in the surrounding area, and PSSR is mainly used to determine the wake of aircraft not equipped with ADS-B. MLAT is used to determine detailed flight positions, especially in the vicinity of ITAMI. Acquisition of the transponder response signal, which is used to calculate the wake, is a standard function of measure- ment stations, so the wake can be measured without additional investment. In the event of a malfunc- tion of the measurement equipment, recovery can be made by referring to data from other stations.

In addition to the identification data measured at the station, the system can display the wake in conjunction with the noise to grasp the flight status of aircraft around the monitoring point at the time of noise occurred and seamlessly investigate the aircraft that caused the noise and its flight path (Fig- ure 5).

worm 2022

+ 2022/04/26 09:35:21 gsm ao [ess on 527 es) coo II ~~

Figure 5: Example of displaying noise in conjunction. (Example of MLAT wake)

5. REALTIME PROCESSING

The new system, with the introduction of real-time processing technology, enables to check the instantaneous noise level at each station, and the wake by ADS-B and PSSR with a 15-second delay (Figure 6). This makes it possible to smoothly grasp the constantly changing noise values and flight positions of aircraft. Especially in ITAMI, where operation time is limited, it is necessary to accu- rately determine the position of aircraft landing near the end of operation time, and real-time moni- toring is a function that has been required by airport operators for a long time.

worm 2022

aie

Figure 6: Integrated real-time display function of noise and wake. (Example of ADS-B)

Furthermore, a quasi-real-time publication function was implemented to publish the available noise figure on the Web with a three-minute delay. By allowing anyone to view the instantaneous noise values at each measurement station on the Kansai Airports website, the system has contributed to more open airport operations (Figure 7).

Figure 7: Noise real-time disclosure (http://www.kansai-airports.co.jp/noise/)

KANSAI AIRPORTS 202020 O15

6. CONCLUSIONS

A new aircraft noise monitoring system was built to assess aircraft noise accurately and system- atically in the areas surrounding the three airports in the Kansai region.

The new system has achieved the same or better aircraft noise assessment accuracy and noise/flight log data matching accuracy as the conventional system, and has also realized labor-saving work re- lated to data scrutiny through the introduction of AI.

The introduction of a wake system has made it possible to monitor the position of an aircraft by only 15-second delay. By operating the system together with the noise system, it is possible to check the noise level at the monitoring point and the aircraft's position in conjunction with each other, ena- bling more detailed analysis of the noise generation situation.

The addition of a quasi-real-time web-based publication function of the noise level has contrib- uted to more open airport management by allowing anyone to view the instantaneous noise level values.

7. REFERENCES

1. UN. Population Division, The World's cities in 2018 , 2018. 2. E. Fujita, T. Higashioka, M. Sugiura and O. Kohashi, Evaluation method of military aircraft noise

using ai analysis of aircraft images. Proceedings of InterNoise 21 , pp. 854-862 Washington, D.C. 2021. 3. Y. Tadahira, K. Yamashita and S. Ohashi, Unattended monitoring technique for identifying the

aircraft noise: the method for correlating the observed aircraft noise with an airport mode and an aircraft operation. Proceedings of Inter-Noise 07 , pp. 398-406. Istanbul TURKEY 2007. 4. T. Higashioka, Y. Tadahira, M. Sugiura, E. Fujita and O. Kohashi, Runway determination using

two-point time difference method. Proceedings of InterNoise 22 , Glasgow Scotland 2022.

worm 2022