Improvement of traffic flow efficiency

Efficient air traffic controller (ATCo) task via reinforcement learning

ATCo tasks are highly specialized, requiring the safe and efficient control of a large number of aircraft. The goal is to create an artificial ATCo that is more efficient and safer than human controllers via reinforcement learning.

Fuel saving by climb thrust control

The aircraft usually climbs with maximum thrust, which is not necessarily optimal in terms of fuel consumption. This research aims at finding the optimal thrust control schedule and the way to realize it considering the current aircraft capability and the pilot's workload with the use of a flight simulator.

Optimization of aircraft departure/taxiing process

A large airport in the world faces a congestion problem of the aircraft on the ground. The fuel consumption during taxiing is not negligible, and appropriate runway scheduling will save fuel. There are several possible solutions, such as taxi route optimization, spot allocation, and pushback time control. These are formulated as an optimization problem, and the solving method is investigated.

Data-driven wind estimation method

The wind strongly affects the aircraft/vessel operation, and its estimation accuracy affects the operational efficiency. Apart from the numerical weather forecast, online wind estimation method is developed using machine learning. The impact of the wind estimation accuarcy to the operational efficiency is also under investigation.

Improvement of traffic safety

Speech recognition of air traffic control communication

ATCos and pilots still rely on voice communication, and this is not expected to change in the near future. The miscommunication between ATCos and pilots could lead to fatal accidents.To prevent such miscommunication, we propose a framework for detecting miscommunication by using voice recognition technology.

Mathematical modeling to determine safety separation

The safety is a first priority in aviation, and the large spacial-separation between aircraft improves the safety. However, the large spearation decreases the operational efficiency, so the minimum separation to guarantee the safety should be known. To do so, the mathematical modeling is used, and the collision probability is calculated under the given conditions. The result is reflected on the international rule on ICAO (International Civil Aviation Organization) and Japanese Aviation Safety Standard.

New operational safety standard for drones and urban air mobility (UAM)

Recently, drones and UAM start their operation, but the safety standard, especially the traffic rule, has not been well determined yet. Using the knowledge of manned-aviation field, the new operational rules will be developed.

Gaining knowledge from operational data

The traffic volume has been increasing in all traffic fields, and tons of data are available. These data include the various hints to improve the traffic efficiency and safety, and the way to gain it is the target of this research. Currently, operational data are obtained from the aircraft operatos in Japan, and the meaningful insights in terms of safety will be obtained.

Relaxation of obstacle construction rules

The height of obstalces such as buildings and trees is limited due to the flight safety near airports. These rules were developed long time ago, and have not been changed so far. A new obstacle assessment method is proposed, and the change of the obstacle construction rules is expected.

Human factors and others

Development of a pilot manual control model

The lack of aircraft/vessels pilots is a known issue, but it takes long time to gain the skill. Such skills are expected to be gained more effectively, which may solve the issue. We propose to develop a control model using the control data, and model parameters are used for the evaluation of the skill.

Development of an automatic procedure design tool

The aircraft must follow the flight procedure, which include route structures and flight restrictions. The flight procedure must be developed under the flight procedure design rules, and the current flight procedures are designed manually. The development of automatic design tool will contribute both workload of procedure designers and flight efficiency, and is under construction with a cooperation of Japan Civil Aviation Bureau (JCAB) and Electronic Navigation Research Institute (ENRI).