Drones and computer vision Team

Principal Investigator:Professor Li-Chun Wang

—

Summary

Our project proposes an “Artificial Intelligence (AI) Drone-Cruiser” base station for helping 5G mobile communication systems rapidly recover the network after the disaster and handle the instant traffic of the flash crowd. The drone-cruiser base station can overcome the communications problem for three types of flash crowds, such as in stadiums, parades, and large plaza so that an appropriate number of aerial base stations can be precisely installed to meet the huge and dynamical traffic requirements. Artificial intelligence can solve such problems by analyzing the collected data, and then adjust the system parameters to achieve the goals of self-configuration, self-optimization, and self-healing under the framework of Self-Organizing Network (SON). With the help of AI technologies, the 5G network can become even more intelligent. This project is to provide a novel service, On-Demand Aerial Base Station as a Service, including the following three technical challenges: (1) rapid air-to-ground 3D wireless channel learning technology; (2) optimal 3d placement for aerial base stations; (3) big data analysis and AI technologies for the automatic UAV management; and (4) innovational design for the long-time hovering drone-cruiser. It is hoped that the outcome of this project can help open up another emerging opportunity for Flying Base Station (Flying Access Point) in the post-5G era for Taiwan’s information and communications industry, following the success story of WiFi in Taiwan.

Keywords

Aerial Base Station, Drone-Cruiser, Artificial Intelligence, Self-Organizing Network, 3D Placement, Flying Access Point.

Innovations

• We analyze the characteristics of the 3D air-to-ground wireless channel and design a new learning mechanism for UAV-BS to provide stable communication services.
• We propose a hybrid training framework for training an effective convolutional neural network to estimate the crowds and density of users in the sensed aerial photographs.
• With the consideration of the above estimation information, we design a fast 3d placement of aerial base stations for guaranteeing the allocated data rate of each served user.
• We integrate the latest LTE, WiGig, and mmWave technologies on the UAV for guaranteeing the backhaul connections while providing network services.
• We propose a deep Q-learning network for the autonomous landing and thus guarantee the safety of the flight.
• We design and develop a new type of UAV, Drone-Cruiser, to provide base station service with dynamic mobility for a long hovering time.

Benefits

• We collect the real 3d wireless channel data and construction a 3D+radio dataset/map to help the prediction of communication link quality.
• We develop a simulation platform to recommend the appropriate 3d placement configuration for improving the performance of UAV-assisted communications systems.
• We propose an on-demand UAV base station placement for guaranteeing the allocated data rate of each user. The proposed method also saves the power consumption of a UAV for communications by more than 28.9%.
• The new type of UAV, Drone-Cruiser, can be one of an important product for the UAV-assisted communications industry.