1. Design of AI-based 3.84 kW Battery Package Using Backpropagation Artificial Neural Network Algorithm for Cargo Drones.
- Author
-
Hartono, Rodi, Sang Min Oh, Sung Won Lim, Kalend, Tshibang Patrick A., Doliev, Jasurbek, Jun Hyuk Lee, and Kyoo Jae Shin
- Subjects
ARTIFICIAL neural networks ,ARTIFICIAL intelligence ,DRONE aircraft delivery ,BATTERY management systems ,FREIGHT & freightage ,ALGORITHMS - Abstract
Despite limitations in payload and range, cargo drones have promising applications in emergency logistics and remote delivery. In this study, we tackle these challenges by developing a high-capacity 3.84 kW battery specifically designed for a 50-kg-payload cargo drone operating in demanding terrains. Focusing on the transport of emergency goods, we investigate key drone design aspects and details of the battery pack development, including cell selection, internal configuration, and critical circuits for cell balancing, charging/discharging, and advanced battery management. A key innovation is the integration of a backpropagation artificial neural network (BPANN) algorithm to predict the depth of discharge (DoD) and the state of charge (SoC). Research results show that BPANN offers highly accurate predictions, with error percentages as low as 0.12% for DoD and 0.02% for SoC, ensuring optimized and safe battery operation. Comprehensive field testing is carried out to evaluate the effectiveness of the proposed cell balancing strategy, robust battery management system (BMS), and BPANN implementation. We investigate the drone's performance in terms of DoD, SoC, and overall field operation with the designed battery pack and demonstrate its feasibility and potential for realworld applications. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF