A Multi-Timescale Method for State of Charge Estimation for Lithium-Ion Batteries in Electric UAVs Based on Battery Model and Data-Driven Fusion DOI Creative Commons

Xiao Cao,

Liu Li

Drones, Journal Year: 2025, Volume and Issue: 9(4), P. 247 - 247

Published: March 26, 2025

This study focuses on the critical problem of precise state charge (SOC) estimation for electric unmanned aerial vehicle (UAV) battery systems, addressing a fundamental challenge in enhancing energy management reliability and flight safety. The current data-driven methods require big data high computational complexity, model-based need high-quality model parameters. To address these challenges, multi-timescale fusion method that integrates technologies SOC lithium-ion batteries has been developed. Firstly, under condition no or insufficient data, an adaptive extended Kalman filtering with multi-innovation algorithm (MI-AEKF) is introduced to estimate based Thévenin fast timescale. Then, hybrid bidirectional time convolutional network (BiTCN), gated recurrent unit (BiGRU), attention mechanism (BiTCN-BiGRU-Attention) deep learning using parameters used correct error relatively slow performance proposed validated various dynamic profiles battery. results show maximum (ME), mean absolute (MAE) root square (RMSE) zero data-driving, sufficient data-driving conditions are below 2.3%, 1.3% 1.5%, 0.9%, 0.4% 0.4%, 0.6%, 0.3% 0.3%, respectively, which showcases robustness remarkable generalization method. These findings significantly advance strategies Li-ion systems UAVs, thereby improving operational efficiency extending endurance.

Language: Английский

Enhancing Sustainable Last-Mile Delivery: The Impact of Electric Vehicles and AI Optimization on Urban Logistics DOI Creative Commons
João C. Ferreira,

Marco Esperança

World Electric Vehicle Journal, Journal Year: 2025, Volume and Issue: 16(5), P. 242 - 242

Published: April 22, 2025

The rapid growth of e-commerce has intensified the need for efficient and sustainable last-mile delivery solutions in urban environments. This paper explores integration electric vehicles (EVs) artificial intelligence (AI) into a combined framework to enhance environmental, operational, economic performance logistics. Through comprehensive literature review, we examine current trends, technological developments, implementation challenges at intersection smart mobility, green logistics, digital transformation. We propose an operational that leverages AI route optimization, fleet coordination, energy management EV-based networks. is validated through real-world case study conducted Lisbon, Portugal, where logistics provider implemented city consolidation center model supported by AI-driven optimization tools. Using key indicators—including time, consumption, utilization, customer satisfaction, CO₂ emissions—we measure pre- post-AI deployment impacts. results demonstrate significant improvements across all metrics, including 15–20% reduction 10–25% gain efficiency, up 40% decrease emissions. findings confirm synergy between EVs provides robust scalable achieving supporting broader mobility climate objectives.

Language: Английский

Citations

0
A Multi-Timescale Method for State of Charge Estimation for Lithium-Ion Batteries in Electric UAVs Based on Battery Model and Data-Driven Fusion DOI Creative Commons

Xiao Cao,

Liu Li

Drones, Journal Year: 2025, Volume and Issue: 9(4), P. 247 - 247

Published: March 26, 2025

This study focuses on the critical problem of precise state charge (SOC) estimation for electric unmanned aerial vehicle (UAV) battery systems, addressing a fundamental challenge in enhancing energy management reliability and flight safety. The current data-driven methods require big data high computational complexity, model-based need high-quality model parameters. To address these challenges, multi-timescale fusion method that integrates technologies SOC lithium-ion batteries has been developed. Firstly, under condition no or insufficient data, an adaptive extended Kalman filtering with multi-innovation algorithm (MI-AEKF) is introduced to estimate based Thévenin fast timescale. Then, hybrid bidirectional time convolutional network (BiTCN), gated recurrent unit (BiGRU), attention mechanism (BiTCN-BiGRU-Attention) deep learning using parameters used correct error relatively slow performance proposed validated various dynamic profiles battery. results show maximum (ME), mean absolute (MAE) root square (RMSE) zero data-driving, sufficient data-driving conditions are below 2.3%, 1.3% 1.5%, 0.9%, 0.4% 0.4%, 0.6%, 0.3% 0.3%, respectively, which showcases robustness remarkable generalization method. These findings significantly advance strategies Li-ion systems UAVs, thereby improving operational efficiency extending endurance.

Language: Английский

Citations

0