MATLAB and Parrot Anafi data

Description / Summary
This dataset supports the research paper proposing an Energy Consumption Reduction Model (ECRM) for autonomous multi-UAV path planning. The dataset comprises two components: MATLAB-based simulation data and real-world field experiment data collected using a Parrot Anafi UAV.
MATLAB Simulation Data: Three CSV files contain time-series flight logs generated from MATLAB simulations of ten UAVs operating over a 5000 m × 5000 m environment. Each file corresponds to one of the three improved path planning algorithms evaluated: Improved Soft Actor-Critic for Discrete Action Spaces Deep Q-Network (SAC-DQN), Improved Dueling Double Deep Q-Network (D3QN), and Improved Rapidly Exploring Random Tree-Connect (RRT-Connect). Logged variables include UAV position (x, y, z), velocity, energy level, stealth effectiveness, risk level, computational time, radar encounters, obstacle avoidances, and mission status, recorded at each simulation time step across all ten UAVs.
Parrot Anafi Field Experiment Data: Three Python scripts were executed on a laptop and transmitted wirelessly to a Parrot Anafi UAV via WiFi connection for outdoor real-world flight trials. The laptop acted as the ground control station, running the three improved path planning algorithms — SAC-DQN, D3QN, and RRT-Connect — and sending real-time flight commands to the Parrot Anafi through the Olympe SDK over a WiFi link. Flight data including GPS-derived path length, battery consumption, and mission completion time were collected directly from the Parrot Anafi via iPhone using the FreeFlight 6 application. These field results validate the practical deployability of the proposed ECRM in a real-world three-dimensional environment.