TY - JOUR
T1 - Optimal Motion Planning Method for Accurate Split-type Flying Vehicle Docking
AU - Wang, Weida
AU - Li, Boyuan
AU - Yang, Chao
AU - Qie, Tianqi
AU - Li, Ying
AU - Cheng, Jiankang
N1 - Publisher Copyright:
IEEE
PY - 2024
Y1 - 2024
N2 - The split-type flying vehicle is a new type of vehicle that can execute different tasks through different combinations of chassis, cabin, and aircraft. When the vehicle switches from ground driving to air flying, the chassis should be able to autonomously and accurately arrive at the target point to complete docking. To realize this switching, a motion planning method considering vehicle dynamics is proposed to achieve accurate docking. Firstly, a hybrid model of the vehicle is established, which is formulated by the kinematics model and the deviation model. Then, the neural network is used to train the deviation model. By setting the inputs as the known reference path, the deviation model is embedded in planning as linear constraints while retaining nonlinear characteristics. Finally, the planning considering the actual vehicle dynamics can be described in the form of quadratic programming so that the motion planning can be easily solved. The proposed method is verified with the split-type flying vehicle. Results show that compared with other methods, the proposed method reduces the mean lateral tracking deviation by 29.7% and the final lateral deviation by 66.7%, and reduces the mean heading angle tracking deviation by 50.0% and the final heading angle deviation by 81.3%.
AB - The split-type flying vehicle is a new type of vehicle that can execute different tasks through different combinations of chassis, cabin, and aircraft. When the vehicle switches from ground driving to air flying, the chassis should be able to autonomously and accurately arrive at the target point to complete docking. To realize this switching, a motion planning method considering vehicle dynamics is proposed to achieve accurate docking. Firstly, a hybrid model of the vehicle is established, which is formulated by the kinematics model and the deviation model. Then, the neural network is used to train the deviation model. By setting the inputs as the known reference path, the deviation model is embedded in planning as linear constraints while retaining nonlinear characteristics. Finally, the planning considering the actual vehicle dynamics can be described in the form of quadratic programming so that the motion planning can be easily solved. The proposed method is verified with the split-type flying vehicle. Results show that compared with other methods, the proposed method reduces the mean lateral tracking deviation by 29.7% and the final lateral deviation by 66.7%, and reduces the mean heading angle tracking deviation by 50.0% and the final heading angle deviation by 81.3%.
KW - Aircraft
KW - Atmospheric modeling
KW - Dynamics
KW - Kinematics
KW - Planning
KW - Transportation
KW - Vehicle dynamics
KW - motion planning
KW - optimization problem
KW - split-type flying vehicle
UR - http://www.scopus.com/inward/record.url?scp=85188015426&partnerID=8YFLogxK
U2 - 10.1109/TTE.2024.3374512
DO - 10.1109/TTE.2024.3374512
M3 - Article
AN - SCOPUS:85188015426
SN - 2332-7782
SP - 1
JO - IEEE Transactions on Transportation Electrification
JF - IEEE Transactions on Transportation Electrification
ER -