Aiming to address the Unmanned Aerial Vehicle(UAV) formation collision avoidance problem in Three-Dimensional(3-D) low-altitude environments where dense various obstacles exist, a fluid-based path planning framework n...Aiming to address the Unmanned Aerial Vehicle(UAV) formation collision avoidance problem in Three-Dimensional(3-D) low-altitude environments where dense various obstacles exist, a fluid-based path planning framework named the Formation Interfered Fluid Dynamical System(FIFDS) with Moderate Evasive Maneuver Strategy(MEMS) is proposed in this study.First, the UAV formation collision avoidance problem including quantifiable performance indexes is formulated. Second, inspired by the phenomenon of fluids continuously flowing while bypassing objects, the FIFDS for multiple UAVs is presented, which contains a Parallel Streamline Tracking(PST) method for formation keeping and the traditional IFDS for collision avoidance. Third, to rationally balance flight safety and collision avoidance cost, MEMS is proposed to generate moderate evasive maneuvers that match up with collision risks. Comprehensively containing the time and distance safety information, the 3-D dynamic collision regions are modeled for collision prediction. Then, the moderate evasive maneuver principle is refined, which provides criterions of the maneuver amplitude and direction. On this basis, an analytical parameter mapping mechanism is designed to online optimize IFDS parameters. Finally, the performance of the proposed method is validated by comparative simulation results and real flight experiments using fixed-wing UAVs.展开更多
The rapid advancement of three-dimensional printed concrete(3DPC)requires intelligent and interpretable frameworks to optimize mixture design for strength,printability,and sustainability.While machine learning(ML)mode...The rapid advancement of three-dimensional printed concrete(3DPC)requires intelligent and interpretable frameworks to optimize mixture design for strength,printability,and sustainability.While machine learning(ML)models have improved predictive accuracy,their limited transparency has hindered their widespread adoption in materials engineering.To overcome this barrier,this study introduces a Random Forests ensemble learning model integrated with SHapley Additive exPlanations(SHAP)and Partial Dependence Plots(PDPs)to model and explain the compressive strength behavior of 3DPC mixtures.Unlike conventional“black-box”models,SHAP quantifies each variable’s contribution to predictions based on cooperative game theory,which enables causal interpretability,whereas PDP visualizes nonlinear and interactive effects between features that offer practical mix design insights.A systematically optimized random forest model achieved strong generalization(R2=0.978 for training,0.834 for validation,and 0.868 for testing).The analysis identified curing age,Portland cement,silica fume,and the water-tobinder ratio as dominant predictors,with curing age exerting the highest positive influence on strength development.The integrated SHAP-PDP framework revealed synergistic interactions among binder constituents and curing parameters,which established transparent,data-driven guidelines for performance optimization.Theoretically,the study advances explainable artificial intelligence in cementitious material science by linking microstructural mechanisms to model-based reasoning,thereby enhancing both the interpretability and applicability of ML-driven mix design for next-generation 3DPC systems.展开更多
Nighttime navigation faces challenges from limited data and interference,especially when satellite signals are unavailable.Leveraging lunar polarized light,polarization navigation offers a promising solution for night...Nighttime navigation faces challenges from limited data and interference,especially when satellite signals are unavailable.Leveraging lunar polarized light,polarization navigation offers a promising solution for nighttime autonomous navigation.Current algorithms,however,are limited by the requirement for known horizontal attitudes,restricting applications.This study introduces an autonomous 3-D attitude determination method to overcome this limitation.Our approach utilizes the Angle of Polarization(AOP)at night to extract neutral points from the AOP pattern.This allows for the calculation of polarization meridian plane information for attitude determination.Subsequently,we present an optimized Polarization TRIAD(Pol-TRIAD)algorithm to acquire the 3-D attitude.The proposed method outperforms the existing approaches in outdoor experiments by achieving lower Root Mean Square Error(RMSE).For one baseline attitude,it improves pitch by 31.7%,roll by 21.7%,and yaw by 2.6%,while for the attitude with a larger tilt angle,the improvements are 64.4%,30.4%,and 9.1%,respectively.展开更多
基金supported in part by the National Natural Science Foundations of China(Nos.61175084,61673042 and 62203046)the China Postdoctoral Science Foundation(No.2022M713006).
文摘Aiming to address the Unmanned Aerial Vehicle(UAV) formation collision avoidance problem in Three-Dimensional(3-D) low-altitude environments where dense various obstacles exist, a fluid-based path planning framework named the Formation Interfered Fluid Dynamical System(FIFDS) with Moderate Evasive Maneuver Strategy(MEMS) is proposed in this study.First, the UAV formation collision avoidance problem including quantifiable performance indexes is formulated. Second, inspired by the phenomenon of fluids continuously flowing while bypassing objects, the FIFDS for multiple UAVs is presented, which contains a Parallel Streamline Tracking(PST) method for formation keeping and the traditional IFDS for collision avoidance. Third, to rationally balance flight safety and collision avoidance cost, MEMS is proposed to generate moderate evasive maneuvers that match up with collision risks. Comprehensively containing the time and distance safety information, the 3-D dynamic collision regions are modeled for collision prediction. Then, the moderate evasive maneuver principle is refined, which provides criterions of the maneuver amplitude and direction. On this basis, an analytical parameter mapping mechanism is designed to online optimize IFDS parameters. Finally, the performance of the proposed method is validated by comparative simulation results and real flight experiments using fixed-wing UAVs.
基金supported by the Ongoing Research Funding Program(Grant No.ORFFT-2025-025-4)at King Saud University,Riyadh,Saudi Arabia.The grant was awarded to Yassir M.Abbas。
文摘The rapid advancement of three-dimensional printed concrete(3DPC)requires intelligent and interpretable frameworks to optimize mixture design for strength,printability,and sustainability.While machine learning(ML)models have improved predictive accuracy,their limited transparency has hindered their widespread adoption in materials engineering.To overcome this barrier,this study introduces a Random Forests ensemble learning model integrated with SHapley Additive exPlanations(SHAP)and Partial Dependence Plots(PDPs)to model and explain the compressive strength behavior of 3DPC mixtures.Unlike conventional“black-box”models,SHAP quantifies each variable’s contribution to predictions based on cooperative game theory,which enables causal interpretability,whereas PDP visualizes nonlinear and interactive effects between features that offer practical mix design insights.A systematically optimized random forest model achieved strong generalization(R2=0.978 for training,0.834 for validation,and 0.868 for testing).The analysis identified curing age,Portland cement,silica fume,and the water-tobinder ratio as dominant predictors,with curing age exerting the highest positive influence on strength development.The integrated SHAP-PDP framework revealed synergistic interactions among binder constituents and curing parameters,which established transparent,data-driven guidelines for performance optimization.Theoretically,the study advances explainable artificial intelligence in cementitious material science by linking microstructural mechanisms to model-based reasoning,thereby enhancing both the interpretability and applicability of ML-driven mix design for next-generation 3DPC systems.
基金supported in part by the National Key Research and Development Program of China(Nos.2020YFA0711200,2022YFB4701301)in part by the Defense Industrial Technology Development Program,China(No.JCKY2021601B016)+1 种基金in part by the Fundamental Research Funds for the Central Universities,China(No.YWF-23-JC-07)in part by the National Natural Science Foundation of China(No.62425302)。
文摘Nighttime navigation faces challenges from limited data and interference,especially when satellite signals are unavailable.Leveraging lunar polarized light,polarization navigation offers a promising solution for nighttime autonomous navigation.Current algorithms,however,are limited by the requirement for known horizontal attitudes,restricting applications.This study introduces an autonomous 3-D attitude determination method to overcome this limitation.Our approach utilizes the Angle of Polarization(AOP)at night to extract neutral points from the AOP pattern.This allows for the calculation of polarization meridian plane information for attitude determination.Subsequently,we present an optimized Polarization TRIAD(Pol-TRIAD)algorithm to acquire the 3-D attitude.The proposed method outperforms the existing approaches in outdoor experiments by achieving lower Root Mean Square Error(RMSE).For one baseline attitude,it improves pitch by 31.7%,roll by 21.7%,and yaw by 2.6%,while for the attitude with a larger tilt angle,the improvements are 64.4%,30.4%,and 9.1%,respectively.
