Abstract:It is impossible for takeaway delivery staff to plan the takeout delivery route balancing rationality and efficiency.To address this problem,an Improved Ant Colony Optimization (IACO) algorithm is proposed.The initial routes are obtained using the Ant Colony Optimization (ACO) algorithm and then optimized using Large Neighborhood Search (LNS) algorithm.The solution quality is improved by combining the ACO algorithm with the LNS algorithm.The proposed algorithm is verified by simulating the delivery routes for different number of takeout orders.Comparative analysis shows that the proposed IACO algorithm can increase the takeout delivery efficiency,according to the optimal distribution plan and the ideal value of the objective penalty function.The proposed strategy can enhance the intelligence and promote the long-term growth of the delivery system of Internet-connected takeout platforms.

Abstract:The rapid and accurate path planning for lane change,effective tracking of the desired path while maintaining vehicle stability during lane change are core technologies to ensure active safety for intelligent vehicles.Here,a path planning strategy based on double quintic polynomial and introduction of intermediate position is proposed for intelligent vehicles,with the purpose to perform smooth and safe lane-changing in real time.The active lane change scenario was analyzed to determine the initial and target positions of the lane change,then a lane-changing path planning strategy of double quintic polynomial was proposed based on critical collision position during lane change.And a co-simulation model was established to carry out active lane change simulation tests in different road conditions.The results show that the proposed lane-changing path planning strategy has more obvious lateral displacement due to the introduction of the intermediate position,which can avoid obstacles in front of the vehicle thus ensure the safe lane changing. Meanwhile, the maximum lateral acceleration of the vehicle at the intermediate position is no more than 2 m/s²,which ensures the vehicle stability during lane change process.Furthermore,the required longitudinal safety distance for lane changing is reduced by about 20 m on both dry and wet roads,which reduces the longitudinal collision risk.The results provide theoretical and practical basis for active lane change path planning of intelligent vehicles.

Abstract:To address the problem of long time-consuming and large memory consumption of A*algorithm in solving path trajectory,this paper proposes an improved A* algorithm based on adaptive step size.First,the priority order of the search direction was set according to the position relationship between the current point and the end point,with the purpose to reduce the redundant planning calculation on unreasonable directions. Second, the judgment condition for reaching the end point was modified to achieve path jumping during path planning.Then an adaptive step size strategy was proposed to improve the efficiency of A* algorithm in path planning.Finally,an eight-directional search approach was proposed to address the issues of large memory usage and possible memory overflow when facing large maps.Experimental results show that compared with original A* algorithm,the improved A* algorithm greatly improves the efficiency of path planning,and solves the problem of large memory usage.

Abstract:A three-stage rectifier containing three diodes is proposed in this paper.These diodes with different threshold voltages are located in three branches to achieve high conversion efficiency for different input power levels.By combining the three branches,the proposed rectifier is enabled to achieve high conversion efficiency over a wide dynamic input power range.A prototype of a 2.4 GHz rectifier is designed,simulated and measured.The efficiency of the three branches remains above 30% in the power range of -10-11 dBm,11-24 dBm and 24-35 dBm,respectively.The measured results show that the conversion efficiency maintains over 40% within the input power range of -6-25 dBm,with the maximum conversion efficiency reaching up to 71.2%.It indicates that the proposed rectifier can achieve low sensitivity to input power variation and high conversion efficiency over an ultra-wide input power range,which can be used in energy harvesting or dynamic wireless power transmission systems.

Abstract:Aiming at the distributed convex optimization problem in multi-agent systems,a zero-gradient-sum optimization algorithm based on adaptive event-triggered mechanism is proposed in this paper.The adaptive event-triggered condition is designed based on virtual clock,and the condition will not be triggered until the virtual clock of each agent meets the condition,which effectively reduces the update times of the controller as well as the communication burden of the system.Through constructing the Lyapunov function,it is proved that the states of all agents can converge asymptotically to the global optimal solution under the algorithm.In addition,the designed event-triggered condition makes the minimum inter-event time designable,which effectively excludes the Zeno behavior.Finally,simulation results verify the effectiveness of the algorithm.

Abstract:The finite-time guaranteed cost control problem is studied for impulsive stochastic Markov jump systems with time-varying delays.By selecting mode-dependent Lyapunov functionals,using techniques such as Linear Matrix Inequality (LMI) and average dwell time,we obtain some new sufficient conditions to ensure finite-time stability and the upper bound performance index of the system,hence design a guaranteed cost controller with state feedback.Finally,the LMI toolbox of Matlab is used for data simulation,and the corresponding mean-square trajectory plots are obtained.Numerical examples show that the obtained simulation results are consistent with the theoretical results,which verifies the effectiveness of the proposed approach.

Abstract:To ensure the output performance and suppress electromagnetic vibration of a Hollow Shafted Permanent Magnet Brushless DC Motor (HSPMBLDCM),a surrogate-assisted model optimization method based on nonlinear multiple regression is proposed.First,the optimal space-filling sampling was determined by the Audze-Eglajs (AE) criterion,and four main variables were selected by the Kernel Principal Components Analysis (KPCA) algorithm to construct the surrogate model.Next,the surrogate model was constructed using nonlinear multiple regression,and the R² values of the coefficients of determination were all greater than 0.9,which verified the accuracy of the surrogate model.Finally,the Robust Multiobjective Optimization Evolutionary Algorithms (RMOEA) were used to solve the surrogate model,and the optimal stator slot parameters were obtained.The results show that by optimizing the stator slot parameters,the average torque of the motor is reduced by 1.3%,which does not affect the output performance,and the maximum vibration acceleration of the motor at no load and rated load is reduced by 19% and 34.5%,respectively,which effectively reduces the electromagnetic vibration and proves the effectiveness and reliability of the optimization method.It provides an alternative means for motor vibration damping design optimization with practical significance.

Abstract:Existing intelligent optimization algorithms for reactive power optimization in distribution networks are perplexed by problems of slow convergence speed and easy falling into local optima.Here,a new approach based on the Multi-Objective Sand Cat Swarm Optimization (MOSCSO) is proposed to solve the reactive power optimization of wind and solar power storage and distribution networks.MOSCSO integrates the update and selection mechanism of external save sets in multi-objective algorithms,and has good global optimization ability.Meanwhile,the unique search and attack population update method of the sand cat swarm algorithm ensures its fast convergence speed and good optimization ability.An IEEE 33 bus system mathematical model with Energy Storage System (ESS) was established as the control variable,and then the MOSCSO was applied for simulation verification.The results demonstrate that the proposed approach can reduce grid loss and improve grid stability while balancing the wind and solar power generation systems,which verify the effectiveness and stability of MOSCSO in reactive power optimization.

Abstract:The existing remote sensing image object detection algorithms have been frustrated by large parameter quantities,slow detection speed and inability to deploy on mobile devices.Here,we propose a lightweight remote sensing image object detection algorithm without anchor frames.First,a DWS-Sandglass lightweight module is designed to reduce the model volume,and the activation function of the model is improved to ensure detection accuracy.Then,a parameter free attention module SimAM is introduced to enable the network to focus on more important feature information.Finally,the redundant channels of the anchor frame free algorithm are pruned to reduce the number of model parameters,and the accuracy is improved by fine tuning.The experimental results on HRSC2016 dataset show that compared with current mainstream detection algorithms free of anchor frame,the proposed algorithm has faster detection speed and smaller model size,making it more suitable for deployment on mobile devices with comparable detection accuracy.

Abstract:The large number of people and vehicles gathered in a short period of time around large-scale events will lead to a differentiated traffic flow.Here,an interpretable machine learning model integrating XGBoost algorithm and partial dependence plots is proposed to capture the nonlinear effects and synergistic influences of large-scale events and their characteristics on the operation of nearby road network,and an empirical study has been conducted in Beijing.The heterogeneity of single factors shows that the distance of road section away from event venue and the event scale have great impact on nearby traffic flow,with relative importance of 27.1% and 25.4%,respectively;time before start and after end of the event has obvious nonlinear characteristics,and the road sections within 3 km from the venue will be significantly affected within 30-60 minutes before the event and 30 minutes after the event.The synergistic effect of two-dimensional factors shows that,if an event attracted more than 30,000 people,holidays and adverse weather have a negative impact on the nearby traffic flow;in rain or haze weather,the road section within 2.5 km from the venue will be affected within 60 minutes before the event and 40 minutes after the event.The findings can provide quantitative data support for identifying the causes of road congestion and formulating reasonable and effective road network control strategies during large events.

Abstract:The high randomness and volatility of photovoltaic (PV) power makes it difficult for single prediction models to accurately analyze the fluctuation patterns in historical data,resulting in low prediction accuracy.To address this issue,a combined model for short-term PV power prediction was proposed,which incorporated Convolutional Neural Network,Gated Recurrent Unit (CNN-GRU) and an Improved Sparrow Search Algorithm (ISSA) for optimizing the eXtreme Gradient Boosting (XGBoost) model.First,the historical data were normalized after outlier removal,and feature selection was carried out via Principal Component Analysis (PCA) so as to better identify the key factors affecting photovoltaic power.Then,the CNN and GRU networks were used to extract the spatial and temporal features of the data,respectively.To address the difficulty in manually configuring parameters and high randomness of the XGBoost model,ISSA was used to optimize the hyperparameters of the model.Finally,the reciprocal error method was used to reduce the error of the results predicted by the two methods (CNN-GRU and ISSA-XGBoost) while the weights were updated to obtain new predicted values to complete the prediction of photovoltaic power.The experimental results show that the proposed CNN-GRU-ISSA-XGBoost model has strong adaptability and high accuracy.

Abstract:The prediction of reservoir water level provides important decision support for reservoir operation,flood control and water resources operation and management.Accurate and reliable prediction plays an important role in the optimal management of water resources.Aiming at the nonlinearity,instability and complex temporal and spatial characteristics of reservoir water level data,a hybrid reservoir water level prediction model integrating adaptive Variational Mode Decomposition (VMD),Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) is proposed.Among them,VMD eliminates noise by decomposing the water level sequence,CNN is used to effectively extract the local features of water level data,and GRU is used to extract the deep time features of water level data.Taking the daily water level prediction of Shenwo reservoir as an example,the proposed model outperforms current deep learning models in accuracy.In terms of computing efficiency,the operation efficiency of GRU selected in this approach is significantly improved compared with Long Short-Term Memory network (LSTM).Therefore,the proposed model has high accuracy and high operation efficiency,and is more suitable for the real-time operation of reservoir water level.

Abstract:Accurate wind speed prediction is the key to large-scale application of wind energy in power system,but the randomness and volatility of wind speed sequence make it difficult to predict.Herein,strategies of Logistic chaotic mapping,adaptive parameter adjustment,and the introduction of mutation are used to improve the Carnivorous Plant Algorithm (CPA),and a short-term wind speed prediction model based on error correction and VMD-ICPA-LSSVM is proposed.First,meteorological factors are used as inputs for Least Squares Support Vector Machine (LSSVM) to predict wind speed and obtain an error sequence.Then,K-L divergence is used to adaptively determine the parameters of Variational Mode Decomposition (VMD) and decompose the error sequence.Then the Improved Carnivorous Plant Algorithm (ICPA) is combined to optimize the adjustable parameters of LSSVM to predict the decomposed subsequences.The prediction results of each subsequence are stacked and error correction is performed on the original prediction sequence to obtain the final wind speed prediction values.The experimental results show that the proposed model has excellent prediction accuracy and generalization performance.

Abstract:This paper uses the dual-frequency GNSS observation data of the France's BRST port and the UK's Severn Bridge monitoring system to perform GPS-IR water level inversion in static and high dynamic environments,respectively,to explore the feasibility and accuracy of traditional GNSS monitoring system for water level inversion.The results show that the inversion accuracy of the L1 band is higher than that of the L2 band;in the static scene,the correlation coefficient between the GPS-IR water level inversion results and the tide gauge data is greater than 0.98;in the high dynamic scene,the retrieval accuracy of F001 station is slightly lower.Then the Empirical Mode Decomposition (EMD) is used to improve the accuracy of GPS-IR water level retrieval results in complex bridge environment,which can reduce the root mean square error by about 50%.The EMD-improved GPS-IR water level retrieval approach improves the applicability of GPS-IR technology in different water environments,and has a good application prospect in water level monitoring.

Abstract:Multipath frequencies need to be estimated for tide level inversion from Global Navigation Satellite System Reflectometry (GNSS-R) signals.However,conventional inversion methods only estimate principal frequency,which results in low data utilization and insufficient temporal resolution of the inversion results.Here,we use the Savitzky-Golay (SG) smoothing filtering to optimize the GNSS-R tide level inversion.First,the Lomb-Scargle Periodogram (LSP) is used to extract the first four frequencies (f₁-f₄) of signal power,which are then inverted for their corresponding tide level values.Then the SG smoothing filtering is used to extract the best inversion results.Finally,the 30-day data from BRST and MAYG stations in France are used to verify the effectiveness of the approach.The results show that,compared with LSP method,the proposed approach increases the number of daily average inversion values by 34.3% and 19.6%,and reduces the maximum time interval of inversion values by 43.2% and 29.4%,for BRST station and MAYG station,respectively;compared with window LSP (WINLSP) method,the proposed approach increases the number of daily average inversion values of BRST station and MAYG station by 24.2% and 45.9%,decreases the maximum time interval of inversion values by 25.4% and 28.6%,respectively,and reduces the RMSE by 7 cm for both stations.It can be concluded that this method increases the number of inversion results,raises the data utilization and improves the temporal resolution of tide level inversion besides adequate accuracy.

Abstract:The timely and accurate monitoring of water quality via remote sensing is of great significance to water environment protection.However,the relationship between water quality parameters and surface reflectance is not a simple linear one.BP neural network and Support Vector Machine (SVM) have been widely used in water quality inversion for their nonlinear simulation characteristics,yet traditional BP neural network is perplexed by slow convergence and being easy to fall into local optimum,while SVM is greatly affected by penalty coefficient and kernel function parameter.Here,a coupled model using Sparrow Search Algorithm (SSA) to optimize BP neural network and SVM is proposed to retrieve water quality parameters of conductivity and turbidity in Yunlong Lake from Sentinel-2 images.SSA is used to optimize the parameters of BP neural network and SVM,the model weight is calculated based on verification set MAE,and the final inversion results are obtained after the weighted calculation of output layer of SSA-BP and SSA-SVM model test group.And comparisons are carried out between the proposed SSA-BP-SVM model and BPNN,SVM,SSA-BP,and SSA-SVM models.The results show that,the sensitive bands of Sentinel-2 image to conductivity and turbidity are visible light and shortwave infrared;the proposed model of SSA-BP-SVM is more precise with the R² of the inverted conductivity and turbidity being 0.92 and 0.89,respectively;the Yunlong Lake is a typical urban water body with conductivity being greatly affected by the drainage from upstream water treatment plant and turbidity being greatly affected by particulate pollutants from social production activities.The proposed SSA-BP-SVM model has good application potential in water quality inversion from Sentinel-2 image,which can provide technical support for water quality monitoring and protection of Yunlong Lake.