Search

Increasing use of Unmanned Aerial Vehicle (UAV) in urban environments poses to an increased risk of fallen UAVs impacting people and vehicles on the ground, as well as colliding with manned aircraft in the vicinity of airports. Risk management of UAV flights for safe operations is essential. We proposed a comprehensive risk assessment model for UAV operation in urban environments. Three risk categories (people, vehicles, and manned aircraft) were considered and each risk cost was quantified using collision probability. We adjusted the risk costs in various magnitudes to a same scale and conducted a sensitivity analysis to determine the optimal coefficients of the three risk cost models. We then computed the total risk and generated a risk cost map for path planning. Modified path planning algorithms were used to produce a cost-effective path, and we compared their performances in terms of total risk cost and computational time. Lastly, we performed simulations to validate the feasibility and effectiveness of our proposed risk assessment model. The results show that the risk-cost-based path planning method can generate safer path for UAV operations than the traditional shortest-distance-based method. Our proposed model can be extended to complex urban environments by including more relevant parameters and data.

Numerous studies employ multi-scale decomposition to improve the prediction performance of neural networks, but the grounds for selecting the decomposition algorithm are not explained, and the effects of decomposition algorithms on other performance of neural networks are also lacking further study. This paper studies the influence of commonly used multi-scale decomposition algorithms including EMD (Empirical Mode Decomposition), EEMD(Ensemble Empirical Mode Decomposition), CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise), VMD (Variational Mode Decomposition), WD (Wavelet Decomposition), and WPD (Wavelet Packet Decomposition) on the performance of Neural Networks. Decomposition algorithms are adopted to decompose traffic flow data into component signals, and then K-means is used to cluster component signals into volatility components, periodic components, and residual components. A Bi-directional LSTM (BiLSTM) neural network is adopted as the standard model for training and forecasting. Finally, three metrics, including prediction performance, robustness, and generalization performance are proposed to evaluate the influence of the multi-scale decomposition algorithm for neural networks comprehensively. By comparing the evaluation results of different hybrid models, this study provides some useful suggestions on proper multi-scale decomposition algorithm selection in short-time traffic flow prediction.