Search

Air route crossing waypoint optimization is one of the effective ways to improve airspace utilization, capacity and resilience in dealing with air traffic congestion and delay. However, research is lacking on the optimization of multiple Crossing Waypoints (CWPs) in the fragmented airspace separated by Prohibited, Restricted and Dangerous areas (PRDs). To tackle this issue, this paper proposes an Artificial Potential Field (APF) model considering attractive forces produced by the optimal routes and repulsive forces generated by obstacles. An optimization framework based on the APF model is proposed to optimize the different airspace topologies varying the number of CWPs, air route segments, and PRDs. Based on the framework, an adaptive method is developed to dynamically control the optimization process in minimizing the total air route cost. The proposed model is applied to busy controlled airspace. And the obtained results show that after optimization the safety-related indicators: conflict number and controller workload reduced by 7.75% and 6.51% respectively. As for the cost-effectiveness indicators: total route length, total air route cost and non-linear coefficient, declined by 1.74%, 3.13% and 1.70% respectively. While the predictability indicator, total flight delay, saw a notable reduction by 7.96%. The proposed framework and methodology can also provide an insight in the understanding of the optimization to other network systems. Civil Aviation Authority of Singapore (CAAS) Published version This research was supported by the Civil Aviation Authority of Singapore and the Nanyang Technological University, Singapore under their collaboration in the Air Traffic Management Research Institute. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of the Civil Aviation Authority of Singapore.

There have been many applications of drones in urban environments, such as delivery, rescue, and surveillance. In a persistent surveillance task, the drones sometimes cannot complete it independently when some regions are required to be covered on the ground. For this purpose, unmanned aerial vehicles and unmanned ground vehicles (UAVs & UGVs) system is introduced to perform such a task in this article, and the goal is to generate the circular paths for the drones and the UGVs, respectively, to minimize their travel time of realizing a complete coverage. First, the cooperative path planning problem of UAVs & UGVs is formulated into a large-scale 0–1 optimization problem, in which the on–off states of the discrete points are to be optimized. Second, a hybrid algorithm integrating the estimation of distribution algorithm (EDA) and the genetic algorithm (GA) algorithm is proposed to solve the problem. The advantages of EDA and GA in the global and local search are fully taken considering the demands in different phases of the iterative process. A simple sweep-based approach is employed to determine the optimal sequence of passing the open points. Then, an online local adjustment strategy is also applied to address the changes of the requirements on covering the ground area. Simulation results demonstrate that the UAVs & UGVs system can enhance the efficiency of the task. The hybrid EDA-GA algorithm can greatly improve the performance of EDA and GA in terms of the quality and the stability of solutions. The online adjustment strategy is effective to maintain a complete coverage while minimizing the impact on the circular paths.

Different from the usual surveillance task in which the goal is to achieve complete coverage of the specified area, the cooperative path planning problem of drones for persistent surveillance task is studied in this paper considering multiple constraints of the covered area. The goal is to maximize the combinational coverage area of drones while giving preference to the area that hasn’t been visited beyond a certain time interval. The influence of shooting resolution and blocking of buildings are considered, and the state information of each grid is defined to record the visit information of the ground area. Considering the characteristic of the established model, the multi-constrained cooperative path planning (MCCPP) algorithm is developed. The grids which have not been visited for a long time are received special attentions, and the drone is led to reducing the flight height to cover the gird which has a special requirement on the shooting resolution. The cooperation mechanism among drones is also set to ensure that all the drones can determine the next path point synchronously. An emergency path planning algorithm with the continuous checking strategy is designed for a drone to fly to the specified area and finish a complete coverage of it.

We propose a causal framework to explain the catastrophic forgetting in Class-Incremental Learning (CIL) and then derive a novel distillation method that is orthogonal to the existing anti-forgetting techniques, such as data replay and feature/label distillation. We first 1) place CIL into the framework, 2) answer why the forgetting happens: the causal effect of the old data is lost in new training, and then 3) explain how the existing techniques mitigate it: they bring the causal effect back. Based on the framework, we find that although the feature/label distillation is storage-efficient, its causal effect is not coherent with the end-to-end feature learning merit, which is however preserved by data replay. To this end, we propose to distill the Colliding Effect between the old and the new data, which is fundamentally equivalent to the causal effect of data replay, but without any cost of replay storage. Thanks to the causal effect analysis, we can further capture the Incremental Momentum Effect of the data stream, removing which can help to retain the old effect overwhelmed by the new data effect, and thus alleviate the forgetting of the old class in testing. Extensive experiments on three CIL benchmarks: CIFAR-100, ImageNet-Sub&Full, show that the proposed causal effect distillation can improve various state-of-the-art CIL methods by a large margin (0.72%--9.06%).

Traffic flow forecasting is a typical problem of spatial-temporal data forecasting. Because traffic data show complex patterns and traffic graph is difficult to process, it is very challenging. And most existing traffic flow prediction methods lack the ability to model the dynamic spatial-temporal features. These methods cannot yield satisfactory prediction results. In the paper, we propose Attention Based Multi-Unit Spatial-Temporal Network (AMU-STN) solve the traffic flow forecasting problem. The model mainly consisting three units, specifically, spatial-temporal attention unit to effectively capture the dynamic spatial-temporal correlations, spatial-temporal feature extraction unit to extract long-range feature, and prediction unit to predict the feature status. For dealing with long-term predictions, we stack multiple layers of attention units and feature extraction units. We use a residual connection to avoid gradient vanish due to deepening of network layers. Experimental results on two real-world datasets from Caltrans Performance Measurement System (PeMS) prove ours model outperforms the state-of-the-art baselines. We also analyze the weight of the attention matrix; the result shows the effectiveness of the attention mechanism and reflects the interpretability of ours model.

Real-world visual recognition requires handling the extreme sample imbalance in large-scale long-tailed data. We propose a "divide&conquer" strategy for the challenging LVIS task: divide the whole data into balanced parts and then apply incremental learning to conquer each one. This derives a novel learning paradigm: class-incremental few-shot learning, which is especially effective for the challenge evolving over time: 1) the class imbalance among the old-class knowledge review and 2) the few-shot data in new-class learning. We call our approach Learning to Segment the Tail (LST). In particular, we design an instance-level balanced replay scheme, which is a memory-efficient approximation to balance the instance-level samples from the old-class images. We also propose to use a meta-module for new-class learning, where the module parameters are shared across incremental phases, gaining the learning-to-learn knowledge incrementally, from the data-rich head to the data-poor tail. We empirically show that: at the expense of a little sacrifice of head-class forgetting, we can gain a significant 8.3% AP improvement for the tail classes with less than 10 instances, achieving an overall 2.0% AP boost for the whole 1,230 classes.

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. Nanyang Technological University Published version This work was supported in part by the National Key R&D Program of China under Grant 2016YFB0502400, and in part by the UAS Programme in the Air Traffic Management Research Institute (ATMRI), Nanyang Technological University (NTU), Singapore.

The demand on the public transportation in many cities has been increasing with the diversity of commuters’ activities that require the regular travelling across the urban and suburb areas. However, the public transport systems still suffer from the traffic congestion and detour, which will then reduce the efficiency. Accordingly, extensive efforts have been made for decades to optimize the operations of metro ground traffic with trains, subways, and buses (2D public transportation). On the other hand, unmanned aerial vehicles (UAVs) have been widely applied in the event-triggered tasks, such as delivery, rescue and surveillance in urban environments, and they have a great potential to relieve the pressure in 2D ground public transportation. Inspired from metro systems, AirMetro is a new concept proposed for future 3D public air transportation service, which carries the passengers commuting from a point to the destination, routinely and periodically. In this paper, the flight scheduling problem for AirMetro is studied based on the flight route of UAVs. The low-altitude airspace in urban environments is divided into several layers by altitude, and AirMetro is conducted in the allocated airspace as the public lanes. First, a pipeline-based route planning algorithm is proposed for the gridded urban environments to further reduce the length of a flight route generated by A* algorithm. The UAV is not required to pass through the centre of a cube in the pipeline-based algorithm when compared with the A* algorithm. This can then potentially result in a shorter flight route. As for resolving the conflict among UAVs, a minimum influence-based approach is developed to delay the flight and reduce the influence of adjustment on the follow-up flights of the UAV. By further introducing a modified simulated annealing (MSA) algorithm, the delay of flight can be further reduced by optimizing the UAV sequence conducting the conflict resolution. Results of the case studies demonstrate that the trajectory-based flight scheduling method can improve the flight efficiency by ensuring the flight safety of UAVs and reducing the average delay of UAVs as well in AirMetro. Ministry of Education (MOE) Nanyang Technological University This research work is supported by the Chongqing Research Program of Basic Research and Frontier Technology (grant number of cstc2020jcyj-msxmX0602) and the Fundamental Research Funds for the Central Universities (grant number of 2020 CDJ-LHZZ-066). Thanks are also due to China Scholarship Council (project reference number 201906055030) for the financial support of the research attachment of the first author to the ATMRI in the NTU, Singapore. This collaborative research, findings presented, and interactions among teams are also relevant to the Ministry of Education (MOE, Singapore) Tier-1 project research grant (Project ID: 2018-T1-002-124) and the UAS Program on “Urban Aerial Transport Traffic Management and Systems” in the ATMRI, NTU, Singapore. The current work also provides the basis and idea for the recently approved National Natural Science Foundation of China (grant number of 52102453).

In this paper, an improved evolutionary prisoner’s dilemma (PD) game model is proposed by considering the weighting effect. Taking into account individual’s perceived payoff (benefits), the evolutionary tendency of the cooperators and three equilibrium points of the proposed model are obtained. We then numerically investigate how different exterior and interior factors influence on individuals’ cooperative behavior and their payoff both in the ER random network and the BA scale-free network. Our results reveal that the heterogeneous network structure is conducive to cooperation. In addition, the existence of leader nodes is an important driving force for promoting individuals’ cooperation. By further analyzing the rationality coefficient which appears in the weighting function, we obtain that a greater of irrationality could lead more people to take cooperative strategies. Finally, two indicators which are used to measure the real average payoff and perceived average payoff are defined. The results show that the real average payoff and perceived average payoff are larger in the heterogeneity network than that in homogeneous network.

Unmanned aerial vehicle (UAV) formation has a wide range of applications as it can increase the success rate of task and improve the reliability of system. Formation control and obstacle avoidance are two related key problems and are studied based on the consensus theory in this paper. First, two forms of kinetic models for UAV are described, and the standard model and the model with autopilot are deployed to apply in obstacle avoidance and formation control respectively. The constraints on the maneuverability of UAV are proposed. Based on the standard consensus algorithm, the three degrees of freedom (DOFs) of UAV in the improved consensus algorithm (ICA) are converted to be consistent with the DOFs describing the relative positions among UAVs, which make it possible to add the formation information to the standard consensus algorithm. The constraints on the maneuverability of UAV are also met by using the proposed minimal adjustment strategy. In the obstacle avoidance problem, the particle swarm optimization (PSO) algorithm is introduced, and the static obstacles and dynamic obstacles are dealt with by the ICA-PSO algorithm and the model predictive control (MPC)-PSO algorithm respectively, and the two algorithms can be combined to cope with more complicated situations. Simulation results demonstrate that the ICA can make UAVs with different initial states form the specified formation while satisfying all the constraints, and a safe and efficient flight for UAV formation can be ensured with the obstacle avoidance algorithm.

System Protection Communication Network (SPCN) is a new type of high-speed, real-time, secure and reliable communication network proposed in China supporting services such as AC/DC control, pumped storage control etc. In order to reduce the impact of SPCN failure on electric power system, this paper proposes a risk modeling and optimization approach. Firstly, we build a risk model to analyze the dynamic link and service risk from aspects of failure probability and its impact value. Then, we construct a risk optimization problem aiming at minimizing the link risk balance degree with service quality and risk constraints, and propose improved genetic algorithm to solve it. Based on part of network topology from a Chinese province, simulation results show that the proposed approach can make SPCN more reliable comparing to other methods when link failure occurs.