The architecture of Integrated Modular Avionics (IMA) provides airborne software with a robust temporal partitioning mechanism, which achieves the reliable fault containment between avionics applications. However, the partition scheduling of an IMA system is a complex nonlinear non-convex optimization problem, making it difficult to solve the optimal temporal allocation for partitions using traditional analytical methods. This paper presents a model-based approach to optimizing the partition scheduling of IMA systems, whose temporal behavior is modeled as a network of timed automata. Given a system model, the optimizer employs a parallel genetic algorithm to search for the optimal partition resource parameters with respect to minimum processor occupancy. For each promising parameter combination, the schedulability constraints and processor occupancy of the system are precisely evaluated by Classical and Statistical Model Checking (i.e., CMC and SMC), respectively. We also apply SMC hypothesis testing to the fast falsification of non-schedulable solutions, thereby speeding up the schedulability verification based on CMC. Two case studies demonstrate that our proposed approach outperforms classical analytical methods on the processor occupancy of typical IMA systems.

With the rapid development of unmanned aerial vehicle (UAV) swarm technology, achieving high-precision, multi-constraint cooperative path planning in complex mountainous environments has become a prominent research focus in intelligent control. However, traditional optimization algorithms often suffer from slow convergence, premature stagnation, and limited global search capabilities when addressing high-dimensional, nonlinear spaces. To overcome these challenges, this paper proposes a novel Adaptive Elite Grey Wolf-guided Crested Porcupine Optimizer (AEGWCPO), designed to enhance cooperative path planning for multi-UAV systems in complex 3D environments. The method first constructs an adaptive CPO by introducing a Good Point Set initialization strategy to improve population uniformity. It also uses a fitness-state-based perturbation mechanism to adaptively adjust step sizes and incorporates periodic Lévy disturbances to enhance local escape capability, improving search accuracy and robustness. AEGWCPO integrates a dynamic convergence trend detection mechanism that activates the Elite Grey Wolf-guided Optimization during stagnation phases. By using elite solution centroids, it performs jump-based updates in global and local exploitation stages, strengthening the algorithm’s guidance across multiple search phases. By combining local perturbation and dynamic elite guidance, AEGWCPO effectively balances exploration and exploitation in high-dimensional, complex path planning, enhancing both convergence speed and global optimization. Simulation results show that AEGWCPO outperforms multiple state-of-the-art algorithms in 3D mountainous environments in terms of convergence rate, path quality, and adaptability, offering an efficient and robust solution for multi-objective UAV swarm path planning.

With the continuous development of Unmanned Aerial Vehicles(UAVs) technology, UAVs are gradually applied to the collection of remote sensing information. In the process of collection, researchers generally choose to use laser point cloud data to complete the automatic driving of UAVs. However, the accuracy of the current laser point cloud semantic segmentation technology is not ideal, which is mainly attributed to the following two aspects: (1) The point cloud of the UAVs perspective will lose a lot of spatial features during the training process; (2) The point cloud of objects with different scales under the UAVs perspective has different density and distribution. To solve this problem, this paper proposes a semantic Segmentation technology for UAVs perspective laser point cloud data, namely Aero Auto Pilot Segmentation Network (AAPSN). The technology is improved on the basis of polar coordinate segmentation network, and by adding a multi-scale feature extraction module, the point cloud data features of objects of different scales can be extracted. In addition, attention mechanisms are introduced in the pooling process to focus more on the spatial structure characteristics of cloud data, thus improving the accuracy and efficiency of segmentation. The network model proposed in this paper improves the mean Intersection over Union (mIoU) index by 3.3% and 19.2% respectively on two real urban LiDAR datasets.