Аңдатпа
Unmanned Aerial Vehicles (UAVs) have emerged as pivotal tools for addressing region-specific challenges in Kazakhstan, a nation characterized by vast geographic diversity, extreme climatic conditions, and infrastructural demands in remote areas. However, deploying UAVs in Kazakhstan’s unique operational environments—marked by temperature extremes (-40°C to +45°C), unpredictable wind gusts (15–20 m/s in the Almaty and Kostanay regions), and frequent GPS signal degradation in mountainous terrain—poses significant technical and logistical challenges. Physical testing of UAV control algorithms under these conditions is not only prohibitively expensive but also constrained by safety regulations, environmental unpredictability, and the sheer scale of operational zones. To address these barriers, this article proposes the development of a Kazakhstan-centric UAV simulation platform, designed to emulate the country’s environmental and operational realities with high fidelity. Built on the Robot Operating System (ROS Noetic) and Gazebo 11, the platform integrates three novel components: (1) physics-based UAV dynamics calibrated using field data from Kazakh agricultural and disaster-response UAV deployments, including mass (1.5 kg), inertia tensor, and rotor thrust profiles; (2) synthetic sensor models (LiDAR, IMU, RGB cameras) with noise profiles tailored to regional conditions, such as dust-induced LiDAR range errors (±0.15 m) and temperature-dependent IMU drift (0.2°/hour at +40°C); and (3) environmental disturbance models derived from meteorological datasets provided by Kazhydromet, Kazakhstan’s national weather agency, including steppe wind dynamics (gusts up to 18 m/s) and probabilistic GPS signal loss (25–35% dropout rates in the Tian Shan mountains). The platform’s modular architecture supports testing of adaptive control algorithms, including Model Predictive Control (MPC) for wind disturbance rejection, swarm coordination strategies for search-and-rescue missions, and reinforcement learning (RL)-based fault tolerance systems, under scenarios mirroring real-world Kazakh challenges. Case studies demonstrate its efficacy: in simulated high-wind scenarios (18 m/s gusts), a decentralized swarm coordination algorithm achieved 88% mission success in maintaining formation over the Tian Shan mountains, while an adaptive PID controller reduced trajectory tracking errors by 35% under +40°C sensor drift conditions. Cross-validation with field data from a DJI Matrice 300 UAV deployed in the Turkestan region confirmed a 94% correlation between simulated and real-world trajectory RMSE (0.12 m vs. 0.15 m), with energy consumption predictions deviating by less than 3% from observed values.