中文/EN
AeroSTY
Technical Specifications
Detailed hardware and software architecture for industrial-grade autonomous safety landing.
Back to Home
Detailed hardware and software architecture for industrial-grade autonomous safety landing.
Back to HomeThe hardware must follow four core principles: redundancy, low power consumption, high reliability, and portability. It is designed as an independent "perception computing pod" or embedded module, physically isolated from the main flight controller (e.g., PX4) for EMI optimization, independent thermal management, simplified system maintenance, and cross-platform rapid integration capability.
硬件端必须遵循"冗余性、低功耗、高可靠性、可迁移性"四大原则。设计为独立封装的"感知计算吊舱"或内嵌模块，与主飞控（如 PX4）物理隔离，实现电磁兼容性优化、热管理独立化、系统维护便利性，以及跨平台快速适配能力。
The software architecture implements a comprehensive six-layer hierarchical system spanning from hardware foundation to safety-critical decision-making. This fully edge-converged architecture eliminates cloud dependency while ensuring deterministic safety response under all network conditions. The layered design enables adaptive computational resource allocation across the six tiers based on mission phase, environmental complexity, and system constraints. During UAV approach to landing target, the system operates in strategy cycles where L0/L1/L2 perception layers are executed with dynamically adjusted ratios rather than fixed sequential order. Flight velocity and direction are system outputs that feed back into the strategy cycle, creating a closed-loop optimization where environmental complexity, battery status, and safety requirements jointly determine the optimal ratio of L0:L1:L2 processing within each cycle to balance real-time responsiveness with environmental understanding depth.
软件架构实现了从硬件基础到安全关键决策的完整六层分层系统。这种完全端侧收敛的架构消除了云端依赖，同时在所有网络条件下确保确定性的安全响应。分层设计使得系统能够根据任务阶段、环境复杂度和系统约束，在六个层级间动态分配计算资源。在无人机向目标着陆点近进过程中，系统以策略循环方式运行，L0/L1/L2感知层以动态调整的比例执行，而非固定的顺序执行。飞行速度和方向作为系统输出，会反馈到策略循环中，形成闭环优化机制——环境复杂度、电池状态和安全要求共同决定每个循环中L0:L1:L2处理的最优比例，以平衡实时响应能力与环境理解深度。
(1-2 years, Industrial Grade Practical Implementation)
（1-2年内，主打工业级实用化）
Positioning: Solve the "survival" problem for industrial drones in complex environments and sudden failures, achieving safe emergency landing and basic obstacle avoidance.
定位： 解决工业无人机在复杂环境和突发故障下的"活下来"问题，实现安全迫降与基础避障。
(2-3 years, High-Reliability Industrial Applications)
（2-3年内，主打高可靠性工业应用）
Positioning: Introduce independent flight controller system on top of Generation 1, achieving fully independent safe landing even when the main flight controller completely fails.
定位： 在第一代基础上引入独立飞控系统，实现主飞控完全失效情况下的独立安全降落。
(3-4 years, VTOL Transformation for Fixed-Wing)
（3-4年内，固定翼 VTOL 改造）
Positioning: Based on Generation 2's independent flight controller, equip with independent quad/hex rotor frame system to extend fixed-wing aircraft into hybrid wing aircraft.
定位： 在第二代独立飞控基础上，配备独立的四/六旋翼框架系统，将固定翼飞行器扩展成复合翼飞行器。
(4-5 years, eVTOL Pilot & All-Weather)
（4-5年内，主打eVTOL试点与全天候）
Positioning: Transition from industrial drones to manned eVTOL, solving all-weather complex perception and preliminary airworthiness certification requirements.
定位： 从工业无人机向载人eVTOL领域过渡，解决全天候复杂感知与初步的适航认证要求。
(5-6 years, Multi-Agent Coordination & Swarm Intelligence)
（5-6年内，主打多机协同与群体智能）
Positioning: Become UAM infrastructure supporting large-scale, high-density low-altitude airspace collaborative operations.
定位： 成为城市空中交通（UAM）的基础设施，支持大规模、高密度的低空空域协同运行。
Extending progressive world modeling capabilities to ground robots and embodied intelligence systems
将渐进式世界建模能力扩展到地面机器人和具身智能系统
Positioning: Provide aviation-grade safety perception capabilities for industrial robots.
定位： 为工业机器人提供航空级安全感知能力。
Positioning: Provide complex indoor environment understanding capabilities for service robots.
定位： 为服务机器人提供复杂室内环境理解能力。
Positioning: Build a general embodied intelligence platform for cross-scenario transfer learning.
定位： 构建通用具身智能平台，实现跨场景迁移学习。