Drones are rapidly evolving from remotely piloted platforms into autonomous systems that must perceive their environment, understand context, make real-time decisions, and act accordingly. High-performance AI and superb vision processing are essential to interpret sensor data for tasks such as obstacle detection, navigation, and scene understanding, especially in dynamic, constantly changing and unstructured environments.
Drone applications are highly sensitive to latency, reliability, and connectivity. Communication links can be intermittent or unavailable, and round-trip latency is incompatible with real-time control loops. In addition, drones operate under strict SWaP-C (Size, Weight, Power, Cost) constraints. For all these reasons, drones require highly efficient on-device processing to ensure long flight time and predictable behavior without dependence on external infrastructure. Hailo’s edge AI processors enable advanced perception and autonomy directly onboard the drone, eliminating dependence on connectivity, while delivering high performance per watt, and working at a very low power consumption, in accordance with battery operation and low weight requirements.
Hailo AI processors deliver the compute required for advanced perception while minimizing power consumption, directly extending flight time
Predictable performance is critical for real-time autonomy and safety-related workloads
Hailo enables advanced AI capabilities at low power consumption which requires no active cooling, therefore reducing payload size and system cost
Autonomous flight requires continuous perception of the environment to detect obstacles, track target objects, perform localization & mapping and understand spatial and temporal context, especially for collision avoidance and BVLOS (Beyond Visual Line of Sight) operation. These capabilities rely on compute-intensive vision models and multi-stage perception pipelines that must execute in real time to support fast control-loop decisions.
Running this AI efficiently on the drone is essential to meet latency and safety requirements. Hailo accelerates these workloads efficiently, enabling robust autonomy within tight SwaP-C limitations and making advanced autonomous flight practical even for compact, entry-level drones
Many drone missions take place in low-light or nighttime conditions, where sensor noise degrades visibility and downstream perception accuracy. AI-based low-light denoising and video enhancement models significantly improve image quality, enabling reliable detection, navigation and decision making in challenging lighting environments.
Hailo enables superior low-light visibility, preserving color, texture and sharpness and reducing motion blur at extreme low-light conditions (<0.01LUX) improving low-light and night-time perception without increasing sensor cost, power consumption, or system complexity.
