The RBX-SAR-H6 is a field-tested autonomous hexacopter built for Search & Rescue operations. It detects humans using onboard Edge AI, executes autonomous area sweeps, and delivers payloads all without internet, all without a skilled pilot.
Pakistan is among the most flood-affected countries in the world. Current emergency response is stretched thin, dependent on manual operators, stable communication links, and roads that floods destroy first.
Ground teams take hours to cover terrain an aerial platform sweeps in minutes, in conditions too dangerous to enter on foot.
Manually piloted drones demand skilled pilots who fatigue over extended missions, causing inconsistency precisely when focus matters most.
Disaster zones have no internet, no roads, no cloud. Systems that depend on connectivity fail at exactly the moment they are needed.
The RBX-SAR-H6 uses a three-node distributed computing architecture: a browser-based Ground Control Station sends mission commands, a Raspberry Pi 5 runs the AI detection pipeline and relays decisions via MAVLink, and a Pixhawk flight controller executes precise autonomous flight. Every node communicates in real time over a local WiFi link , no internet, no cloud.
Unlike traditional drones controlled by a single flight controller, the RBX-SAR-H6 distributes intelligence across three nodes that communicate in real time.
The operator's command hub. Runs entirely in a browser , no app installation. Connects over a local WiFi hotspot. Displays live telemetry, camera feed, detection markers, and flight mode. Issues mission commands to the rest of the system.
A Raspberry Pi 5 , palm-sized, yet powerful enough to run YOLOv8 at 15+ FPS entirely on-device. It receives commands from the GCS, runs the detection pipeline, and translates mission decisions into MAVLink commands sent directly to the Pixhawk.
The Pixhawk 2.4.8 running ArduCopter handles all low-level flight , motor mixing, stabilisation, GPS navigation, and failsafe. If battery drops critically or signal is lost, it autonomously triggers Return-to-Launch via GPS without any operator input.
Automatically lands at the exact takeoff point using GPS. Triggers on signal loss or critically low battery.
Maintains a fixed altitude against external forces , ideal during search sweeps over uneven terrain.
Holds position against wind , useful when hovering over a detection point before payload deployment.
The primary autonomous mode. The Raspberry Pi sends real-time waypoints via MAVLink during missions.
Locks both altitude and GPS position simultaneously , maximum precision hovering in degraded conditions.
Self-calibrates PID flight parameters by sampling real environmental data , ensures stable flight in new conditions.
Sum of all components per manufacturer spec sheets
2,170g dry + 500g minimum SAR payload
This platform establishes the foundational framework for higher-capability variants: Nvidia Jetson Orin for faster inference, thermal imaging, swarm coordination, custom 3D-printed frames, and low-noise propellers.
The custom GCS runs in any browser connected to the drone's local WiFi hotspot , no installation, no internet. It gives the operator full situational awareness: live camera feed with AI detection overlay, real-time telemetry, interactive map, and one-tap mission controls.
Actual GCS running live during field test at IUB campus
The RBX-SAR-H6 is available as a complete hardware solution, a software-only license for existing UAV platforms, or a custom-configured variant. Get in touch to discuss your operational requirements.
Complete assembled platform , airframe, avionics, GCS software. Ready to fly.
GCS + AI pipeline licensed for integration into your existing drone platform.
Custom detection models, payload systems, or comms protocols for your use case.
Built by Published researchers · Registered under IPO Pakistan