Division 01

Perception, autonomy
and flight —
engineered end to end.

YTK Robotics takes machines from sensing to acting. We build the visual intelligence, the radio link, the onboard inference and the autonomy layer that lets a humanoid, an industrial arm or a drone operate in the real world — where light is bad, the network drops out and the cost of being wrong is physical.

AI Vision Drones RF Technology AI Agents Edge AI
01 The division

The hard part isn't
the model. It's everything
around it.

A detector that scores well on a clean dataset tells you almost nothing about a machine working at 45 °C with a smeared lens and a vibration problem. So we engineer the whole chain: sensor selection, calibration, timing, compute budget, comms and the fallback behaviour when any of it degrades.

That is why the division holds five disciplines in one team rather than handing a spec between vendors. Vision, radio and inference are coupled problems; treating them separately is how robotics programmes stall.

perception_pipeline · live streaming
HUMANOID 0.97 DEPTH 2.41 m INFER 11.2 ms CONF 0.97 FPS 90 LATENCY · 60 s WINDOW p99 14.8 ms

Fig. 01 — perception and latency, as instrumented in the field

02 Five capabilities

What YTK Robotics
actually builds.

Each capability stands alone as an engagement, and all five are designed to interlock. Most programmes start with one and end up touching three.

01 / VISION

AI Vision & visual intelligence

Detection, segmentation, depth estimation, tracking and pose — trained on your footage and validated against the failure modes that matter: glare, dust, occlusion, motion blur, night operation.

02 / AIR

Drones

Multi-rotor and hybrid platforms, flight control tuning, mission autonomy, precision positioning and payload integration for inspection, survey and patrol.

03 / RF

RF technology

Link budget analysis, antenna selection, mesh and long-range telemetry, spectrum planning and interference mitigation for contested airspace and industrial sites.

04 / AGENTS

AI agents

Mission-level autonomy: task decomposition, tool and payload control, replanning on failure, and supervision that keeps a human accountable for consequential actions.

05 / EDGE

Edge AI

Smaller models running locally on Jetson-class and embedded accelerators — quantised, pruned and scheduled to fit real thermal and power envelopes without a cloud round trip.

COMBINED

Integrated platforms

Humanoids, industrial robots and drones as complete systems — perception, compute, radio and autonomy brought together and tested as one machine.

03 System architecture

Sensors in.
Motion out.

Every platform we build follows the same five-layer spine. It keeps the failure boundaries explicit: perception never commands an actuator directly, and the radio layer is treated as unreliable by design.

platform_stack sensors → actuation
L05 · SENSORS L04 · PERCEPTION L03 · PLANNING L02 · CONTROL L01 · ACTUATION RGB · DEPTH · IMU · GNSS · RADAR DETECT · DEPTH · TRACK · FUSE MISSION · AGENT · REPLAN PID · MPC · SAFETY ENVELOPE MOTORS · JOINTS · PAYLOAD RF LINK MESH · LORA 5.8 GHz VIDEO LINK BUDGET FALLBACK PATH TELEMETRY CLOSED LOOP · 200 Hz
04 Why edge

A round trip to
the cloud is a
safety margin lost.

For a machine that moves, latency is not a performance metric — it is the distance travelled before it reacts. We shrink the model until it fits the compute envelope on the device, and keep the network for coordination rather than control.

Typical trade at the deployment stage: a modest accuracy delta in exchange for an order-of-magnitude reduction in reaction time, plus operation that continues when the link does not.

Cloud inference round trip180–420 ms
On-device edge inference8–16 ms
Dependent on network availabilityYes
Dependent on network availabilityNo

Indicative figures from lab benchmarks. Actual budgets are set per platform.

05 Where it applies

Operating environments
we design for.

Not a list of markets — a list of conditions. If your deployment looks like one of these, the engineering is already close.

A

Industrial & energy inspection

Assets that are tall, hot, pressurised or hard to reach. Close-range visual capture, thermal overlay and defect flagging, with the drone holding position in wind.

B

Infrastructure & asset survey

Corridors, sites and structures mapped repeatedly to a consistent georeferenced baseline, so change between surveys is measurable rather than anecdotal.

C

Security & perimeter patrol

Autonomous routes, anomaly detection and human escalation — designed so an operator can verify any alert quickly enough to act on it.

D

Warehouse & fleet autonomy

Indoor positioning without reliable GNSS, mixed human and machine traffic, docking and handover, and pick-and-place where misidentification is expensive.

E

Humanoid & mobile manipulation

Whole-body perception, grasp planning and safe physical interaction in spaces built for people — with torque and proximity limits enforced below the planner.

06 Toolchain

Built on systems
that survive contact
with hardware.

ROS 2 TensorRT ONNX Runtime NVIDIA Jetson CUDA PX4 ArduPilot MAVLink SLAM Kalman / particle filters YOLO-class detectors Stereo depth LiDAR LoRa / mesh radio 5.8 GHz video RTK GNSS C++ / Rust Python

Have a machine that needs to see, decide or fly?

Send the operating conditions. We will tell you what is feasible and what is not.

cto@ytkresearch.com +971 56 230 0805 Emirates Tower, Dubai — UAE