Research

Research first

Brobot is research-first. Before we sell anything, we are building depth in the domains that full-stack autonomy requires. This page lists the areas we actively work in, what each covers, and how mature our work is — honestly labelled.

Method

How a research thread runs

Each domain runs as a thread: literature and prior art, a minimal working prototype, tests with recorded data, and a written note. Threads feed the roadmap — nothing graduates to 'product' until it has survived the field.

Active means we are doing hands-on work now. Exploratory means we prototype and study but it is not the critical path. Planned means it is on the roadmap and we are explicit that work has not started.

Domains

Engineering domains

Status labels are maintained, not aspirational: active work, exploratory prototyping, or plainly planned.

Autonomous navigation

active

Getting a vehicle from A to B without a human in the loop: path planning, obstacle handling, and behaviour under degraded GPS.

Path planning
State estimation
GNSS-degraded flight
Failsafe design

Swarm & multi-agent systems

active

Decentralised coordination for fleets: inter-agent communication, task allocation, and emergent collective behaviour across distributed vehicles.

ROS 2 multi-agent comms
Task allocation
Consensus & formation
Fleet simulation

UAV & VTOL platforms

active

The aircraft themselves: airframe design, flight-control logic, propulsion sizing, and the embedded systems that keep them in the air.

VTOL flight dynamics
Flight controllers
Embedded firmware
Hardware prototyping

Geospatial intelligence (GeoAI)

active

Turning sensor data into spatial understanding: mapping, terrain analysis, and location-aware decision-making for mission planning.

Spatial data pipelines
Terrain analysis
Mission planning
Remote sensing

Computer vision & perception

exploratory

What the machine sees: detection, tracking, and the perception stack that feeds navigation and mapping.

Object detection
Visual tracking
Sensor fusion
Camera calibration

SLAM & 3D mapping

exploratory

Simultaneous localisation and mapping: building 3D representations of unknown environments while navigating them.

Visual-inertial SLAM
Point-cloud processing
3D reconstruction
Loop closure

Edge AI & embedded intelligence

exploratory

Running real inference on real constraints: model optimisation and deployment on flight-grade compute, power, and thermal budgets.

Model quantisation
On-device inference
Compute selection
Power budgets

IoT & fleet connectivity

exploratory

The links between machines and the ground: telemetry, lossy-network tolerance, and secure fleet communication.

Telemetry links
Mesh networking
Protocol design
Link security

Future robotics & aerospace systems

planned

Where the stack goes next: ground vehicles, larger airframes, and higher-altitude platforms that reuse the same autonomy core.

Ground robotics
Larger UAV classes
High-altitude concepts