At SuperFolders Robotics, we are building a robot orchestration SaaS
platform — a unified control layer that abstracts the complexity of
multi-robot systems. Our mission is to empower businesses to seamlessly
connect their autonomous robots, assign high-level tasks through a
natural chat-based interface, and receive structured results, reports,
and media outputs — without needing deep robotics expertise.
Our platform enables organizations to deploy robots as reliable
teammates, capable of understanding intent, navigating complex
environments, and collaborating with humans and other machines.
System Architecture
Robot Orchestration Layer
We provide a centralized orchestration engine in the cloud that:
- Manages fleets of heterogeneous robots
- Handles task planning, dispatching, and feedback loops
-
Supports semantic commands such as: "Inspect the west wall,"
"Deliver coffee to the front gate," or "Capture a thermal image of
the transformer"
Edge Runtime Stack
Each robot runs a modular runtime built on:
-
NVIDIA Jetson (Orin/AGX/Xavier): Primary onboard
compute unit for mission execution, sensor fusion, SLAM, and AI
inference
-
ROS 2 Middleware: For modular, scalable robot
software integration
-
MQTT/WebSocket Communication Layer: For
persistent, low-latency two-way communication with our cloud
controller
-
Local Storage & Buffering: For real-time logging,
media caching, and offline resilience
Modular Accessory Kits & Cross-Platform Integration
To enable rapid integration with a wide variety of robot platforms
(quadrupeds, UGVs, UAVs), we are developing SuperFolders Modular
Accessory Kits that include:
- Mount-ready compute & sensor modules (Jetson-based)
- Quick-attach LiDAR and multi-camera arrays
- Unified sensor/power/data harnesses
-
Preconfigured firmware for instant registration to our orchestration
cloud
These kits serve as plug-and-play adapters — turning diverse robotic
hardware into smart, cloud-connected agents ready to receive tasks and
return mission results via the SuperFolders platform.
Perception, Navigation & Mapping
SLAM & World Modeling
-
3D LiDAR Integration: Robosense LiDARs enable
real-time 3D SLAM for both indoor and outdoor navigation
-
Map Stitching & Localization: Supports persistent
maps and global localization via RTK-enhanced GNSS
-
Digital Twin Generation: LiDAR data can be exported
as point clouds or meshed into 3D environments for:
- Forestry and terrain modeling
- Construction dimension verification
- Industrial asset inspection
RTK Positioning
Centimeter-Level Accuracy: Using GNSS RTK correction
data (via NTRIP/UBlox F9P), we enable 1–2 cm precision path planning for
critical deployments.
Computer Vision & Edge AI
Edge AI Inference
We utilize YOLOv8, EfficientDet, and custom-trained models for real-time
detection and classification:
- Human presence detection
- Structural anomaly detection
- Equipment identification
Running on:
- NVIDIA Jetson GPU (CUDA/TensorRT)
- Google Coral TPU (Edge TPU)
Future-Specific Models
We are building vertical-specific vision models optimized for:
- Industrial inspection
- Agricultural health analysis
- Building diagnostics
Semantic & Environmental Awareness
We are teaching robots to understand and reason about their
environments:
-
Semantic Mapping Layer: Robots tag elements in the
world (gates, bunkers, stairwells) with semantic labels
-
Natural Language Command Parser: Converts user
phrases into structured navigation and inspection commands
-
Shared Cognitive Maps: Robots share awareness with
one another, enabling collaborative missions
Example: Instead of "Go to waypoint 42.389, -71.128",
users can say:
"Go to the south gate and photograph the control panel."
Context-Aware Mobility & Skill Training
We're training robots to navigate and operate safely in complex,
variable environments:
Terrain-Adaptive Navigation
-
Forest paths: Avoiding roots, rocks, hidden holes
-
Construction sites: Handling debris, elevation,
narrow spaces
Environmental Task Modules
- Periodic patrols
- Object pickup/delivery in unstructured zones
- Multimodal inspection workflows
Sensor & Payload Expansion
We are integrating a suite of sensors that dramatically expand robot
capabilities:
| Sensor Type |
Purpose |
| Thermal Cameras (FLIR) |
Electrical inspection, heat loss analysis, HVAC diagnostics |
| Near-Infrared Cameras (NIR) |
Vegetation health, plant monitoring, agri-inspection |
| VOC Sensors |
Air quality, gas leak detection, environmental compliance checks
|
| IMU / Wheel Encoders |
Odometry and local stabilization |
| External RTK Base Stations |
Standalone GNSS correction for mobile deployment |
| Custom Sensor Add-ons |
Configurable based on client use-cases and environments |
Human-Robot Interaction (HRI)
Our robots will gain social and situational awareness through behavior
modules:
Social Skills Engine (Planned)
- Requesting help (e.g. "Please open the door")
- Communicating ambiguity ("What's behind the curtain?")
- Elevator interaction (ask humans to push buttons when needed)
What's Next
Our long-term vision includes:
- Multi-agent collaboration with shared goals
- Site-wide mission planning
- Self-updating semantic maps
-
Domain-specific robotic "skills" — e.g. warehouse delivery, forest
surveying, industrial QA
Whether it's an outdoor patrol bot, an indoor inspection agent, or a
multi-sensor mobile scanner — our stack adapts and evolves to power
the future of autonomous business operations.