GPS-Denied Autonomy is Powered by Machine Vision Systems
Multiple machine vision based technologies combined with navigation algorithms to deliver GPS-Denied autonomy. Without detecting and tracking objects and features in the environment around the UAV, navigation algorithms wouldn't be able to guide in accordance with its mission.
Autonomy in New, Unmapped Flight Terrain
In unknown and unmapped flight systems, the same methodology is deployed on the edge, and optionally across mesh-networked UAV. When no pre-flight spatial data exists, drones map the area they've flown through and use extracted visual feature-sets for precise reference of movement versus an initial location. The map created during flight can be used on subsequent missions and any other connected UAV.
Autonomy & Positioning without GPS
Rhoman's shared memory system allows drones to use 3D spatial maps (Google Earth, pre-created Digital Twins from reconnaissance, other gathered data) to perform autonomous missions without GPS. The system let's drones
-
Use their existing cameras to identify objects and terrain feature-sets in the real world
-
Compare that environmental detections with stored 3D data (locally stored or updated over 5G for long, commercial flights)
-
Estimate a continuously running GPS-free position estimate
-
Maintain positioning and awareness to complete missions and maintain full capabilities when GPS is blocked or spotty
Spatial Awareness with Existing 3D Maps and Digital Twins
Status Quo
Standard GPS systems show 2D location when connected
Rhoman Cloud
Synthetic GPS provides 3D flight context to UAV for safe & aware operations
Enhanced control systems merge with Synthetic GPS for safe route margin
UAV Shared Memory & Autonomy
Every drone flight is just as safe, or just as dangerous, as that drone's first flight. Shouldn't drones learn from their experience to get better and safer over time?
The Rhoman system provides Flight Muscle Memory, Perception & Awareness, and UAV Memory & Self-Learning to make drones safer, able to understand their environment, and able to execute complex flight maneuvers for performance and safety with and without GPS.
Shared situational awareness (SA) can support decision makers, individual drones flying through a repeat area... and collaborative swarms; including in GPS-Denied environments.
The unique combination of 1) ML-adaptive controls, 2) tuning from prior and shared flight data, and 3) macro-system data-share allow Rhoman and our partners to create a system that fills future needs as commercial UAV expand in urban, suburban, and contested flight-space areas.
Synthetic GPS
Rhoman uses it's network and experience with underground terrain mapping to merge environmental data into a shared network of geo-tagged point clouds and a recognizable camera/LIDAR digital twin of flight engagement areas shared UAV network-wide.
Synthetic GPS
Drone-Drone Interaction
"If every drone had experienced hundreds of autonomous drone-drone interactions, and learned from the encounters - imagine the smooth, autonomous aerial traffic we'd have." Rhoman's system shares our UAV-UAV interaction data network wide.
Drone-Drone Interaction & Swarms
Trustworthy Autonomy
Constraint Barrier Functions
deployed through Rhoman's adaptive controllers - and tuned using Rhoman's network amd shared data systems - provide trustworthy autonomy for BLOS autonomous missions in complex airspace.
Safe & Trustworthy Autonomy
New Capabilities
Because we deploy our capabilities through software and maintain networks with UAV, our systems are uniquely able to connect with drone hardware, sensors, and compute elements in order to deploy future AI technologies and algorithms as they arise.
Future AI & 3rd Party Tech
Enhanced Control Solutions Technology
Rhoman Aerospace deploys innovative adaptive flight control solutions that are uniquely able to benefit from machine-learning tuning systems applied to embedded tunable parameters and control system gains.
Our machine learning adaptive control systems let us 1) get UAV flying faster and at a lower cost than other development options, 2) insert optimizable functions within control-equations and sensor-input-systems to enhance capabilities and obstacle avoidance through ML tuning, and 3) enable UAV self-learning over a shared network for fleet-wide power savings, flight route optimization, and environmental data-share.
Enabling Technologies
Essential Capabilities
Machine Learning Adaptive Controls
ML-Adaptive Controls:
-
Self-tune to unique vehicle configurations
-
Uniquely suited for ML tuning w/ prior flight data
-
Integrate enviro-sensor data to multiple levels of the control system
Adaptive CG Algorithms:
-
Auto-account for off-center CG
-
Auto-account for live CG deltas
-
Handle ad hoc payload
-
Lean-there Go-there controls
ML-Adaptive Controls:
-
Get any unique UAV flying ASAP
-
Enable system optimization
-
Trustworthy obstacle avoidance
Adaptive CG Algorithms:
-
Maintain safety and stability with hanging cargo
-
Deploy payloads to precise drop-zones w/out landing
-
Make any drone multi-purpose
-
Smooth intuitive flight
Software Download
​​
Software Download
​​
Tuning & Optimization
ML-Adaptive Controls:
-
Self-tune to unique vehicle configurations
-
Uniquely suited for ML tuning w/ prior flight data
-
Integrate enviro-sensor data to multiple levels of the control system
Adaptive CG Algorithms:
-
Auto-account for off-center CG
-
Auto-account for live CG deltas
-
Handle ad hoc payload
-
Lean-there Go-there controls
Power Saving Algorithms:
-
Reduce RPM-differentials
-
Optimally orient vehicle
Control Barrier Functions:
-
Use prior flight data for tuning
-
Create theoretical flight path guarantees
ML-Adaptive Controls:
-
Get any unique UAV flying ASAP
-
Enable system optimization
-
Trustworthy obstacle avoidance
Adaptive CG Algorithms:
-
Maintain safety and stability with hanging cargo
-
Deploy payloads to precise drop-zones w/out landing
-
Make any drone multi-purpose
-
Smooth intuitive flight
Power Saving Algorithms:
-
Keep motor in most efficient RPM range
-
Reduce power use
Control Barrier Functions:
-
Create trustworthy flight path routes
Enable trustworthy obstacle avoidance
Software Download
​​
Software Download
​​
Fleetwide UAV Capabilities
ML-Adaptive Controls:
-
Self-tune to unique vehicle configurations
-
Uniquely suited for ML tuning w/ prior flight data
-
Integrate enviro-sensor data to multiple levels of the control system
Adaptive CG Algorithms:
-
Auto-account for off-center CG
-
Auto-account for live CG deltas
-
Handle ad hoc payload
-
Lean-there Go-there controls
Power Saving Algorithms:
-
Reduce RPM-differentials
-
Optimally orient vehicle
Control Barrier Functions:
-
Use prior flight data for tuning
-
Create theoretical flight path guarantees
Flight Navigation Layer:
-
Target layer between waypoints and vehicle navigation that is tunable with out ML systems and prior flight data
Shared Data Systems:
-
Fleet-wide data increases fidelity and ML model robustness
-
Networked UAV paths can be shared with the cloud and UAV
-
Environmental data is saved and sharable to connected UAV
ML-Adaptive Controls:
-
Get any unique UAV flying ASAP
-
Enable system optimization
-
Trustworthy obstacle avoidance
Adaptive CG Algorithms:
-
Maintain safety and stability with hanging cargo
-
Deploy payloads to precise drop-zones w/out landing
-
Make any drone multi-purpose
-
Smooth intuitive flight
Power Saving Algorithms:
-
Keep motor in most efficient RPM range
-
Reduce power use
Control Barrier Functions:
-
Create trustworthy flight path routes
Enable trustworthy obstacle avoidance
Flight Navigation Layer:
-
100% Precision flight path following
-
Hit waypoints and security perimeters with 100% accuracy
-
Whole fleet gets cumulative benefits
Shared Data Systems:
-
Environmental data share: geo-tagged, network-distributed 3D point clouds and enviro maps
-
Backup shared optical positioning
-
Optimized tuning parameters reach new drones in fleet
-
2X Safety positioning redundancy
Software Download
​​
Software Download
​​
Building off of our machine learning adaptive controllers, and leveraging the unique advantages of these systems, we provide value-add control systems to individual drones and fleetwide, scalable benefits to complete drone networks and operations.
Software Download Solutions
Our machine learning adaptive control systems let us 1) get UAV flying faster and at a lower cost than other development options, 2) insert optimizable functions within control-functions and sensor-input-systems to enhance capabilities and obstacle avoidance through ML tuning, and 3) enable UAV self-learning over a shared network for fleet-wide optimization and data-share.
Our machine learning adaptive control systems let us 1) get UAV flying faster and at a lower cost than other development options, 2) insert optimizable functions within control-functions and sensor-input-systems to enhance capabilities and obstacle avoidance through ML tuning, and 3) enable UAV self-learning over a shared network for fleet-wide optimization and data-share.
