latest release: AD Swarm Manager v1.0
- Multiplatform support – quadcopter, fixed wing and boat
- Compatible with DJI, Pixhawk / Ardupilot autopilot systems
- Fully autonomous security and inspection missions available
- Sensor data / video feeds from all connected vehicles
- User interface optimised for non specialists
- Mission builder for easy mission creation and customisation
- Swarm collision avoidance v1.0
- Path planning v1.0
- Drone learning v1.0
Swarm Manager demo
Check out Swarm Manager in action below.
Autonomous inspection mission
Autonomous perimeter security mission
AD Swarm Manager features
Vehicle platform agnostic
Swarm Manager is designed to be completely agnostic of the hardware it is controlling. We prepare device drivers for the robot platforms you want to use: quad-copter, fixed wing, boat, car, submarine, satellite. In addition, we support external IoT devices to facilitate better sensing and decision making.
3rd party tool integration
Swarm Manager integrates seamlessly with data analysis tools, such as video analytics processors, photo stitching software, etc.
Game like controls
Swarm Manager treats drones and other unmanned assets as non-playing characters (NPC) in a computer game, where each asset is controlled by a central computer running our AI software. The operator can drag and select unmanned assets on the map, supervise their health (battery status) and actions, dispatch missions, have access to their camera feeds through an interactive Accelerated Dynamics UI. The UI provides unmatched situational awareness and allows to assess the fleet status at a glance without being overwhelmed by information.
Variable autonomy levels
Swarm Manager’s AI system is modelled with four levels of autonomy that are available to the user:
- 1. The AI offers no assistance – all controls manual.
- 2. The AI suggests a plan, and executes if the user approves.
- 3. The AI suggests a plan and allows the user a restricted time to veto before automatic execution.
- 4. The user is not involved in the decision-making process, the system decides and executes autonomously.
3D collision avoidance
Our system provides autonomous three dimensional waypoint generation for all the drones within the fleet when working in the mission and ensures that they do not collide while traversing their paths. In addition, we provide the user with manual point click waypoint collision-free navigation for semi-manual mission tasks.
Deployed on the cloud or local machines
Swarm manager can be deployed on a local machine or private servers depending on your preferences. When it is running on remote servers multiple users can be allowed to access the drone fleet as operators or observers.
Accelerated Dynamics Swarm Manager uses deep reinforcement learning technology for multi-agent behaviour optimisation. This enables collective learning capability of agents in the fleet improving their mission performance continuously at all times. In addition, you can choose to teach agents according to your input to do missions the way you like.