It is tailored for a context-aware decision-making process, enabling the generation of efficient maneuvers and optimal local trajectories to navigate dynamic environments and avoid obstacles seamlessly. The model-based control schemes ensure precise trajectory tracking. The solution also has configurable traversal policies, enhancing adaptability to a wide range of deployment areas.
- Offline mission generation
- Mission contains PoI and waypoints
- Actions can be added to each way points
- Actions: Wait, Orientation control, Move backward or forward for configurable distance, triggering attachment jobs
- Mission upload to the vehicle
- Real time mission parsing and execution
- For a mission the optimal path is calculated based on mission PoI’s
- A* based optimal path
- Cubic spline fitting for the optimal path
- Transforms waypoints (path) into vehicle coordinates
- Waypoint tracking and vehicle motion compensation
- RL based maneuver selection
- Modified Hybrid A* in continuous space
- A* based pre-search for heuristics to avoid local minima
- Usage of kinematically feasible primitives
- Takes care of path switch, path cost, maneuver cost, potential fields
- Customizable weights and vehicle footprint
- Scenario sensitive primitive selection
- Feasibility checks in discrete space
- Vehicle model-based controller
- Velocity profile generation
- Pure pursuit controller
- PID based yaw-rate and velocity control
- MPC based control (under development)