Level Zero
GPS and INS (Inertial Navigation System):
Sensors:
- IMU,
- GPS,
- Wheel Odometry
Properties:
- Safety Level: Moderate to high.
- Description: The robot uses GPS to navigate between predefined waypoints autonomously.
- Safety Considerations: While GPS navigation reduces the risk of human error, it does not account for unforeseen obstacles or changes in the terrain. It also relies on GPS signal availability and accuracy.
Example: A GPS-guided vehicle might navigate safely in an open field but could get stuck or cause damage if unexpected obstacles like fallen branches are present.
Level One
Vision-Augmented Navigation:
Sensors:
- RGB-D NIR Cameras
- IMU,
- GPS,
- Wheel Odometry
Properties:
- Safety Level: High (based on how dynamic is the environment).
- Description: The robot navigates using a preloaded map of the environment (considering online or predefined desired trajectories).
- Safety Considerations: It can plan paths and avoid static obstacles. However, its ability to adapt to dynamic changes is limited, which can be a safety concern. Hence the environment must be kept static while the robot is operating in the field.
Example: The vehicle can safely navigate around permanent structures like trees or buildings (which have been present in the preloaded map), but might not detect a moving animal or a person stepping into its path.
Level Alpha
Dynamic Environment Navigation:
Sensors:
- Lidar
- RGB-D NIR Cameras
- IMU,
- GPS,
- Wheel Odometry
Properties:
- Safety Level: Very high (depending on how advanced the navigation algorithm is).
- Description: The most advanced level, where the robot could have a pre-defined map of the environment and consider new observations to update local features while operating in the environment, also responds in real-time to environmental changes by adapting motion trajectories.
- Safety Considerations: This level includes advanced sensors and algorithms that allow the robot to perceive and react to dynamic obstacles and changes, greatly enhancing safety
Example: The vehicle can detect and avoid a person walking into its path (e.g. in row-crop farms, the only valid action is stopping for the obstacle). Also, in areas like the row-switching part of the fields with more maneuverable spaces, it could avoid sudden dynamic changes like walking persons, cars, or broken branches of trees.
The page is being created, in case you want to know more about our navigation technology please check out our open-source versions on Git Hub or Contact us.