Object-tracking algorithm detects the object and recognizes it.
It keeps tracking till it disappears from its sight.
Once it detects a human. Once a human is specified and identified, it keeps tracking the human and doesn't lose its ID even if the target is blind for some time.
It memorizes all identified objects.
Our SLAM maps surroundings without twists in coordinates.
It maps seen/unseen environments and plans a path to a goal.
With a known 3D map, it finds a path to a goal in real time and avoids obstacles that were previously unrecognized.
It can semantically segment and recognize obstacles in various weather conditions.
We train the robot in simulation using reinforcement learning and implement on an application robot.
We train robots in simulation using reinforcement learning and implement on an application robot.
Coming soon