Use case
Following a project conducted with a large local industrial group, Logiroad.ai was able to prove the feasibility of solutions allowing a robot to identify loose parts, choose one that can be grabbed and calculate a trajectory to retrieve this part.
Several technologies are mobilized for this purpose.
- First, the robot performs a 3D scan of the parts using an adapted camera.
- The result in the form of colors and 3D points is sent to a neural network, which identifies the shapes of the parts.
- From there, a second algorithm decides which part to pick up first.
- The robot is then guided into the right position to catch the part.
Sensors and computing resources can be fully integrated into the robot, allowing it to perform rapid acquisitions and adapt to changes in the environment.
These devices can be adapted to massive parts that are difficult to handle, as well as to fragile parts or high production rates.
Once implemented, they can be adapted from one production run to another for different part shapes, making the production line agile.