By now, you should be an expert of the following components 😉:
We want you to develop a module that employs the Joint Interface, the Cartesian Interface
and the Gaze Interface to accomplish the following tasks:
- Make iCub look down at the table.
- Detect the object lying on the table.
- Retrieve the object position in the Cartesian domain.
- Let iCub look at the object.
- Based on the object position, select the best hand.
- Control the iCub fingers to achieve suitable grasp configurations.
- Let iCub approach the object to enable a top grasp.
- Ask iCub to grasp the object.
- Lift the object.
This assignment is peculiar in that we can interface the simulator and the real robot indifferently, being the code capable of dealing with object location retrieval in both conditions, transparently to the user.
The outcome should look like the animation below with the simulator:
Some of the points reported above have been already addressed in the code (e.g. object detection and location retrieval),
so you need to fill in the missing gaps highlighted by the comment // FILL IN THE CODE in the src/main.cpp module.
⚠ Don't speed up the movements by reducing the trajectory time of the operational controllers: they're already all set to work with the real robot.
Once done, you can test your code in two ways along with the simulator:
- Manually: running the yarpmanager scripts provided from within app/scripts and yielding corresponding commands to the module rpc port.
- Automatically: running the script test.sh in the smoke-test directory. Take into account these important points:
- We use a timeout of 240 seconds to check the status of rpc communication, meaning that you have 240 seconds max to accomplish each rpc command.
- When you reply to rpc commands, we assume the robot has finished the movement.
- The smoke-test will add a random displacement to the initial position of the ball in order to force the use of both hands 😉
If you pass the test on the simulator, 🕒 book the robot 🤖 to get a real experience!