Deadlines

Milestones

Milestones have been defined to help you meet the requirements of the project on time (we call them milestones because we see them as your milestones rather than our tollgates). You do not have to meet the milestones before the deadlines but years of experience shows that it helps the groups a lot and you will be rewarded for it.

The tasks corresponding to the milestones may seem simple but, do not underestimate them and use your time wisely. Do not leave the things for the final week, this is a real world situation where your intelligence is not the only thing that matters but your ability to organize your work and to account for the unpredictable.

You need to show your TA that you robot meet the requirements of the milestone and be ready to explain how it works. Note that the TAs are not sitting and waiting doing nothing. You need to schedule time with them.

Everything you show for a milestone should be running on the robot with no cables attached.

Milestone 1 (MS1) 

Meeting this milestone means that:

• Your robot should be able to drive in the maze. Keyboard control is OK. This means that you can control the motors and the motion of the robot.
• Your robot should have a dead reckoning system (odometry) on the robot.
• You should be able to display the position of the robot in rviz as it moves through the environment.
• You should be able to display the current scan from the laser in rviz as the robot moves. 

Your robot should now be able to move, use standard messages to visualise information in the ROS standard visualiser (rviz) and you have defined a proper transformation between the robot frame and the laser.

Milestone 2 (MS2) 

In addition to MS1, meeting this milestone means that:

• Using the RGBD camera your robot can detect and locate one of the objects (of your choice) that the robot can pick up.
• You should be able to pick an object with the arm given a pose estimate of the object. The robot can do this in places where you have made sure that the the arm can move freely but you need to show that the pose estimate of the object is with respect to the robot and not camera (i.e. you can handle transformation between robot, arm and camera).
• You should display the detected object in rviz so that one can see its pose with respect to the robot (means you can estimate its pose).
• You should be able to display every Nth scan in rviz. This should clear show the performance of your dead reckoning system. This could also be done in the form of an early version of a map if it fits your plans (that is, do not build a map just for this).

Your robot should now also be able to use the camera to detect objects. You can also use your arm and you will have started realising that the arm is in the way when you move and when you are close to obstacles.

Milestone 3 (MS3)

In addition to MS1 and MS2, meeting this milestone means that:

• You can read the map file and turn that into your own map representation and display it along with the position of the robot.
• You have a localisation system that can track the pose of the robot when it starts from a know position.
• You can command the robot to any position in the map. Path planning can be based on the given map and thus the robot may not succeed in reaching the position. If the path is free your robot should be able to reach the position.
• You can detect and identify at least one object and mark it in the map.
• You can pick up a detected object.
• You can record an evidence rosbag with camera images of the detected/identified objects and their positions following the specification in this wiki page.
• Your robot speaks every time the robot detects an object and explain what it sees. 

Milestone 4 (MS4)

In addition to MS1, MS2 and MS3, meeting this milestone means that:

• You should be able to start from an unknown position and localize successfully on the map.
• You should have a strategy for handling localisation errors resulting from for example rubbing up against a wall.
• You should be able to find a good path to the gate and get out, clearly showing that the robot is able to use information that was gathered via mapping the environment (rubble, objects, etc) when it plans the path.