Exercise: Robot planning and behavior trees

In this challenge you will work in groups of at most two members. The goal is to design a Behavior Tree (BT) that controls a robot to make it complete a mobile manipulation task. 

The task for the robot is to pick heavy and light objects from two conveyors and place them in the delivery area. The robot has a limit in the weight it can carry and has a battery that is drained continuously (see the image below). 

The whole environment for the challenge is already set up for you in a notebook to run in Google Colaboratory. 

There is a set of behaviors that you can use to build the BT. The BT that you design will then be evaluated in the environment and a fitness score will be assigned. To complete the challenge you have to submit your best scoring tree, that will be run in a robustness test and its final fitness score published in this document Links to an external site. and the winning team gets to publicly bask in the glory of their success as well as a somewhat more real reward. 

You will find the instructions to run the notebook and more detailed information about the task, the environment, the behaviors as well as how the fitness is computed in this GitHub repository Links to an external site..

Please do not post your solutions publicly so that the exercise can be used in other contexts.