This challenge focuses on assisting the elderly or impaired. Robots will be helping Grannie Annie - an elderly lady who needs some support to keep her healthy and doing the things she loves. Set in Granny Annie's apartment, tasks for the robot include identifying and reacting appropriately to four different visitors to the apartment, getting familiar with its surroundings and helping Granny Annie with everyday activities such as opening and closing the window blinds.
Below you will find a summary of the tasks, functionalities, platform and benchmarking. Full details on these items can be found in the RoCKIn@Home Rulebook + Nutshell document. You can read more about the RoCKIn@Home testbed here.
All necessary resources for participants can be found on the RoCKIn Wiki and the RoCKIn GitHub repository.
Here the robot will be required to generate a map of its environment and detect random changes in the location of furniture and other items that will be made before each task.
This task assesses the robot's capability to interact effectively with humans and to demonstrate different behaviors when dealing with known and unknown people.
The objective of this functionality benchmark is to assess the capabilities of a robot in processing sensor data in order to extract information about observed objects from a given list. It requires the robot to detect the object's presence, estimate their class, identity and location.
This functionality benchmark assesses the robot's capability to traverse a sequence of way points without colliding with obstacles and reaching its target within reasonable time.
This functionality benchmark aims at evaluating the ability of a robot to understand speech commands that a user gives in a home environment. A list of commands will be selected among the set of predefined recognizable commands (i.e. commands that the robot should be able to recognize within the tasks of the competition or in similar situations).
Teams can use either one or more fully autonomous robots to solve the tasks. Robots must conform to certain size, weight, safety restrictions and are wirelessly networked with other devices, such as motorised blinds + surveillance cameras. Teams are not allowed to modify the environment or add additional devices.
Benchmarking in RoCKIn has been designed so as to minimise subjective evaluations and to allow for the quality of sub-system components to be recognised as well as success in performing the overall task. Though the careful design and revision of the challenge rules we hope to allow for the comparison of performances across years.
Tasks will be benchmarked against whether they successfully complete the required actions (such as understanding commands, correctly operating and/or identifying objects and devices etc.) and how effectively they do so, with penalties being given for actions such as bumping into Granny Annie or furniture for RoCKIn@Home. Penalties or disqualification in RoCKIn@Work will be given for actions such as damaging objects or the testbed.