Humanoids_2013_Logo

Dates & Venue

Atlanta (USA), IEEE-RAS HUMANOIDS 2013 Conference

Sept 10 – Deadline for submission of two-page abstract (EXTENDED)
Sept 15 – Notification of acceptance
Oct 15 – Workshop day

Website: http://www.h2rproject.eu/humanoids2013

Motivation & Goals

The difficulty in defining standard benchmarks for human-like robots is an acknowledged problem. Benchmarking research is inherently difficult since results are typically reported only for a specific robotic system and a self-chosen set of tasks. This makes it very difficult to compare the results with other systems developed in different labs and tested for different specific tasks. Moreover, there is no clear common view on what benchmarking is and how one should evaluate a system against a particular benchmark.

The recently started EU project H2R “Integrative approach for human-like locomotion” (ww.h2rproject.eu) aims to foster the international discussion on benchmarking scheme for bipedal robots, specifically focused on locomotion and posture. The ultimate goal of the project is to define solutions that can be realistically adopted by the scientific community to assess and compare human-like skills of humanoids and walking machines.

If you are a researcher in robotics, gait analysis, musculoskeletal modelling, or neuroscience, and you are interested in answering the questions “What does human-like mean?” and “How can we measure human-likeness?” please participate. We look forward to sharing new ideas with you!

Important Dates

Sept 10 – Deadline for submission of two-page abstract (EXTENDED)
Sept 15 – Notification of acceptance
Oct 15 – Workshop day (half or full day, depending on the number of contributions)

Topics & Discussion

This workshop will cover several aspects related to the assessment of human-like walking and postural skills:

1) Stability. How can human-like stability be effectively measured and described, in both unperturbed and perturbed conditions?

2) Energy efficiency. What method of energy consumption estimation can be better used across different robotic platforms? Where does the trade-off between energy efficiency and functionality lie?

3) Cognitive abilities. How much are prediction and estimation of disturbances relevant to achieve human-like walking and posture? Which features should be included in the ideal benchmarking scheme?

4) Kinematic complexity. To what extent, and in what conditions, is high-DOF better than low-DOF?

5) Compliant actuation. How can compliance be measured in humans and robots? Is the co-contraction ability a useful feature to be measured?

6) Learning. Can machine learning improve the assessment of human-like features?

7) Qualitative vs quantitative. To what extent quantitative measures really improve the assessment of human-like features?

8) Simulation. Which are the potential and limitations of simulation tools compared to real-life robotic benchmarking?

9) Clinical translational potential. How much can we learn for clinical assessment of human gait? And vice versa, how much can clinical scenarios take advantage of methods used in robotic benchmarking?

10) Challenges and competitions. The ideal benchmarking scheme is made of fast, simple, repeatable and reliable procedures that, most importantly, should be applied across different bipedal machines. How far are we to this ambitious goal? To this aim, what kind of competitions can be envisioned? What can we learn from existing competitions?

To encourage the discussion, oral presentations will be interspersed with activities to stimulate the debate between the attendees. We are open to include any other topic or aspect that may be relevant. Please send us a list of questions or new topics, we will be happy to prepare a debate on your inputs.

 

About the author:

Neural Rehabilitation Group, Cajal Institute
Spanish National Research Council (CSIC), Spain

Twitter LinkedIn Vimeo 

Workshop “Benchmarking Human-like robotic locomotion” @ Humanoids 2013

Leave a Reply

Related Posts