Uploaded March 2018 | Updated September 2026, 1 hour ago
During Pepper World Paris 2017, some of our Certified Partners were introducing a broad range of B2B applications and solutions developed for Pepper. These applications are answering several business needs in Retail, Banking, Hospitality or Healthcare. Discover them on video now.
During Pepper World Paris 2017, some of our Certified Partners were introducing a broad range of B2B applications and solutions developed for Pepper. These applications are answering several business needs in Retail, Banking, Hospitality or Healthcare. Discover them on video now.
![[Atalk] Oliver Brock - Variability : The main obstacle for general robotics? 1/2](https://i.ytimg.com/vi/iIXXcNne9ho/mqdefault.jpg "[Atalk] Oliver Brock - Variability : The main obstacle for general robotics? 1/2
Abstract:
The biggest success story of robotics remains factory automation. But our field is struggling to repeat this success in any other domain, especially in domains we call unstructured. Among the main culprits for this difficulty have been uncertainty and the high-dimensionality of configuration spaces associated with versatile robots. In my talk, I will argue that our field has found successful solutions for uncertainty and high-dimensionality, as long as they pertain to the robot itself. I will argue that after alleviating these two problems, a new problem has become the next bottleneck: the variability of the environment. On our path towards general robotics, we must focus on handling this variability. This might seem completely obvious, but I believe the most important implications of this realization have been largely ignored. I will present some attempts to address the variability in the environment, including interactive perception and learning with robotic priors. I will also place these solutions in the context of other attempts to address variability, such as big data and deep learning.
Bio:
Oliver Brock is the Alexander-von-Humboldt Professor of Robotics in the School of Electrical Engineering and Computer Science at the Technische Universität Berlin in Germany. He received his Diploma in Computer Science in 1993 from the Technische Universität Berlin and his Masters and Ph.D. in Computer Science from Stanford University in 1994 and 2000, respectively. He also held post-doctoral positions at Rice University and Stanford University. Starting in 2002, he was an Assistant Professor and Associate Professor in the Department of Computer Science at the University of Massachusetts Amherst, before to moving back to the Technische Universität Berlin in 2009. The research of Brocks lab, the Robotics and Biology Laboratory, focuses on autonomous mobile manipulation, interactive perception, grasping, manipulation, soft hands, interactive learning, motion generation, and the application of algorithms and concepts from robotics to computational problems in structural molecular biology. He is also the president of the Robotics: Science and Systems foundation.")
![[AI Lab] - Pepper learning Velcro Darts](https://i.ytimg.com/vi/i_JBRojCqcc/mqdefault.jpg "[AI Lab] - Pepper learning Velcro Darts
After our previous video showing how Pepper the Robot learns to play the ball-in-a-cup game (https://www.youtube.com/watch?v=jkaRO8J_1XI), realized by the AI Lab of SoftBank Robotics, here we see how Pepper learns to play Velcro Darts in the same way.
The movement is first demonstrated to the robot by guiding its arm. From there, Pepper has to improve its performance through trial-and-error learning. Even though the initial demonstration does not hit the center of the dartboard, Pepper can still learn to play the game successfully.
After 60 trials, Pepper scores 100 points, most of the times!
The movement is represented as a so-called dynamic movement primitive and optimized using an evolutionary algorithm. Our implementation uses the freely available software library dmpbbo: https://github.com/stulp/dmpbbo.")
![[AI Lab] Pepper robot learning ball in a cup](https://i.ytimg.com/vi/jkaRO8J_1XI/mqdefault.jpg "[AI Lab] Pepper robot learning ball in a cup
This video realized by the AI Lab of SoftBank Robotics shows how Pepper robot learns to play the ball-in-a-cup game (bilboquet in French). The movement is first demonstrated to the robot by guiding its arm.
From there, Pepper has to improve its performance through trial-and-error learning. Even though the initial demonstration does not land the ball in the cup, Pepper can still learn to play the game successfully.
The movement is represented as a so-called dynamic movement primitive and optimized using an evolutionary algorithm. Our implementation uses the freely available software library dmpbbo: https://github.com/stulp/dmpbbo.
After 100 trials, Pepper has successfully optimized its behavior and is able to repeatedly land the ball in the cup.")
