Gait optimization for roombots modular robots &amp;#x2014; Matching simulation and reality

Moeckel, Rico; Perov, Yura; Vespignani, Massimo; Bonardi, Stephane; Pouya, Soha; Sproewitz, Alexander; Kieboom, Jesse van den; Wilhelm, Frédéric; Ijspeert, Auke Jan

doi:10.1109/iros.2013.6696820

Cited by 18 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Approaches based on CPGs have been successfully applied in several evolutionary robotics and learning contexts, such as for bipedal walking (Reil and Husbands, 2002), salamanderlike robots (Crespi and Ijspeert, 2008), and modular robots (Sproewitz et al, 2008;Moeckel et al, 2013). Clune et al (2011) used controllers based on more general artificial neural networks to develop controllers for efficient locomotion.…”

Section: Locomotion Controller Approachesmentioning

confidence: 99%

See 1 more Smart Citation

Lamarckian Evolution of Simulated Modular Robots

et al. 2019

View full text Add to dashboard Cite

We study evolutionary robot systems where not only the robot brains but also the robot bodies are evolvable. Such systems need to include a learning period right after 'birth' to acquire a controller that fits the newly created body. In this paper we investigate the possibility of bootstrapping infant robot learning through employing Lamarckian inheritance of parental controllers. In our system controllers are encoded by a combination of a morphology dependent component, a Central Pattern Generator (CPG), and a morphology independent part, a Compositional Pattern Producing Network (CPPN). This makes it possible to transfer the CPPN part of controllers between different morphologies and to create a Lamarckian system. We conduct experiments with simulated modular robots whose fitness is determined by the speed of locomotion, establish the benefits of inheriting optimized parental controllers, shed light on the conditions that influence these benefits, and observe that changing the way controllers are evolved also impacts the evolved morphologies.

show abstract

Section: Locomotion Controller Approachesmentioning

confidence: 99%

“…Approaches based on CPGs have been successfully applied in several evolutionary robotics and learning contexts, such as for bipedal walking (Reil and Husbands, 2002 ), salamander-like robots (Crespi and Ijspeert, 2008 ), and modular robots (Sproewitz et al, 2008 ; Moeckel et al, 2013 ).…”

Section: Related Workmentioning

confidence: 99%

Lamarckian Evolution of Simulated Modular Robots

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The proposed solutions range from adding noise to the simulation [38], [40] to automatically improving simulators [41], [11]. One of the most successful approaches is the "transferability approach" [16], [15], [39]: instead of attempting to correct the simulator, the transferability approach hypothesizes that the simulator is accurate for some behaviors and not others; it is therefore possible to (1) learn the limits of the simulation, which is a supervised learning problem, and (2) encourage the learning process to select behaviors that are within these limits.…”

Section: Transferability In Robot Learningmentioning

confidence: 99%

Trial-and-error learning of repulsors for humanoid QP-based whole-body control

Spitz

Bouyarmane²,

Ivaldi

et al. 2017

2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)

View full text Add to dashboard Cite

Abstract-Whole body controllers based on quadratic programming allow humanoid robots to achieve complex motions. However, they rely on the assumption that the model perfectly captures the dynamics of the robot and its environment, whereas even the most accurate models are never perfect. In this paper, we introduce a trial-and-error learning algorithm that allows whole-body controllers to operate in spite of inaccurate models, without needing to update these models. The main idea is to encourage the controller to perform the task differently after each trial by introducing repulsors in the quadratic program cost function. We demonstrate our algorithm on (1) a simple 2D case and (2) a simulated iCub robot for which the model used by the controller and the one used in simulation do not match.

show abstract

“…The Roombots system (MOECKEL et al, 2013;BONARDI et al, 2013;VESPIGNANI et al, 2013;SPRöWITZ et al, 2014) (shown in Figure 2.9), designed by the Biorobotics Laboratory at École Polytechnique Fédérale de Lausanne (Switzerland), focus on building blocks for adaptive pieces of furniture able to move, self-assemble and self-reconfigure. Two half-spheres linked together with revolute joints forms a 110 mm diameter module weighting 1.4 kg.…”

Section: Roombotsmentioning

confidence: 99%

“…However, there are no additional sensors for sensing the alignment of a module with the grid. Figure 2.9: Roombots (MOECKEL et al, 2013;BONARDI et al, 2013;VESPIGNANI et al, 2013;SPRöWITZ et al, 2014).…”

Section: Roombotsmentioning

confidence: 99%

Sensors in modular robotics for pipeline inspection : design and test of erekobot- σ module

Sousa¹

View full text Add to dashboard Cite

Here, I am using the opportunity to express my gratitude to everyone who supported me throughout these years. I am thankful for their aspiring guidance, constructive criticism and friendly advices during my project, not only in the academic sense, but also in a personal sense. I am sincerely grateful to them for sharing their truthful and illuminating views on a number of issues related to the project. I express my warm thanks to my academic advisors Dr. Carla Koike and Dr. Dianne Magalhães for their expertise, understanding, and patience in the Ereko Group, they provided me with the required facilities and conductive condictions for my dissertation project. I appreciate their vast knowledge and skill in many areas (e.g., robotics, computer science, electronics, mechanics, interaction with participants and contributors), and their assistance in writing reports (i.e., proceedings articles and this dissertation). I also want to express my gratitude to all participants of the Ereko Group, especially Ricardo, whose volunteer work was fundamental for this project. I would also like to thank the other members of the chair, Dr. Carlos Humberto Llanos Quintero and Dr. Renato Alves Borges for the careful evaluation of my project. I must also acknowledge the Brazilian institutions: Agência Nacional do Petróleo, Gás Natural e Biocombustíveis (ANP), Financiadora de Estudos e Projetos (FINEP), Ministério da Ciência e Tecnologia (MCT), and Petróleo Brasileiro S.A. (Petrobras) for financial support in the program PRH-PB 223, especially in the present study. I would also like to thank the coordinator of the program Dr. Eugênio Liborio Feitosa Fortaleza for providing me support and knowledge. I also express my eternal gratitude to the University of Brasilia (UnB) and the Department of Mechanical Engineer (ENM); I will always consider these places home. Finally, I would also like to thank my parents for the support they provided me though my entire life, specially my time in UnB. Their constant love and support were fundamental in order to finish this dissertation, I love you guys. I must acknowledge my best friend and love, Patricia, without whose love, encouragement and editing assistance (mainly spoken words, not only written), I would not have finished this dissertation. Trust me, you have made me a better person. I would also like to thank Ethel,

show abstract

Gait optimization for roombots modular robots — Matching simulation and reality

Cited by 18 publications

References 16 publications

Lamarckian Evolution of Simulated Modular Robots

Lamarckian Evolution of Simulated Modular Robots

Trial-and-error learning of repulsors for humanoid QP-based whole-body control

Sensors in modular robotics for pipeline inspection : design and test of erekobot- σ module

Contact Info

Product

Resources

About