Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555678
R. Iizuka, D. Nenchev, D. Sato
A method for distribution of the desired angular momentum of the robot among the body segments is described. The optimization problem is solved with the help of weights that are proportional to the moments of inertia of the body segments. The method is implemented with the Relative Angular Acceleration (RAA) controller, whereby emergent movements are generated in the pelvis, torso and arms. It is shown that these movements yield a human-like performance of highly dynamic tasks such as flips, long jumps and jumps with twist. The validity of the method and its generality as well as the robustness of the control approach were confirmed with simulations.
{"title":"Motion Generation and Control of Acrobatic Motion Synergies Emerging From the Momentum Equilibrium Principle","authors":"R. Iizuka, D. Nenchev, D. Sato","doi":"10.1109/HUMANOIDS47582.2021.9555678","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555678","url":null,"abstract":"A method for distribution of the desired angular momentum of the robot among the body segments is described. The optimization problem is solved with the help of weights that are proportional to the moments of inertia of the body segments. The method is implemented with the Relative Angular Acceleration (RAA) controller, whereby emergent movements are generated in the pelvis, torso and arms. It is shown that these movements yield a human-like performance of highly dynamic tasks such as flips, long jumps and jumps with twist. The validity of the method and its generality as well as the robustness of the control approach were confirmed with simulations.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117345320","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555771
Daniel García-Vaglio, Javier Peralta-Sáenz, Federico Ruiz-Ugalde
Competent assistant robots need to understand their environment to be able to perform skilled manipulation tasks. One way of achieving it is to have knowledge about the physical behaviour of objects and how they respond to certain stimuli. The proposed approach is to have a library of compact predictor-controller pairs for different simple tasks that chained together enable the robot to execute complex actions. In this paper we expand the previous work by adding the compact models for pushing a cylinder over a table to a goal pose. The model was tested and compared against our previous box model with a real humanoid robot by pushing two cylinder shaped and a box shaped object from random starting positions to predefined goals. The robot successfully pushed the objects towards the goals. With this, as the main contribution fo this paper, we have successfully expanded the systems presented on previous works with new capabilities — namely the ability to work with cylinder shaped objects.
{"title":"Pushing cylinders: Expanding on object-based manipulation","authors":"Daniel García-Vaglio, Javier Peralta-Sáenz, Federico Ruiz-Ugalde","doi":"10.1109/HUMANOIDS47582.2021.9555771","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555771","url":null,"abstract":"Competent assistant robots need to understand their environment to be able to perform skilled manipulation tasks. One way of achieving it is to have knowledge about the physical behaviour of objects and how they respond to certain stimuli. The proposed approach is to have a library of compact predictor-controller pairs for different simple tasks that chained together enable the robot to execute complex actions. In this paper we expand the previous work by adding the compact models for pushing a cylinder over a table to a goal pose. The model was tested and compared against our previous box model with a real humanoid robot by pushing two cylinder shaped and a box shaped object from random starting positions to predefined goals. The robot successfully pushed the objects towards the goals. With this, as the main contribution fo this paper, we have successfully expanded the systems presented on previous works with new capabilities — namely the ability to work with cylinder shaped objects.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129670814","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555775
Gabriel García, Robert J. Griffin, J. Pratt
Recently we have seen a lot of progress done in dynamic locomotion with quadrupedal robots using the Single-Rigid Body Model, which contains simplified dynamics that considers the robot a single “potato”. This approach performs poorly when the robot contains heavy links, because those links take a considerable momentum to move and because they also change the overall inertia of the robot. In this paper, we generalize the SRBM using the Centroidal Dynamics model plus an orientation variable, whose dynamics contain the linearized effects of other links’ momentum and variable inertia. We are designing this Enhanced Centroidal Dynamics using the Full-Body Dynamics, so the trajectories we obtain are instantaneously dynamically feasible. We show our approach in a full-body dynamic simulation of the MIT Humanoid, a biped with line-feet contact, and we show a simplification in the modeling of the wrenches that can be applied with line-feet.
{"title":"MPC-based Locomotion Control of Bipedal Robots with Line-Feet Contact using Centroidal Dynamics.","authors":"Gabriel García, Robert J. Griffin, J. Pratt","doi":"10.1109/HUMANOIDS47582.2021.9555775","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555775","url":null,"abstract":"Recently we have seen a lot of progress done in dynamic locomotion with quadrupedal robots using the Single-Rigid Body Model, which contains simplified dynamics that considers the robot a single “potato”. This approach performs poorly when the robot contains heavy links, because those links take a considerable momentum to move and because they also change the overall inertia of the robot. In this paper, we generalize the SRBM using the Centroidal Dynamics model plus an orientation variable, whose dynamics contain the linearized effects of other links’ momentum and variable inertia. We are designing this Enhanced Centroidal Dynamics using the Full-Body Dynamics, so the trajectories we obtain are instantaneously dynamically feasible. We show our approach in a full-body dynamic simulation of the MIT Humanoid, a biped with line-feet contact, and we show a simplification in the modeling of the wrenches that can be applied with line-feet.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125338956","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555793
Johannes Tenhumberg, B. Bäuml
High absolute accuracy is an essential prerequisite for a humanoid robot to autonomously and robustly perform manipulation tasks while avoiding obstacles. We present for the first time a kinematic model for a humanoid upper body incorporating joint and transversal elasticities. These elasticities lead to significant deformations due to the robot’s own weight, and the resulting model is implicitly defined via a torque equilibrium. We successfully calibrate this model for DLR’s humanoid Agile Justin, including all Denavit–Hartenberg parameters and elasticities. The calibration is formulated as a combined least-squares problem with priors and based on measurements of the end effector positions of both arms via an external tracking system. The absolute position error is massively reduced from 21 mm to 3.1 mm on average in the whole workspace. Using this complex and implicit kinematic model in motion planning is challenging. We show that for optimization-based path planning, integrating the iterative solution of the implicit model into the optimization loop leads to an elegant and highly efficient solution. For mildly elastic robots like Agile Justin, there is no performance impact, and even for a simulated highly flexible robots with 20 times higher elasticities, the runtime increases by only 30%.
{"title":"Calibration of an Elastic Humanoid Upper Body and Efficient Compensation for Motion Planning","authors":"Johannes Tenhumberg, B. Bäuml","doi":"10.1109/HUMANOIDS47582.2021.9555793","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555793","url":null,"abstract":"High absolute accuracy is an essential prerequisite for a humanoid robot to autonomously and robustly perform manipulation tasks while avoiding obstacles. We present for the first time a kinematic model for a humanoid upper body incorporating joint and transversal elasticities. These elasticities lead to significant deformations due to the robot’s own weight, and the resulting model is implicitly defined via a torque equilibrium. We successfully calibrate this model for DLR’s humanoid Agile Justin, including all Denavit–Hartenberg parameters and elasticities. The calibration is formulated as a combined least-squares problem with priors and based on measurements of the end effector positions of both arms via an external tracking system. The absolute position error is massively reduced from 21 mm to 3.1 mm on average in the whole workspace. Using this complex and implicit kinematic model in motion planning is challenging. We show that for optimization-based path planning, integrating the iterative solution of the implicit model into the optimization loop leads to an elegant and highly efficient solution. For mildly elastic robots like Agile Justin, there is no performance impact, and even for a simulated highly flexible robots with 20 times higher elasticities, the runtime increases by only 30%.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"88 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126360707","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555777
Alexander Toedtheide, Johannes Kühn, Edmundo Pozo Fortunic, S. Haddadin
This paper presents a novel 2 degrees of freedom humanoid wrist with solely three tendons, based on an integrated 3S$overline{P}$S-1U parallel kinematics (the $overline{P}$ indicates the active degree of freedom) and driven by three electromechanical motors. Tendon-force measurement and control, combined with nonlinear kinematics mapping enable the implementation of a oint-level impedance controller, virtual walls, a momentum oberver and a virtual joint-torque sensor. The novel mechanical esign, especially the drive-train and the tendon force measurenent module, are discussed in the paper. Simulations show the easibility of the control methods. A static workspace analysis - eveals a configuration depending torque with a theoretical maximum torque of 6.1 Nm or 9.3 Nm (depending on the used gear ratio). All control modes as well as human interaction are validated experimentally. First experiments on a humanoid arm are shown. An experimental performance of a maximum speed of 400 deg/s, a maximum payload of 2.5 kg (lever arm 16 cm) and an accuracy of less than 0.1 deg (95% confidence interval) are achieved.
{"title":"An Integrated, Force-Sensitive, Impedance Controlled, Tendon-Driven Wrist: Design, Modeling, and Control","authors":"Alexander Toedtheide, Johannes Kühn, Edmundo Pozo Fortunic, S. Haddadin","doi":"10.1109/HUMANOIDS47582.2021.9555777","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555777","url":null,"abstract":"This paper presents a novel 2 degrees of freedom humanoid wrist with solely three tendons, based on an integrated 3S$overline{P}$S-1U parallel kinematics (the $overline{P}$ indicates the active degree of freedom) and driven by three electromechanical motors. Tendon-force measurement and control, combined with nonlinear kinematics mapping enable the implementation of a oint-level impedance controller, virtual walls, a momentum oberver and a virtual joint-torque sensor. The novel mechanical esign, especially the drive-train and the tendon force measurenent module, are discussed in the paper. Simulations show the easibility of the control methods. A static workspace analysis - eveals a configuration depending torque with a theoretical maximum torque of 6.1 Nm or 9.3 Nm (depending on the used gear ratio). All control modes as well as human interaction are validated experimentally. First experiments on a humanoid arm are shown. An experimental performance of a maximum speed of 400 deg/s, a maximum payload of 2.5 kg (lever arm 16 cm) and an accuracy of less than 0.1 deg (95% confidence interval) are achieved.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114579934","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper proposes a vertical jump optimization strategy for a one-legged robot with consideration of its variable reduction ratio joints. Firstly, the characteristic of the joint is derived to obtain its influence on jump motion, which is similar to the reduction ratio. Secondly, referring to the joint’s characteristic, the initial posture of jumping is optimized to maximize the initial acceleration of jumping. Then, to generate the trajectory of the center of mass (CoM) and make the jump motion more efficient, nonlinear optimization of CoM is adopted with respect to human jumping data. Full-body dynamics is considered to track the trajectory with virtual force control. For flight phase, joint PD controller is adopted to decelerate and maintain the posture. A contrast simulation is implemented to demonstrate the characteristics of the variable reduction ratio joint. Vertical jump experiment on a one-legged robot platform is realized with a height of 30 cm.
{"title":"A vertical jump optimization strategy for one-legged robot with variable reduction ratio joint","authors":"Haoxiang Qi, Xuechao Chen, Zhangguo Yu, Gao Huang, Libo Meng, K. Hashimoto, Wen-Xiong Liao, Qiang Huang","doi":"10.1109/HUMANOIDS47582.2021.9555679","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555679","url":null,"abstract":"This paper proposes a vertical jump optimization strategy for a one-legged robot with consideration of its variable reduction ratio joints. Firstly, the characteristic of the joint is derived to obtain its influence on jump motion, which is similar to the reduction ratio. Secondly, referring to the joint’s characteristic, the initial posture of jumping is optimized to maximize the initial acceleration of jumping. Then, to generate the trajectory of the center of mass (CoM) and make the jump motion more efficient, nonlinear optimization of CoM is adopted with respect to human jumping data. Full-body dynamics is considered to track the trajectory with virtual force control. For flight phase, joint PD controller is adopted to decelerate and maintain the posture. A contrast simulation is implemented to demonstrate the characteristics of the variable reduction ratio joint. Vertical jump experiment on a one-legged robot platform is realized with a height of 30 cm.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121064714","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555685
Anna Lena Emonds, K. Mombaur
A comprehensive knowledge of the underlying criteria of sprint motions, both of non-amputee and below-knee amputee athletes, is helpful for design improvement of prosthetic devices or towards a fair judgment of (dis-)advantage due to the running-specific prosthesis. For the study of sprint motions with and without running-specific prostheses, we created rigid multi-body system models of three non-amputee and one unilateral transtibial amputee athlete. We restricted the motions to the sagittal plane, ending up with 16 degrees of freedom (DOFs). The internal rotational DOFs are controlled by joint torque actuators. As the prosthetic device has to be passive, the joint torque actuator is replaced by a linear spring-damper system in the prosthetic ankle joint. The aim of our study is to identify the optimal weight factors which combine five elementary optimization criteria in such a way that the resulting synthesized motions comes as close as possible to recorded reference motions. To this aim, we formulated an inverse optimal control problem (IOCP) as a bi-level problem: In the outer loop, the weight factors are adapted such that the comparison of the reference motion and the solutions of the optimal control problem (OCP) which computes sprint motions in the inner loop match each other as close as possible. In contrast to previous studies, we investigated more athletes, added subject-specific joint torque limits based on Muscle Torque Generators and left the average velocities of the running motions free. For all four athletes, we identified a set of optimal weights that generates sprint motions which closely match the recorded reference motion. Significant differences in the identified weights between amputee and non-amputee sprinting have been found. Especially angular momentum control plays a decisive role in unilateral transtibial amputee sprinting.
{"title":"Using Subject-Specific Models to find Differences in Underlying optimization Criteria of Sprinting with and without Prostheses","authors":"Anna Lena Emonds, K. Mombaur","doi":"10.1109/HUMANOIDS47582.2021.9555685","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555685","url":null,"abstract":"A comprehensive knowledge of the underlying criteria of sprint motions, both of non-amputee and below-knee amputee athletes, is helpful for design improvement of prosthetic devices or towards a fair judgment of (dis-)advantage due to the running-specific prosthesis. For the study of sprint motions with and without running-specific prostheses, we created rigid multi-body system models of three non-amputee and one unilateral transtibial amputee athlete. We restricted the motions to the sagittal plane, ending up with 16 degrees of freedom (DOFs). The internal rotational DOFs are controlled by joint torque actuators. As the prosthetic device has to be passive, the joint torque actuator is replaced by a linear spring-damper system in the prosthetic ankle joint. The aim of our study is to identify the optimal weight factors which combine five elementary optimization criteria in such a way that the resulting synthesized motions comes as close as possible to recorded reference motions. To this aim, we formulated an inverse optimal control problem (IOCP) as a bi-level problem: In the outer loop, the weight factors are adapted such that the comparison of the reference motion and the solutions of the optimal control problem (OCP) which computes sprint motions in the inner loop match each other as close as possible. In contrast to previous studies, we investigated more athletes, added subject-specific joint torque limits based on Muscle Torque Generators and left the average velocities of the running motions free. For all four athletes, we identified a set of optimal weights that generates sprint motions which closely match the recorded reference motion. Significant differences in the identified weights between amputee and non-amputee sprinting have been found. Especially angular momentum control plays a decisive role in unilateral transtibial amputee sprinting.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122632207","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555799
Kentaro Uno, Naomasa Takada, T. Okawara, Keigo Haji, Arthur Candalot, Warley F. R. Ribeiro, K. Nagaoka, Kazuya Yoshida
This study presents the design and sequential control strategies of a novel lightweight climbing robot. The quadruped robot with a left–right and front–hind symmetric insect-type configuration has three degrees of freedom (3-DOF) actuated joints in each limb, a 3-DOF passive compliant spine gripper at each foot, and an actuator to open/close the gripper. First, we present the mechanical design and minimal hardware integration of the robot, which have helped successfully reduce the entire mass of the robot to 3 kg with a base height of 0.16 m. Next, a sequential strategy to process stable climbing locomotion is introduced. The implemented software architecture that realizes climbing motion is described. With the successful result of a teleoperation experiment on an indoor test field simulating the Martian uneven slalom (local max. inclination: 45°), we proved that the proposed sequential control strategy enables the robot to stably climb challenging terrain.
{"title":"HubRobo: A Lightweight Multi-Limbed Climbing Robot for Exploration in Challenging Terrain","authors":"Kentaro Uno, Naomasa Takada, T. Okawara, Keigo Haji, Arthur Candalot, Warley F. R. Ribeiro, K. Nagaoka, Kazuya Yoshida","doi":"10.1109/HUMANOIDS47582.2021.9555799","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555799","url":null,"abstract":"This study presents the design and sequential control strategies of a novel lightweight climbing robot. The quadruped robot with a left–right and front–hind symmetric insect-type configuration has three degrees of freedom (3-DOF) actuated joints in each limb, a 3-DOF passive compliant spine gripper at each foot, and an actuator to open/close the gripper. First, we present the mechanical design and minimal hardware integration of the robot, which have helped successfully reduce the entire mass of the robot to 3 kg with a base height of 0.16 m. Next, a sequential strategy to process stable climbing locomotion is introduced. The implemented software architecture that realizes climbing motion is described. With the successful result of a teleoperation experiment on an indoor test field simulating the Martian uneven slalom (local max. inclination: 45°), we proved that the proposed sequential control strategy enables the robot to stably climb challenging terrain.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133077797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555808
Marc Bestmann, Jasper Güldenstein, Florian Vahl, Jianwei Zhang
We present our open humanoid robot platform Wolfgang. The described hardware focuses on four aspects. Firstly, the robustness against falls is improved by integrating 3D printed elastic elements. Additionally, a high control loop frequency is achieved by using new custom control electronics. Furthermore, a torsion spring is applied to reduce the torque on the knee joints. Finally, computational power is provided through the combination of different processors. The paper also presents the ROS-based software stack that is used in RoboCup.
{"title":"Wolfgang-OP: A Robust Humanoid Robot Platform for Research and Competitions","authors":"Marc Bestmann, Jasper Güldenstein, Florian Vahl, Jianwei Zhang","doi":"10.1109/HUMANOIDS47582.2021.9555808","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555808","url":null,"abstract":"We present our open humanoid robot platform Wolfgang. The described hardware focuses on four aspects. Firstly, the robustness against falls is improved by integrating 3D printed elastic elements. Additionally, a high control loop frequency is achieved by using new custom control electronics. Furthermore, a torsion spring is applied to reduce the torque on the knee joints. Finally, computational power is provided through the combination of different processors. The paper also presents the ROS-based software stack that is used in RoboCup.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123621786","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555784
Claudio S. Ravasio, F. Iida, A. Rosendo
Walking robots have been a thriving topic for years, and their impact in our future is undeniable. Through different walking techniques machines match their control parameters to environmental conditions, and in both simulation and real-world this control optimization always requires many iterations to find the best parameter. We benchmark four optimization methods and two fitness shaping methods to assess how fast a locomotion model, with two control parameters, can converge to stability. We find that a best overall solution does not exist, with inference-based methods such as Bayesian Optimization in some cases being as inefficient as Random Search. Fitness Shaping using additional information provided by the simulation after termination is shown to improve optimization speed in the presence of running gaits. Additionally, our results validate Bayesian optimization as the fastest optimization method for walking gaits, and present Neural Networks as the fastest for running gaits and. In the presence of so many methods and models, this comparative study aims to clarify the potential gains for optimization methods in bipedal locomotion.
{"title":"Fitness Shaping on SLIP Locomotion Optimization","authors":"Claudio S. Ravasio, F. Iida, A. Rosendo","doi":"10.1109/HUMANOIDS47582.2021.9555784","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555784","url":null,"abstract":"Walking robots have been a thriving topic for years, and their impact in our future is undeniable. Through different walking techniques machines match their control parameters to environmental conditions, and in both simulation and real-world this control optimization always requires many iterations to find the best parameter. We benchmark four optimization methods and two fitness shaping methods to assess how fast a locomotion model, with two control parameters, can converge to stability. We find that a best overall solution does not exist, with inference-based methods such as Bayesian Optimization in some cases being as inefficient as Random Search. Fitness Shaping using additional information provided by the simulation after termination is shown to improve optimization speed in the presence of running gaits. Additionally, our results validate Bayesian optimization as the fastest optimization method for walking gaits, and present Neural Networks as the fastest for running gaits and. In the presence of so many methods and models, this comparative study aims to clarify the potential gains for optimization methods in bipedal locomotion.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131819470","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：