بحث بعنوان Design and Modeling of a Lightweight and Low Power Consumption Full-Scale Biped Robot

بحث بعنوان Design and Modeling of a Lightweight and Low Power Consumption Full-Scale Biped Robot
اسم المؤلف
Michele Folgheraiter , and Bauyrzhan Aubakir
التاريخ
المشاهدات
521
التقييم
(لا توجد تقييمات)
Loading...
التحميل

بحث بعنوان
Design and Modeling of a Lightweight and Low Power Consumption Full-Scale Biped Robot
Michele Folgheraiter , and Bauyrzhan Aubakir
Robotics and Mechatronics Department,
School of Science and Technology,
Nazarbayev University,
53 Kabanbay Batyr Ave., Astana, Kazakhstan
[email protected]
[email protected]
Received 6 June 2018
Revised 20 July 2018
Accepted 31 July 2018
Published 6 September 2018
This paper introduces the design methodology, the modeling and the power consumption tests
for a newly developed biped robot equipped with 12 DOFs. The robot is 1.1 meters tall
(lower limbs) which makes it comparable in dimension with other state-of-the-art full-scale
humanoids. By using a combination of 3D printing techniques and lightweight materials, the
system weighs only 10.8 kg (without batteries) while retaining high links strength and rigidity.
Without compromising the workspace dimension, the robot presents a very low weight-toheight ratio (9.8 kg/m) that translates into a safer operation and reduced energy consumption.
To perform elementary locomotion primitives, e.g., changing the support from one foot to the
other or lifting its body, the robot prototype consumes only 65 watts. Simulation results
demonstrate the suitability of the robot’s kinematics to perform walking motion and predict an
average power consumption of 200 watts. The direct kinematics of the robot is presented
together with its inverse dynamics based on a Chaotic Recurrent Neural Network (CRNN). The
adaptive model is identi¯ed using a recursive least squares algorithm that allows the CRNN to
predict the torques at di®erent step lengths with a MSE of 0.0057 on normalized data.
Keywords: Humanoid robotics; biped robot; kinematic design; 3d printing; dynamic modeling;
recurrent neural networks.

  1. Introduction
    Humanoid robotics is a relatively new research ¯eld which aims at developing
    anthropomorphic robotics systems intended to be used in public and household
    ‡Corresponding author.
    This is an Open Access article published by World Scienti¯c Publishing Company. It is distributed under
    the terms of the Creative Commons Attribution 4.0 (CC-BY) License. Further distribution of this work is
    permitted, provided the original work is properly cited.
    International Journal of Humanoid Robotics
    Vol. 15, No. 5 (2018) 1850022 (32 pages)
    °c The Author(s)
    DOI: 10.1142/S0219843618500226
    1850022-1
    Int. J. Human. Robot. 2018.15.
    Conclusion
    In this paper, we introduce the design and the model of a newly developed lightweight and full-size biped robot intended for application in household and public
    environments. The robot presents a total of 12 rotational joints, 6 for each leg, that
    allow the system to perform static and dynamic walking. The ¯rst three joints in the
    hip have axes that intersect in a common point such that the inverse kinematics
    admits a closed-form solution. The pitch joint of the ankle is actuated by a servomotor located in the upper part of the lower leg through an elastic synchronous belt.
    This has the e®ect to reduce the e®ective sti®ness of the joint and absorb impact
    forces while the robot is walking. Four force sensors are integrated in each foot to
    calculate the center of pressure while standing or walking. The robot is equipped with
    on-board computational units that are capable to implement low-level control
    strategies, and that allow to control the robot remotely by using a graphical user
    interface or a console. By using a combination of 3D printing techniques and lightweight materials, we could design a robot which is only 10.8 kg and of comparable
    dimensions with other state-of-the-art full size humanoid robots. By having a weightto-length ratio of only 9.8 kg/m, the system is inherently safe while interacting with
    humans. Furthermore, with a low links’ weight and inertia, the average power
    consumption is estimated by simulation to be 200 W, while performing a static gait.
    Experiments conducted on the real prototype con¯rm this ¯gure, thus the robot
    absorbs 45 W to move its weight from the center to the left or right foot and 65 W to
    perform a squatting motion primitive. The direct kinematics of the robot is introduced together with its inverse dynamic model based on a CRNN. By using the RLS
    algorithm, joint torques can be predicted from instantaneous joint positions at different step lengths with a MSE of 0.0057.
    To perform balance and gait control, there are di®erent algorithms and techniques available in literature. Many of them are model-based and verify the stability of
    the robot by using the concept of Zero Moment Point (ZMP).41–43 Others rely on the
    estimation of the Center of Pressure (COP) by using the data from sensors installed
    under the feet.44–46 In Ref. 47 a motion primitives switching methodology is introduced
    that provides a more e±cient posture control to compensate for di®erent disturbances
    patterns. The dynamic model of the robot together with the information of the COP
    Design and Modeling of Full-Scale Biped Robot
    1850022-27
    Int. J. Human. Robot. 2018.15. Downloaded from www.worldscientific.com
    by 217.54.46.56 on 06/29/22. Re-use and distribution is strictly not permitted, except for Open Access articles.and the torques limits are used as constraints for a quadratic program algorithm in
    order to ¯nd the optimal joint trajectories. In Ref. 48 a passivity-based control method
    is presented, which using minimal actuation is capable to perform stable gaits for a
    biped equipped with compliant ankle’s joints. By reducing at the minimum, the active
    control phases to compensate for the e®ect of gravity and by employing the intrinsic
    oscillatory dynamics of the biped, it was possible to minimize the energy consumption
    and to walk on arbitrary slopes.
    In comparison with Ref. 47, where the motion primitives are the outcome of an
    o®line optimization process, our methodology to control the robot will be based on an
    adaptive model that can self-adjust during the robot operation. It will be possible to
    introduce an additional input to the CRNN that will consider the disturbance force
    estimation in order to adapt the gait accordingly. As we had already demonstrated
    the capability of a single CRNN to generate gates with di®erent step lengths, it will
    also be possible to adjust the gate to compensate for the di®erent disturbances
    present in the robot environment. If in Ref. 47, the primitives are recorded as separated models, by using a CRNN of su±cient complexity, it will be possible to
    generate di®erent robot behaviors by using a single neural network. Furthermore,
    thanks to the generalization properties of a CRNN, switching between di®erent gates
    will be performed smoothly. In comparison with Ref. 48, where an underactuated
    biped is considered that integrates a 3 DOFs compliant ankle, our biped is instead
    fully actuated and includes only one compliant DOF in the ankle. Our solution will
    therefore require more energy. However, on the other hand, it will also allow a better
    controllability of the robot. This is particularly important when the robot is supposed to move in a household environment, where di®erent static and dynamic
    obstacles are present.
    As a future work, we intend to do extensive testing of the robot prototype while
    performing di®erent gaits in order to measure the energy consumption. It will be
    necessary to install additional sensors like IMUs and depth cameras that will allow to
    implement balancing motion primitives and identify and avoid obstacles while
    walking. Furthermore, we plan to use the developed adaptive inverse dynamics
    to implement a model-based feedback control system able to compensate for nonlinearities and changing dynamic conditions while the robot is operating.
    Acknowledgment
    This work was supported by the Ministry of Education and Science of the Republic of
    Kazakhstan under the grant and target funding scheme agreement #328/239-2017
    and by Nazarbayev University under the Faculty Development Competitive
    Research Grants Program award #090118FD5343.
    M. Folgheraiter & B. Aubakir
    1850022-28
    Int. J. Human. Robot. 2018.15.

كلمة سر فك الضغط : books-world.net
The Unzip Password : books-world.net

تحميل

يجب عليك التسجيل في الموقع لكي تتمكن من التحميل

تسجيل | تسجيل الدخول