当前位置: X-MOL 学术Soft Comput. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Optimization of stability of humanoid robot NAO using ant colony optimization tuned MPC controller for uneven path
Soft Computing ( IF 4.1 ) Pub Date : 2021-01-05 , DOI: 10.1007/s00500-020-05515-1
Abhishek Kumar Kashyap , Dayal R. Parhi

The primary conventional method for simplifying legged robots’ complex walking dynamics involves using low-dimensional models such as the linear inverted pendulum model (LIPM). This paper emphasizes utilizing the LIPM plus flywheel model (LIPPFM) for analysis of the complete dynamic motion of the humanoid robot. Inclining toward a more realistic case, the model is improvised to remove the COM’s height constraint (center of mass) and consider the effect of the upper body part using the mass of the pendulum. Furthermore, the double support phase is being discussed in the locomotion phase of the humanoid robot. MPC (model predictive control) approach has been used in this paper, which is tuned with the ACO (ant colony optimization) technique. The desired trajectory, joint angles, has been imparted to the MPC, which provides the robot’s joint motion. This joint motion has been further transferred to ACO, optimizing the step adjustment and providing an expected trajectory to walk over an uneven surface. The simulation has been carried out in an uneven environment based on ACO tuned MPC controller, and further, it has been validated using real-time experiments on humanoid robot NAO. The controller shows a reasonable degree of efficiency in both the real NAO and simulated NAO with a deviation under 5%. The comparative study among various controller shows that the proposed controller lowers the peak overshoot and the settling time. In comparison with the previously developed controller, the deviation in roll angle and pitch angle justifies the selection of the proposed controller.



中文翻译:

基于蚁群优化的MPC控制器优化人形机器人NAO的稳定性

简化腿式机器人复杂步行动力学的主要常规方法包括使用低维模型,例如线性倒立摆模型(LIPM)。本文强调利用LIPM加飞轮模型(LIPPFM)来分析类人机器人的完整动态运动。朝着更现实的情况倾斜,该模型被简化以消除COM的高度约束(重心),并使用摆的质量来考虑上身部分的影响。此外,在人形机器人的运动阶段中正在讨论双支撑阶段。本文采用了MPC(模型预测控制)方法,并通过ACO(蚁群优化)技术对其进行了调整。所需的轨迹,关节角度已赋予MPC,它提供了机器人的关节运动。该关节运动已进一步传递给ACO,优化了步距调节并提供了在不平坦表面上行走的预期轨迹。基于ACO调整的MPC控制器,在不平坦的环境中进行了仿真,此外,还通过在类人机器人NAO上进行的实时实验对其进行了验证。控制器在实际NAO和模拟NAO中均显示合理程度的效率,偏差低于5%。各种控制器之间的比较研究表明,所提出的控制器降低了峰值过冲和稳定时间。与以前开发的控制器相比,侧倾角和俯仰角的偏差证明了所建议控制器的选择是合理的。优化步进调整并提供在不平坦表面上行走的预期轨迹。基于ACO调整的MPC控制器,在不平坦的环境中进行了仿真,此外,还通过在类人机器人NAO上进行的实时实验对其进行了验证。控制器在实际NAO和模拟NAO中均显示出合理的效率水平,偏差在5%以下。各种控制器之间的比较研究表明,所提出的控制器降低了峰值过冲和稳定时间。与以前开发的控制器相比,侧倾角和俯仰角的偏差证明了所建议控制器的选择是合理的。优化步进调整并提供在不平坦表面上行走的预期轨迹。基于ACO调整的MPC控制器,在不平坦的环境中进行了仿真,此外,还通过在类人机器人NAO上进行的实时实验对其进行了验证。控制器在实际NAO和模拟NAO中均显示出合理的效率水平,偏差在5%以下。各种控制器之间的比较研究表明,所提出的控制器降低了峰值过冲和稳定时间。与以前开发的控制器相比,侧倾角和俯仰角的偏差证明了所建议控制器的选择是合理的。它已通过人形机器人NAO上的实时实验进行了验证。控制器在实际NAO和模拟NAO中均显示出合理的效率水平,偏差在5%以下。各种控制器之间的比较研究表明,所提出的控制器降低了峰值过冲和稳定时间。与以前开发的控制器相比,侧倾角和俯仰角的偏差证明了所建议控制器的选择是合理的。它已通过人形机器人NAO上的实时实验进行了验证。控制器在实际NAO和模拟NAO中均显示出合理的效率水平,偏差在5%以下。各种控制器之间的比较研究表明,所提出的控制器降低了峰值过冲和稳定时间。与以前开发的控制器相比,侧倾角和俯仰角的偏差证明了所建议控制器的选择是合理的。

更新日期:2021-01-05
down
wechat
bug