Citation
Share
Abstract
Humanoid robots have been researched during the last few years because of the useful applications they can be implemented in the future. Some of these include medicine, education, military, industry, and support areas. A humanoid can be used to rescue people and animals from natural disasters, be a support in a production line, teach students, form part of military defense, and many other applications. Biped robots are known as the lower part of the humanoid robot and the principal characteristic is that it has 2 feet. To generate movements in a biped robot requires following specific trajectories in all actuators. A control algorithm is responsible to follow the specific references in the DC motors individually. Some basic motions for a biped robot are walking and squats because are the principal movements to travel and lift heavy things. Additionally, the analysis of the implementation of control algorithms to individual actuators for biped robot applications is not completely studied. This thesis proposes the development of an embedded system to apply controllers and the implementation of a second-order Super Twisting Sliding Mode control (STSMC) in the actuators for a constructed biped robot with 12 degrees of freedom actuated by 12 DC motors and an individual comparison with a PID (Proportional, Integrative, and Derivative control) controller. First, the DC motor model is identified using the least square method. Then, the model is validated to ensure representation in simulation. Next, the Sliding Mode control (SMC) is analyzed to evaluate the functional theory in simulation. Additionally, the STSMC is simulated and then implemented in one DC motor. Moreover, to generate walking and squatting patterns the model of the three-dimensional inverted pendulum is applied using the stability criterion. The inverse kinematics provides the joint angles to generate the trajectories. Finally, the mean squared error (MSE) is implemented to measure the effectiveness of the control algorithms. The results show that the PID and STSMC controls have an MSE average of 0.4704 and 0.0228 respectively when replicating the squat trajectory. Meaning an improvement of 1961%. The maximum error with the STSMC was was 0.5° when the joint reached the change of direction because of the inertial force. This error represents an error of 6.71mm at the end of the foot. This error is inside the allowed limits, the robot can reproduce the squat without falling. Finally, this research can be applied to manipulator robots that use DC motors to control the position.
Description
https://orcid.org/0000-0003-2559-539X