Design and Implementation of Intelligent and Autonomous AI Operated Humanoid Robot (DINKINESH)

Abstract

The premature possibilities of the impression of machine intelligence systematized do not contain the partition of the emerging field of Artificial Intelligence. The organizers of AI proposed the concept of embedded intelligence as being adjoined between perception, reasoning and actuation. Until now the fields of AI and Robotics drifted apart. Professionals of AI focused on problems and algorithms preoccupied from the real world. Roboticists, generally with a background in mechanical and electrical engineering, focused sensory motor functions. That departure was gradually being joined with the advancement of both fields and with the rising interest in the autonomous systems. The design and application of intelligent and autonomous AI operated humanoid robot (Dinkinesh) with the design and implementation of bipedal locomotion motion and balancing strategy for humanoid robot, path planning, design and integrated artificial intelligence. In the locomotion and balancing strategies, the dynamic modeling in the inverse kinematics problem with robot frame singularity, this can be improved by closed form inverse kinematics, with two stability measurements like, ZMP and COM based on MPC controller. In the conditions when the biped-robot operate in some useful operation, the path trajectory in accordance with the environments from the initial point to the goal point for those task the model predictive control with simultaneous gait pattern generation and path planning was implemented and obtained successfully. In the case of unstructured environment in the conditions of what a controller not understand what to do in known and unknown environments through the method like deep reinforcement learning, with the agent deep deterministic policy gradient (DDPG) which was just general artificial intelligence. The report introduces a method of implementing of intelligent and AI-operated humanoid robots Dinkinesh with a several capabilities of design parameters, tools, and techniques was used to integrate the robotic software architectures with its hardware system that allow a humanoid robot Dinkinesh to become more autonomous, intelligent and AI-powered. To cope with this issue the global perspective on main characteristics, design choices and constraints was briefly explained and the methods are utilized and the required output was discussed based on the hardware design requirements. The works had used their own the hardware design method from the scratch and all the material used to design the Dinkinesh biped humanoid robot was tested with postural stability and inverse kinematics problem was tested accordingly and gives good results.