Team of researchers are putting collective efforts in creating strategies in order to test the physical capabilities of human robots.
By using a combination of algorithms and machine learning, the team has developed test robots which are capable of reacting to unknown changes within a simulated environment. This experiment has thus also helped in improving the robots odds of functioning in the real world.
Human robots can be defined as machines that resemble the human body shape in build and design. These humanoid robots have similar human attributes such as a torso, the head, two arms and feet. Moreover, they can also communicate with other robots and humans. These humanoid machines possess sensors, in addition to other input devices and can perform limited activities based on the outside input.
According to experts, these robots are pre-programmed for certain activities, for which they rely upon two learning methods, such as model-free and model-based. In the model-based method, the robot is taught a set of models which it uses to behave in a particular scenario; the model-free method on the other hand, does not teach the robot.
In order to make the robot ready to face a real-world scenario, which often changes unpredictably and constantly, research suggests, paradigm of each method has not proven to be sufficient. In order to overcome this difficulty, experts developed a new learning structure which incorporated both parts of the model-based and model-free learning methods in order to balance a two-legged robot.
The changes thus made helped in reducing the gap between the two learning paradigms. As a result, the transition between learning the model and learning the procedure could be effectively completed.
Based on the new simulation results, the proposed algorithm was able to stabilize the robot on a moving platform especially under unknown rotations. The rotations thus demonstrated that the robot was able to adapt to various unpredictable situations which could also be implemented to robots functioning outside the laboratory environment.
Owning to the current success, researchers hope to expand their work in developing more complex environments involving more changing variables and other unpredictable scenarios for the machines to adapt to. This could be further broadly used in intensifying the work of humanoid robots in the field of education, entertainment, disaster response and healthcare.
“Our ultimate goal will be to see how our method enables the robot to have control over its entire body as it is exposed to unmeasurable and unexpected disturbances such as a changing terrain. We would also like to see the robot’s ability to learn how to imitate human motion, such as ankle joint movement, without having been given prior information.”
The present study was published in Journal of Automatica Sinica.