Evolution of prehension ability in an anthropomorphic neurorobotic arm

Abstract

In this paper, we show how a simulated anthropomorphic robotic arm controlled by an artificial neural network can develop effective reaching and grasping behaviour through a trial and error process in which the free parameters encode the control rules which regulate the fine-grained interaction between the robot and the environment and variations of the free parameters are retained or discarded on the basis of their effects at the level of the global behaviour exhibited by the robot situated in the environment. The obtained results demonstrate how the proposed methodology allows the robot to produce effective behaviours thanks to its ability to exploit the morphological properties of the robot's body (i.e. its anthropomorphic shape, the elastic properties of its muscle-like actuators and the compliance of its actuated joints) and the properties which arise from the physical interaction between the robot and the environment mediated by appropriate control rules.

Neuromodelling based on evolutionary robotics: on the importance of motor control for spatial attention

“…The only difference is that the real eye is controlled by three couple of muscles, whereas in our simplified model, we used two couples (i.e., inferior, superior, medial and lateral rectus). The muscles were simulated using the Hills model (Massera et al 2007(Massera et al , 2014Massera 2010). In a previous study (Gigliotta et al 2015), we demonstrated that the use of zoom actuator improves the performance of a neurorobot, controlled in velocity, in a simpler variant of cancellation task in which target stimuli disappeared as soon as they were cancelled.…”

The Facilitatory Role of Linguistic Instructions on Developing Manipulation Skills

“…For each condition, the evolutionary process has been have been simulated by using Newton Game Dynamics (NGD, see: www.newtondynamics.com), a library for accurately simulating rigid body dynamics and collisions. For related approaches, see [23], [22], [24].…”

“…This in turn allows the robot to exploit sensory-motor coordination (i.e., the possibility to act in order to later experience useful sensory states) as well as the properties arising from the physical interactions between the robot and the environment. In [22] it is shown how this approach allows the robot to distinguish objects of different shapes by self-selecting useful stimuli through action, and in [23] it is shown how this approach allows for the exploiting of properties arising from the physical interaction between the robot body and the environment for the purpose of manipulating the object.…”

The Utility of Evolving Simulated Robot Morphology Increases with Task Complexity for Object Manipulation

“…A range sensor returns a value between zero and one commensurate with the length of the ray emitted by the sensor. Clearly these range sensors detract from the anthropomorphic aspect of the robot, but as the goal of the current work is to demonstrate the reasons for co-optimization of morphology and control rather than evolve a realistic robot, this is a minor detail: As has been shown previously [14,27,42], it is possible to evolve object manipulation for anthropomorphic robot arms without range sensors. The shoulder contains a proprioceptive sensor that measures the sagittal rotation of the arm: High positive values indicate the arm is raised, values near zero indicate the arm is horizontal, and high negative values indicate the arm is rotated downward.…”

“…Grasping requires the robot to minimize the distance between its fingers and the object, lifting requires it to maximize the vertical displacement of the object, and active perception requires it to interact with objects of different shapes in order to distinguish between them. Object manipulation has been previously studied in robotics (e.g., [14,27,21,46,42]), and methods for enabling a robot to perform multiple behaviors simultaneously and in sequence constitute a popular area of study (e.g., [46]). Here, however, the goal is not to demonstrate that a robot can accomplish these tasks, but rather the task domain is used to show that there is a positive correlation between the likelihood of successfully performing complex tasks and the amount of the robotʼs morphology placed under evolutionary control.…”