Psychoanalytical model for automation and robotics

Abstract

Research in automation focuses on systems which are capable of solving very complex tasks and problems. Artificial Intelligence and especially Cognitive Science have brought remarkable successes; however, in some areas the boarders of feasibility and further extension are reached. Compared to human intelligence the range of capabilities of the solutions is still modest. In the following we will argue why we see the necessity to introduce a novel approach for creating models, which possibilities and tools computer engineering can offer, why a psychoanalytical template is considered meaningful, and which open problems could be tackled or even broken through with this approach, respectively. The article is based on comprehensive research results in the course of several research projects including a European one. Involved persons originate from a number of research institutions in Austria, South Africa, and Canada.

A decision unit inspired by cognitive sciences for future robotic applications

“…Current publications of the model like [14] prove the validity and importance of the model for engineering purposes. The team is now focusing on evaluating the model concerning the psychoanalytic correctness and performance in comparison to state of the art decision units.…”

“…Using the top down design process a first functional model was introduced in [11] and [14] and described in more detail in [2] using the second topographical model of S. Freud. The functional model describes in detail how the different input parameters of the system form the intention to do something (a wish).…”

Experience driven adaptive behaviour of autonomous humanoid robots

“…The feelings are not utilised for expressing emotions or responding to human emotional cues [1] but for enabling the robot to adapt its behaviour by means of experiences in order to achieve its objectives. For the related algorithms must run on the limited resources of the microcontroller embedded in the robot, our approach focuses on lean algorithms rather than modelling functions of the brain [2,3].…”

Humanoid robot control by intelligent sensors and self-generating will

“…Although it is intelligent in a sense that it can adapt by learning new object classes and words, it cannot cope with conflicting requirements and cannot use behavioural experiences it has made previously. In order to enable the robot to act intelligently [12][13] when boundary conditions have changed and requirements contradict, we propose to equip the robot with a self-generating will based on the algorithms we use so far.…”