Modeling Autonomic-Dependent Human-Robotic Cooperation
as an Ecosystemic Communication Channel
 

Ron Cottam, Willy Ranson & Roger Vounckx

Abstract

            

True, complete, persistent autonomy in a robot is neither feasible nor desirable. Both survival computation and natural (bio-) semiotics model the successful correlation of persistently differentiable entities as an autonomy negotiation, rather than an imposition. John Collier, for example, has suggested that the human brain cedes bio-support autonomy to the body in favor of gaining information-processing autonomy. The required relationship is far from being a one-way street: useful human-robot interaction implies that human and robot are mutually partially autonomous of each other and partially inter-dependent.

Robotic operation necessitates the involvement of human partners. Systems are only systems if they are unified by Quantum Mechanical-entanglement (QMe). High-level formal artificial information-processing assemblies are always and only unified by their interactions with the QMe of the human brain. This unification is projected by us onto artificial information “systems” so effectively that in virtually all cases we neither notice our involvement in their operation nor include ourselves in their analysis.

Psychologists have long modeled the human use of tools as a “transfer of presence” from a human operator to the tool itself. The manner in which we negotiate difficult passages in an automobile by “becoming the automobile” is an excellent example of this effect. Given the necessity of an autonomy-dependence negotiation between human and robot, adoption of this approach to cooperative development makes far more sense than that of imposing a necessarily incomplete and therefore fragile “autonomy” on a robot.

We regularly learn complex skills by attention to their details, in a way which is often slow and painstaking. Having finally learnt how to integrate these details into a goal (for example, in how to swing a golf club successfully), we then subjugate the detail to a more holistic approach in which we transfer our attention to the target rather than the mechanism of its achievement. The resulting “automation” of a local goal creates the freedom to pay attention to higher-level motives which require the successful “less-than-conscious” embodiment of intermediate techniques.

We propose that the thorny problem of “autonomy, partial autonomy, relinquished control, imposed autonomy, or what?” can best be dealt with by modeling any long-term developmental interaction between human and robot as a bidirectional communication channel, which not only permits human intentions to be transferred to the robot, but also makes its controller aware of the context within which the robot is operating. This recursive “presence correlation” would entail the implementation of an existing birational hierarchical scheme of abstraction to transfer and correlate data, information, understanding, intelligence and wisdom bi-directionally between a robot and its “controller”. The ultimate aim would be to enable a human to operate “less-than-consciously” as him or herself, in the guise of a robot, in a way which permits the cooperative survival of both and ultimately the blurring or removal of their differentiation.
 

______________________________________________
______________________________________________