Self-organizing Neural Controller Research Articles

Evolving Modularity in Soft Robots Through an Embodied and Self-Organizing Neural Controller.

Modularity is a desirable property for embodied agents, as it could foster their suitability to different domains by disassembling them into transferable modules that can be reassembled differently. We focus on a class of embodied agents known as voxel-based soft robots (VSRs). They are aggregations of elastic blocks of soft material; as such, their morphologies are intrinsically modular. Nevertheless, controllers used until now for VSRs act as abstract, disembodied processing units: Disassembling such controllers for the purpose of module transferability is a challenging problem. Thus, the full potential of modularity for VSRs still remains untapped. In this work, we propose a novel self-organizing, embodied neural controller for VSRs. We optimize it for a given task and morphology by means of evolutionary computation: While evolving, the controller spreads across the VSR morphology in a way that permits emergence of modularity. We experimentally investigate whether such a controller (i) is effective and (ii) allows tuning of its degree of modularity, and with what kind of impact. To this end, we consider the task of locomotion on rugged terrains and evolve controllers for two morphologies. Our experiments confirm that our self-organizing, embodied controller is indeed effective. Moreover, by mimicking the structural modularity observed in biological neural networks, different levels of modularity can be achieved. Our findings suggest that the self-organization of modularity could be the basis for an automatic pipeline for assembling, disassembling, and reassembling embodied agents.

Autonomous Identification and Goal-Directed Invocation of Event-Predictive Behavioral Primitives

Voluntary behavior of humans appears to be composed of small, elementary building blocks, or behavioral primitives. While this modular organization seems crucial for the learning of complex motor skills and the flexible adaption of behavior to new circumstances, the problem of learning meaningful, compositional abstractions from sensorimotor experiences remains an open challenge. Here, we introduce a computational learning architecture, termed as surprise-based behavioral modularization into event-predictive structures (SUBMODES) that explores behavior and identifies the underlying behavioral units completely from scratch. The SUBMODES architecture bootstraps sensorimotor exploration using a self-organizing neural controller. While exploring the behavioral capabilities of its own body, the system learns modular structures that predict the sensorimotor dynamics and generate the associated behavior. In line with recent theories of event perception, the system uses unexpected prediction error signals, i.e., surprise, to detect transitions between successive behavioral primitives. We show that, when applied to two robotic systems with completely different body kinematics, the system manages to learn a variety of complex behavioral primitives. Moreover, after initial self exploration the system can use its learned predictive models progressively more effectively for invoking model predictive planning and goal-directed control in different tasks and environments.

Self-organizing neural control system design for dynamic processes

Abstract This paper discusses the achievable nominal performance of a well-parametrized neural feedback control system, and proposes an efficient training method for parametrizing such a controller. A self-organizing neural control (SONC) system is presented in which a layered feedforward neural network is adopted as the controller structure in order to apply directly existing back-propagated learning techniques. A self-organizing methodology is introduced to provide the training set for adjusting parameters of the neural controller. One important feature of the proposed adaptive mechanism is that, though it should lack extensive knowledge of the process dynamics at the outset of controller design, it will still be able to achieve its desired results by employing the subjective experience of control specialists as its training aids. Tuning variables of the SONC system are reviewed through exploring their effects on five typical transfer functions. The applicability of the SONC system is also demonstrated on a continuous stirred tank reactor. Simulation results show that a well-parametrized neural controller can improve nominal performance for a wide variety of different processes, and the proposed self-organizing mechanism can direct a controller to achieve the desired final parametrization.