Biological systems thinking for control engineering design

In this paper it is shown how a simple neuro-mechanical network can be configured and tuned to perform a certain task, using optimization based on simulation. This would have the potential of creating structures that could be trained and adapted to perform certain actions. The basic components hardly exist at present. The objective here, however, is to study what they could be like and how they could be configured. Although the neuromechanical network systems are partly aimed towards future technologies, the concept of neuro-mechanical networks can be useful much sooner as it has meaning also at a micro or even macro level in order to realize highly robust flexible actuator systems. Another potential use is for design of more conventional system, with a minimum of components. Also it can be used as an explanatory model for some of the mechanics found in human body tissue.

The Journal of the Acoustical Society of America, 1995

A biologically inspired control approach for reducing vibrations in distributed elastic systems has been derived and experimentally verified for narrow-band excitation. The control paradigm approximates natural biological systems for initiating movement, in that a low number of signals are sent from an advanced, centralized controller (analogous to the motor cortex of the brain) and are then distributed by local simple rules to multiple control actuators (analogous to muscle fiber). Both theoretical and experimental investigations of three different local rules were carried out including a stability analysis for reducing beam vibrations. In general the results have demonstrated that the biological control approach has the potential to control multimodal response in distributed elastic systems using an array of many actuators with a reduced order main controller. Thus significant reductions in control system computational complexity have been realized by this approach.

Proceedings of the 6th International Conference on Frontiers of Information Technology - FIT '09, 2009

The complexity of technical systems in automation is rapidly growing and is causing increasing costs. Research has to identify appropriate measures to deal with the increasing complexity; Artificial Intelligence (AI) and Cognitive Science (CS) already show noteworthy results. However, compared to human intelligence those results can only be valued as modest. In the following the authors will try to verify the necessity of developing new principles to model control systems and will try to show, which technical principles and tools are suitable for the defined requirements. However, the challenge will be to accept a new kind of thinking in the bionic world, which will be in accord with scientists of other specific scientific faculties such as neuropsychoanalysis. The goal must be a formal approach in specifying the psyche. 1

Proceedings of PerAc '94. From Perception to Action, 1994

This paper provides a high-level review of current and recent work in the use of genetic algorithm based techniques to develop sensorimotor control systems for autonomous agents. It focuses on network-based controllers and genetic encoding issues associated with them. The paper closes with a discussion of the possibility of using arti cial evolutionary techniques to help tackle more speci cally scienti c questions about natural sensorimotor systems.

Proceedings of the 9th annual conference on Genetic and evolutionary computation - GECCO '07, 2007

A primary goal of evolutionary robotics is to create systems that are as robust and adaptive as the human body. Moving toward this goal often involves training control systems that process sensory information in a way similar to humans. Artificial neural networks have been an increasingly popular option for this because they consist of processing units that approximate the synaptic activity of biological signal processing units, i.e. neurons. In this paper we train a nonlinear recurrent spino-neuromuscular system (SNMS) model and compare the performance of genetic algorithms (GA)s, particle swarm optimizers (PSO)s, and GA/PSO hybrids. Several key features of the SNMS model have previously been modeled individually but have not been combined into a single model as is done here. The results show that each algorithm produces fit solutions and generates fundamental biological behaviors, such as tonic tension behaviors and triceps activation patterns, that are not explicitly trained.

Frontiers in neurorobotics, 2009

Modelling is continuously being deployed to gain knowledge on the mechanisms of motor control. Computational models, simulating the behaviour of complex systems, have often been used in combination with soft computing strategies, thus shifting the rationale of modelling from the description of a behaviour to the understanding of the mechanisms behind it. In this context, computational models are preferred to deterministic schemes because they deal better with complex systems. The literature offers some striking examples of biologically inspired modelling, which perform better than traditional approaches when dealing with both learning and adaptivity mechanisms. Can these theoretical studies be transferred into an application framework? That is, can biologically inspired models be used to implement rehabilitative devices? Some evidences, even if preliminary, are presented here, and support an affi rmative answer to the previous question, thus opening new perspectives.

The main objective of this work is to analyze the computing capabilities of human neuroblastoma cultured cells and to define stimulation patterns able to modulate the neural activity in response to external stimuli for controlling an autonomous robot. Multielectrode Arrays Setups have been designed for direct culturing neural cells over silicon or glass substrates, providing the capability to stimulate and record simultaneously populations of neural cells. This paper tries to modulate the natural physiologic responses of human neural cells by tetanic stimulation of the culture. If we are able to modify the selective responses of some cells with a external pattern stimuli over different time scales, the neuroblastoma-cultured structure could be trained to process pre-programmed spatio-temporal patterns. We show that the large neuroblastoma networks developed in cultured MEAs are capable of learning: stablishing numerous and dynamic connections, with modifiability induced by external stimuli.

2012

This paper presents a model of a control system for robot systems inspired by the functionality and organisation of human neuroregulatory system. Our model was specified using software agents within a formal framework and implemented through Web Services. This approach allows the implementation of the control logic of a robot system with relative ease, in an incremental way, using the addition of new control centres to the system as its behaviour is observed or needs to be detailed with greater precision, without the need to modify existing functionality. The tests performed verify that the proposed model has the general characteristics of biological systems together with the desirable features of software, such as robustness, flexibility, reuse and decoupling.

Biomedical Engineering / 765: Telehealth / 766: Assistive Technologies, 2012

Biologically inspired algorithms were used in this work to approach different components of pattern recognition applied to the control of robotic prosthetics. In order to contribute with a different training paradigm, Evolutionary (EA) and Particle Swarm Optimization (PSO) algorithms were used to train an Artificial Neural Network (ANN). Since the optimal input set of signal features is yet unknown, a Genetic Algorithm (GA) was used to approach this problem. The training length and rate of convergence were considered in the search of an optimal set of signal features, as well as for the optimal time window length. The ANN proved to be an accurate pattern recognition algorithm predicting 10 movements with over 95% accuracy. Moreover, new combinations of signal features with higher convergence rates than the commonly found in the literature were discovered by the GA. It was also found that the PSO had better performance that the EA as a training algorithm but worse than the well established Back-propagation. The latter considered accuracy, training length and convergence. Finally, the common practice of using 200 ms time window was found to be sufficient for producing acceptable accuracies while remaining short enough for a real-time control.

Patterns, 2021

Three dissimilar methodologies in the field of artificial intelligence (AI) appear to be following a common path toward biological authenticity. This trend could be expedited by using a common tool, artificial nervous systems (ANS), for recreating the biology underpinning all three. ANS would then represent a new paradigm for AI with application to many related fields.