Motor babbling

Last updated

Motor babbling is a process of repeatedly performing a random motor command for a short duration. It is similar to the vocal babbling of infants, where the brain learns the relation between vocal muscle activities and the resulting sounds. However, it was found that the general motor-control system is already exploring itself in the womb, in animals, in a similar way. Originally, the random spasms and convulsions of the embryo were seen as the non-functional consequences of growth. Later it was realized that the motor system is already calibrating its sensorimotor system before birth. [1] After birth, motor babbling in primates continues in the random grasping movements towards visual targets, training the hand–eye coordination system. These insights are used since the early nineteen nineties in models of biological movement control and in robotics. [2] [3] In robotics, it is a system of robot learning whereby a robotic system can autonomously develop an internal model of its self-body and its environment. Early work is by Kuperstein (1991) [4] using a robot randomly positioning a stick in its workspace, while being observed by two cameras, using a neural network (multilayer perceptron) to associate poses of the stick with joint angles of the arm. This type of research has led to the research field of developmental robotics. [5] [6] [7]

The random exploration of the motor-control state and its effects can lead to the identification of parameters for a inverse kinematics and inverse dynamics model, which the robot can use to find appropriate control signals, given a particular task that requires positioning and/or force control. [8] [9]

See also

Related Research Articles

Neuromorphic computing is an approach to computing that is inspired by the structure and function of the human brain. A neuromorphic computer/chip is any device that uses physical artificial neurons to do computations. In recent times, the term neuromorphic has been used to describe analog, digital, mixed-mode analog/digital VLSI, and software systems that implement models of neural systems. The implementation of neuromorphic computing on the hardware level can be realized by oxide-based memristors, spintronic memories, threshold switches, transistors, among others. Training software-based neuromorphic systems of spiking neural networks can be achieved using error backpropagation, e.g., using Python based frameworks such as snnTorch, or using canonical learning rules from the biological learning literature, e.g., using BindsNet.

Motor learning refers broadly to changes in an organism's movements that reflect changes in the structure and function of the nervous system. Motor learning occurs over varying timescales and degrees of complexity: humans learn to walk or talk over the course of years, but continue to adjust to changes in height, weight, strength etc. over their lifetimes. Motor learning enables animals to gain new skills, and improves the smoothness and accuracy of movements, in some cases by calibrating simple movements like reflexes. Motor learning research often considers variables that contribute to motor program formation, sensitivity of error-detection processes, and strength of movement schemas. Motor learning is "relatively permanent", as the capability to respond appropriately is acquired and retained. Temporary gains in performance during practice or in response to some perturbation are often termed motor adaptation, a transient form of learning. Neuroscience research on motor learning is concerned with which parts of the brain and spinal cord represent movements and motor programs and how the nervous system processes feedback to change the connectivity and synaptic strengths. At the behavioral level, research focuses on the design and effect of the main components driving motor learning, i.e. the structure of practice and the feedback. The timing and organization of practice can influence information retention, e.g. how tasks can be subdivided and practiced, and the precise form of feedback can influence preparation, anticipation, and guidance of movement.

Developmental robotics (DevRob), sometimes called epigenetic robotics, is a scientific field which aims at studying the developmental mechanisms, architectures and constraints that allow lifelong and open-ended learning of new skills and new knowledge in embodied machines. As in human children, learning is expected to be cumulative and of progressively increasing complexity, and to result from self-exploration of the world in combination with social interaction. The typical methodological approach consists in starting from theories of human and animal development elaborated in fields such as developmental psychology, neuroscience, developmental and evolutionary biology, and linguistics, then to formalize and implement them in robots, sometimes exploring extensions or variants of them. The experimentation of those models in robots allows researchers to confront them with reality, and as a consequence, developmental robotics also provides feedback and novel hypotheses on theories of human and animal development.

Sensory processing is the process that organizes and distinguishes sensation from one's own body and the environment, thus making it possible to use the body effectively within the environment. Specifically, it deals with how the brain processes multiple sensory modality inputs, such as proprioception, vision, auditory system, tactile, olfactory, vestibular system, interoception, and taste into usable functional outputs.

Motor control is the regulation of movement in organisms that possess a nervous system. Motor control includes reflexes as well as directed movement.

Cognitive Robotics or Cognitive Technology is a subfield of robotics concerned with endowing a robot with intelligent behavior by providing it with a processing architecture that will allow it to learn and reason about how to behave in response to complex goals in a complex world. Cognitive robotics may be considered the engineering branch of embodied cognitive science and embodied embedded cognition, consisting of Robotic Process Automation, Artificial Intelligence, Machine Learning, Deep Learning, Optical Character Recognition, Image Processing, Process Mining, Analytics, Software Development and System Integration.

Robot learning is a research field at the intersection of machine learning and robotics. It studies techniques allowing a robot to acquire novel skills or adapt to its environment through learning algorithms. The embodiment of the robot, situated in a physical embedding, provides at the same time specific difficulties and opportunities for guiding the learning process.

In physiology, motor coordination is the orchestrated movement of multiple body parts as required to accomplish intended actions, like walking. This coordination is achieved by adjusting kinematic and kinetic parameters associated with each body part involved in the intended movement. The modifications of these parameters typically relies on sensory feedback from one or more sensory modalities, such as proprioception and vision.

<span class="mw-page-title-main">Spiking neural network</span> Artificial neural network that mimics neurons

Spiking neural networks (SNNs) are artificial neural networks that more closely mimic natural neural networks. In addition to neuronal and synaptic state, SNNs incorporate the concept of time into their operating model. The idea is that neurons in the SNN do not transmit information at each propagation cycle, but rather transmit information only when a membrane potential—an intrinsic quality of the neuron related to its membrane electrical charge—reaches a specific value, called the threshold. When the membrane potential reaches the threshold, the neuron fires, and generates a signal that travels to other neurons which, in turn, increase or decrease their potentials in response to this signal. A neuron model that fires at the moment of threshold crossing is also called a spiking neuron model.

Reservoir computing is a framework for computation derived from recurrent neural network theory that maps input signals into higher dimensional computational spaces through the dynamics of a fixed, non-linear system called a reservoir. After the input signal is fed into the reservoir, which is treated as a "black box," a simple readout mechanism is trained to read the state of the reservoir and map it to the desired output. The first key benefit of this framework is that training is performed only at the readout stage, as the reservoir dynamics are fixed. The second is that the computational power of naturally available systems, both classical and quantum mechanical, can be used to reduce the effective computational cost.

A cultured neuronal network is a cell culture of neurons that is used as a model to study the central nervous system, especially the brain. Often, cultured neuronal networks are connected to an input/output device such as a multi-electrode array (MEA), thus allowing two-way communication between the researcher and the network. This model has proved to be an invaluable tool to scientists studying the underlying principles behind neuronal learning, memory, plasticity, connectivity, and information processing.

Neurorobotics is the combined study of neuroscience, robotics, and artificial intelligence. It is the science and technology of embodied autonomous neural systems. Neural systems include brain-inspired algorithms, computational models of biological neural networks and actual biological systems. Such neural systems can be embodied in machines with mechanic or any other forms of physical actuation. This includes robots, prosthetic or wearable systems but also, at smaller scale, micro-machines and, at the larger scales, furniture and infrastructures.

Gail Alexandra Carpenter is an American cognitive scientist, neuroscientist and mathematician. She is now a "Professor Emerita of Mathematics and Statistics, Boston University." She had also been a Professor of Cognitive and Neural Systems at Boston University, and the director of the Department of Cognitive and Neural Systems (CNS) Technology Lab at Boston University.

In machine learning, particularly in the creation of artificial neural networks, ensemble averaging is the process of creating multiple models and combining them to produce a desired output, as opposed to creating just one model. Frequently an ensemble of models performs better than any individual model, because the various errors of the models "average out."

Sensory-motor coupling is the coupling or integration of the sensory system and motor system. Sensorimotor integration is not a static process. For a given stimulus, there is no one single motor command. "Neural responses at almost every stage of a sensorimotor pathway are modified at short and long timescales by biophysical and synaptic processes, recurrent and feedback connections, and learning, as well as many other internal and external variables".

In neuroscience and motor control, the degrees of freedom problem or motor equivalence problem states that there are multiple ways for humans or animals to perform a movement in order to achieve the same goal. In other words, under normal circumstances, no simple one-to-one correspondence exists between a motor problem and a motor solution to the problem. The motor equivalence problem was first formulated by the Russian neurophysiologist Nikolai Bernstein: "It is clear that the basic difficulties for co-ordination consist precisely in the extreme abundance of degrees of freedom, with which the [nervous] centre is not at first in a position to deal."

Neurocomputational speech processing is computer-simulation of speech production and speech perception by referring to the natural neuronal processes of speech production and speech perception, as they occur in the human nervous system. This topic is based on neuroscience and computational neuroscience.

<span class="mw-page-title-main">Daniel Wolpert</span> British neuroscientist

Daniel Mark Wolpert FRS FMedSci is a British medical doctor, neuroscientist and engineer, who has made important contributions in computational biology. He was Professor of Engineering at the University of Cambridge from 2005, and also became the Royal Society Noreen Murray Research Professorship in Neurobiology from 2013. He is now Professor of Neurobiology at Columbia University.

The bi-directional hypothesis of language and action proposes that the sensorimotor and language comprehension areas of the brain exert reciprocal influence over one another. This hypothesis argues that areas of the brain involved in movement and sensation, as well as movement itself, influence cognitive processes such as language comprehension. In addition, the reverse effect is argued, where it is proposed that language comprehension influences movement and sensation. Proponents of the bi-directional hypothesis of language and action conduct and interpret linguistic, cognitive, and movement studies within the framework of embodied cognition and embodied language processing. Embodied language developed from embodied cognition, and proposes that sensorimotor systems are not only involved in the comprehension of language, but that they are necessary for understanding the semantic meaning of words.

<span class="mw-page-title-main">Interlimb coordination</span> Coordination of the left and right limbs

Interlimb coordination is the coordination of the left and right limbs. It could be classified into two types of action: bimanual coordination and hands or feet coordination. Such coordination involves various parts of the nervous system and requires a sensory feedback mechanism for the neural control of the limbs. A model can be used to visualize the basic features, the control centre of locomotor movements, and the neural control of interlimb coordination. This coordination mechanism can be altered and adapted for better performance during locomotion in adults and for the development of motor skills in infants. The adaptive feature of interlimb coordination can also be applied to the treatment for CNS damage from stroke and the Parkinson's disease in the future.

References

  1. ↑ Bernstein, N.A. (1967). The Co-Ordination and Regulation of Movements. Oxford: Pergamon Press.
  2. ↑ Bullock, Daniel; Grossberg, Stephen; Guenther, Frank (1993). "A Self-Organizing Neural Model of Motor Equivalent Reaching and Tool Use by a Multijoint Arm". Journal of Cognitive Neuroscience. 5 (4): 408–435. doi:10.1162/jocn.1993.5.4.408. PMID   23964916. S2CID   207610285.
  3. ↑ Villa, Alessandro E. P.; Masulli, Paolo; Rivero, Antonio Javier Pons (2016). Artificial Neural Networks and Machine Learning – ICANN 2016: 25th International Conference on Artificial Neural Networks, Barcelona, Spain, September 6-9, 2016, Proceedings. Cham: Springer. p. 167. ISBN   978-3-319-44777-3.
  4. ↑ Kuperstein, Michael (1991). "INFANT neural controller for adaptive sensory-motor coordination". Neural Networks. 4 (2): 131–145. doi:10.1016/0893-6080(91)90001-L. ISSN   0893-6080.
  5. ↑ Saegusa, Ryo; Metta, Giorgio; Sandini, Giulio; Sakka, Sophie. "Active Motor Babbling for Sensory-Motor Learning" (PDF).{{cite journal}}: Cite journal requires |journal= (help)
  6. ↑ Demiris, Yiannis, and Anthony Dearden. "From motor babbling to hierarchical learning by imitation: a robot developmental pathway." (2005): 31-37.
  7. ↑ Saegusa, Ryo, et al. "Active motor babbling for sensorimotor learning." Robotics and Biomimetics, 2008. ROBIO 2008. IEEE International Conference on. IEEE, 2009.
  8. ↑ Arik, Sabri; Huang, Tingwen; Lai, Weng Kin; Liu, Qingshan (2015). Neural Information Processing: 22nd International Conference, ICONIP 2015, Istanbul, Turkey, November 9–12, 2015, Proceedings. Cham: Springer. p. 27. ISBN   978-3-319-26554-4.
  9. ↑ McFarland, David; Stenning, Keith; McGonigle, Maggie (2012). The Complex Mind: An Interdisciplinary Approach. New York: Palgrave Macmillan. p. 119. ISBN   978-0-230-35445-6.
This page is based on this Wikipedia article
Text is available under the CC BY-SA 4.0 license; additional terms may apply.
Images, videos and audio are available under their respective licenses.