A5066791342- Modular Robots and Swarm Intelligence
- Reinforcement Learning in Robotics
- Evolutionary Algorithms and Applications
- Robot Manipulation and Learning
- Robotic Locomotion and Control
- Micro and Nano Robotics
- Soft Robotics and Applications
- Advanced Materials and Mechanics
- Neural dynamics and brain function
- Mobile Crowdsensing and Crowdsourcing
- Metaheuristic Optimization Algorithms Research
- Evolutionary Game Theory and Cooperation
- Cellular Automata and Applications
- Embodied and Extended Cognition
- Evolution and Genetic Dynamics
- Neural Networks and Applications
- Gene Regulatory Network Analysis
- Artificial Intelligence in Games
- Robotic Path Planning Algorithms
- Advanced Memory and Neural Computing
- Machine Learning and Algorithms
- Machine Learning and Data Classification
- Computability, Logic, AI Algorithms
- Advanced Sensor and Energy Harvesting Materials
- Cellular Mechanics and Interactions
University of Vermont
2015-2024
Morpho (United States)
2011-2024
INFICON (United States)
2023
Radboud University Nijmegen
2016
Cornell University
2004-2007
University of Sussex
1999-2007
University of Coimbra
2007
Saarland University
2007
Carnegie Mellon University
2007
Indiana University Bloomington
2007
Complex nonlinear dynamics arise in many fields of science and engineering, but uncovering the underlying differential equations directly from observations poses a challenging task. The ability to symbolically model complex networked systems is key understanding them, an open problem disciplines. Here we introduce for first time method that can automatically generate symbolic coupled dynamical system series data. This applicable any be described using sets ordinary equations, assumes...
Animals sustain the ability to operate after injury by creating qualitatively different compensatory behaviors. Although such robustness would be desirable in engineered systems, most machines fail face of unexpected damage. We describe a robot that can recover from change autonomously, through continuous self-modeling. A four-legged machine uses actuation-sensation relationships indirectly infer its own structure, and it then this self-model generate forward locomotion. When leg part is...
Living systems are more robust, diverse, complex, and supportive of human life than any technology yet created. However, our ability to create novel lifeforms is currently limited varying existing organisms or bioengineering organoids in vitro. Here we show a scalable pipeline for creating functional lifeforms: AI methods automatically design diverse candidate silico perform some desired function, transferable designs then created using cell-based construction toolkit realize living with the...
Taking a biologically inspired approach to the design of autonomous, adaptive machines.
We introduce a system that combines ontogenetic development and artificial evolution to automatically design robots in physics-based, virtual environment. Through lesion experiments on the evolved agents, we demonstrate genetic regulatory networks from successful evolutionary runs are more modular than those obtained unsuccessful runs.
Here we introduce one simulated and two physical three-dimensional stochastic modular robot systems, all capable of self-assembly self-reconfiguration.We assume that individual units can only draw power when attached to the growing structure, have no means actuation.Instead they are subject random motion induced by surrounding medium unattached.We present a simulation environment with flexible scripting language allows for parallel serial selfassembly self-reconfiguration processes.We...
Most animals exhibit significant neurological and morphological change throughout their lifetime. No robots to date, however, grow new structure while behaving. This is due technological limitations but also because it unclear that provides a benefit the acquisition of robust behavior in machines. Here I show evolving populations simulated robots, if from anguilliform into legged during lifetime early stages evolution, body plan gradually lost later gaits are evolved for final, form robot...
Whether, when, how, and why increased complexity evolves in biological populations is a longstanding open question. In this work we combine recently developed method for evolving virtual organisms with an information-theoretic metric of morphological order to investigate how the morphologies, which are evolved locomotion, varies across different environments. We first demonstrate that selection locomotion results evolution morphologies increase over evolutionary time beyond what would be...
Taking a biologically inspired approach to the design of autonomous, adaptive machines.
All living systems perpetuate themselves via growth in or on the body, followed by splitting, budding, birth. We find that synthetic multicellular assemblies can also replicate kinematically moving and compressing dissociated cells their environment into functional self-copies. This form of perpetuation, previously unseen any organism, arises spontaneously over days rather than evolving millennia. show how artificial intelligence methods design postpone loss replicative ability perform...
Organisms result from adaptive processes interacting across different time scales. One such interaction is that between development and evolution. Models have shown sweeps over several traits in a single agent, sometimes exposing promising static traits. Subsequent evolution can then canalize these rare Thus, can, under the right conditions, increase evolvability. Here, we report on previously unknown phenomenon when embodied agents are allowed to develop evolve: Evolution discovers body...
Evolution sculpts both the body plans and nervous systems of agents together over time. By contrast, in artificial intelligence robotics, a robot's plan is usually designed by hand, control policies are then optimized for that fixed design. The task simultaneously co-optimizing morphology controller an embodied robot has remained challenge. In psychology, theory cognition posits behaviour arises from close coupling between sensorimotor control, which suggests why these two subsystems so...
Advances in science and engineering often reveal the limitations of classical approaches initially used to understand, predict, control phenomena. With progress, conceptual categories must be re-evaluated better track recently discovered invariants across disciplines. It is essential refine frameworks resolve conflicting boundaries between disciplines such that they facilitate, not restrict, experimental capabilities. In this essay, we address specific questions critiques which have arisen...
New robotics is an approach to that, in contrast traditional robotics, employs ideas and principles from biology. While the there are generally accepted methods (e.g., control theory), designing agents new still largely considered art. In recent years, we have been developing a set of heuristics, or design principles, that on one hand capture theoretical insights about intelligent (adaptive) behavior, other provide guidance actually building systems. this article overview all but focus...
We present a coevolutionary algorithm for inferring the topology and parameters of wide range hidden nonlinear systems with minimum experimentation on target system. The synthesizes an explicit model directly from observed data produced by intelligently generated tests. is composed two coevolving populations. One population evolves candidate models that estimate structure second informative tests either extract new information system or elicit desirable behavior it. fitness their ability to...
Object The authors describe the artificial neural network (ANN) as an innovative and powerful modeling tool that can be increasingly applied to develop predictive models in neurosurgery. They aimed demonstrate utility of ANN predicting survival following traumatic brain injury compare its ability with regression clinicians. Methods designed predict in-hospital injury. model was generated 11 clinical inputs a single output. Using subset National Trauma Database, “trained” outcome by providing...
An ensemble is a set of learned models that make decisions collectively. Although an usually more accurate than single learner, existing methods often tend to construct unnecessarily large ensembles, which increases the memory consumption and computational cost. Ensemble pruning tackles this problem by selecting subset members form subensembles are subject less resource response time with accuracy similar or better original ensemble. In paper, we analyze accuracy/diversity trade-off prove...