An official website of the United States government
In a complex system, the interactions between individual agents often lead to emergent collective behavior such as spontaneous synchronization, swarming, and pattern formation. Beyond the intrinsic properties of the agents, the topology of the network of interactions can have a dramatic influence over the dynamics. In many studies, researchers start with a specific model for both the intrinsic dynamics of each agent and the interaction network and attempt to learn about the dynamics of the model. Here, we consider the inverse problem: given data from a system, can one learn about the model and the underlying network? We investigate arbitrary networks of coupled phase oscillators that can exhibit both synchronous and asynchronous dynamics. We demonstrate that, given sufficient observational data on the transient evolution of each oscillator, machine learning can reconstruct the interaction network and identify the intrinsic dynamics.
more »
« less
An output-driven approach to design a swarming model for architectural indoor environments
We introduce a novel tool for designing a swarming behavior model for a set of virtual agents to automatically capture an initially unknown indoor architectural environment. Our key idea is to use an output-driven optimization to create targeted swarming behavior. The input to our model is a simple rectangular proxy of the target area and desired acquisition indicator values. The final outputs are the parameters for a swarming behavior model that is autonomous and decentralized, uses only local exploration, and is robust to agent failure. We show and compare the swarming performance in several simulated environments of up to several hundred square meters, 100 agents, and under various conditions.
more »
« less
- Award ID(s):
- 1835739
- PAR ID:
- 10211196
- Date Published:
- Journal Name:
- Computers graphics
- Volume:
- 87
- ISSN:
- 0097-8493
- Page Range / eLocation ID:
- 103-110
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
It is now possible to deploy swarms of drones with populations in the thousands. There is growing interest in using such swarms for defense, and it has been natural to program them with bio-mimetic motion models such as flocking or swarming. However, these motion models evolved to survive against predators, not enemies with modern firearms. This paper presents experimental data that compares the survivability of several motion models for large numbers of drones. This project tests drone swarms in Virtual Reality (VR), because it is prohibitively expensive, technically complex, and potentially dangerous to fly a large swarm of drones in a testing environment. We model the behavior of drone swarms flying along parametric paths in both tight and scattered formations. We add random motion to the general motion plan to confound path prediction and targeting. We describe an implementation of these flight paths as game levels in a VR environment. We then allow players to shoot at the drones and evaluate the difference between flocking and swarming behavior on drone survivability.more » « less
-
At the microscale, coupled physical interactions between collectives of agents can be exploited to enable self-organization. Past systems typically consist of identical agents; however, heterogeneous agents can exhibit asymmetric pairwise interactions which can be used to generate more diverse patterns of self-organization. Here, we study the effect of size heterogeneity in microrobot collectives composed of circular, magnetic microdisks on a fluid–air interface. Each microrobot spins or oscillates about its center axis in response to an external oscillating magnetic field, in turn producing local magnetic, hydrodynamic, and capillary forces that enable diverse collective behaviors. We demonstrate through physical experiments and simulations that the heterogeneous collective can exploit the differences in microrobot size to enable programmable self-organization, density, morphology, and interaction with external passive objects. Specifically, we can control the level of self-organization by microrobot size, enable organized aggregation, dispersion, and locomotion, change the overall shape of the collective from circular to ellipse, and cage or expel objects. We characterize the fundamental self-organization behavior across a parameter space of magnetic field frequency, relative disk size, and relative populations; we replicate the behavior through a physical model and a swarming coupled oscillator model to show that the dominant effect stems from asymmetric interactions between the different-sized disks. Our work furthers insights into self-organization in heterogeneous microrobot collectives and moves us closer to the goal of applying such collectives to programmable self-assembly and active matter.more » « less
-
Neurobiological theories of spatial cognition developed with respect to recording data from relatively small and/or simplistic environments compared to animals’ natural habitats. It has been unclear how to extend theoretical models to large or complex spaces. Complementarily, in autonomous systems technology, applications have been growing for distributed control methods that scale to large numbers of low-footprint mobile platforms. Animals and many-robot groups must solve common problems of navigating complex and uncertain environments. Here, we introduce the NeuroSwarms control framework to investigate whether adaptive, autonomous swarm control of minimal artificial agents can be achieved by direct analogy to neural circuits of rodent spatial cognition. NeuroSwarms analogizes agents to neurons and swarming groups to recurrent networks. We implemented neuron-like agent interactions in which mutually visible agents operate as if they were reciprocally connected place cells in an attractor network. We attributed a phase state to agents to enable patterns of oscillatory synchronization similar to hippocampal models of theta-rhythmic (5–12 Hz) sequence generation. We demonstrate that multi-agent swarming and reward-approach dynamics can be expressed as a mobile form of Hebbian learning and that NeuroSwarms supports a single-entity paradigm that directly informs theoretical models of animal cognition. We present emergent behaviors including phase-organized rings and trajectory sequences that interact with environmental cues and geometry in large, fragmented mazes. Thus, NeuroSwarms is a model artificial spatial system that integrates autonomous control and theoretical neuroscience to potentially uncover common principles to advance both domains.more » « less
-
Deploying decentralized control strategies for outdoor multi-agent Uncrewed Aircraft Systems (UASs) is challenging due to timing variations, packet loss, and computing resource limitations. In this work we address robustness to these conditions through a novel co-regulated control strategy that varies the periodicity of control inputs and communication with other agents. Co-regulation is applied to a decentralized hierarchical controller consisting of a global component governing inter-group coordination to multiple targets while a local component governs intra-group coordination of the agents as they progress to the target of interest. The control gains are “gain scheduled” according to current conditions while a cyber controller schedules the control and communication tasks for execution based on swarm performance. The control gains are found via reinforcement learning and the entire algorithm is deployed on a swarm consisting of 7 custom agents. Our results show the impact of rethinking swarming algorithms with computation and communication resource limitations in mind and indicate we can provide exceptional swarm control utilizing fewer resources while also improving the quality of service for an onboard, anytime collision avoidance algorithm.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1016/j.cag.2020.02.003
