[Paper Review] Emergent oscillations assist obstacle negotiation during ant cooperative transport

Aviram Gelblum, Itai Pinkoviezky+3 more|arXiv (Cornell University)|Jul 20, 2021

Insect and Arachnid Ecology and Behavior89 references27 citations

TL;DR

The study shows that when ants cooperatively transport a load encountering an obstacle, constraint-induced relaxation oscillations emerge as a group-level solution, not from changes in individual behavior, and are explained by a two-forces model (informed pull toward the nest and uninformed motion-alignment) whose interactions produce order, oscillations, or complete rotations depending on system size.

ABSTRACT

Collective motion by animal groups is affected by internal interactions, external constraints and the influx of information. A quantitative understanding of how these different factors give rise to different modes of collective motion is, at present, lacking.} Here, we study how ants that cooperatively transport a large food item react to an obstacle blocking their path. Combining experiments with a statistical physics model of mechanically coupled active agents, we show that the constraint induces a deterministic collective oscillatory mode that facilitates obstacle circumvention. We provide direct experimental evidence, backed by theory, that this motion is an emergent group effect that does not require any behavioral changes at the individual level. We trace these relaxation oscillations to the interplay between two forces; informed ants pull the load towards the nest while uninformed ants contribute to the motion's persistence along the tangential direction. The model's predictions that oscillations appear above a critical system size, that the group can spontaneously transition into its ordered phase, and that the system can exhibit complete rotations are all verified experimentally. We expect that similar oscillatory modes emerge in collective motion scenarios where the structure of the environment imposes conflicts between individually held information and the group's tendency for cohesiveness.

Motivation & Objective

• Investigate how obstacle constraints affect collective transport by ant groups.
• Determine whether obstacle-induced oscillations are emergent at the group level rather than driven by individual behavioral changes.
• Test a minimal, physically grounded model against constrained experimental data to assess its predictive power.

Proposed method

• Combine experiments with tethered-load and obstacle setups to provoke constrained cooperative transport.
• Use a statistical physics model of mechanically coupled active agents distinguishing informed (nest-directed) and uninformed (motion-aligned) ants.
• Modify the model with a tether constraint to reproduce obstacle-induced dynamics without changing individual decision rules.
• Fit model parameters to experimental observables such as angular velocity, amplitude distributions, and phase-space trajectories.
• Analyze a simplified three-phase (stationary, oscillations, rotations) dynamical system to elucidate bifurcation structure and transitions.

Experimental results

Research questions

• RQ1 Do obstacle constraints induce emergent collective oscillations in cooperative transport?
• RQ2 Can a minimal two-force model (informed vs. uninformed ants) reproduce obstacle-induced oscillations without changing individual behavior?
• RQ3 How does system size affect the transition from disordered to oscillatory to rotational motion in constrained ant groups?
• RQ4 What is the relationship between tether length and oscillation period, and can the model predict regime transitions (including complete rotations)?

Key findings

• Obstacle constraints induce nearly deterministic tangential oscillations that facilitate obstacle circumvention.
• Oscillations arise without any individual ant altering its behavioral program; they are an emergent group-level effect of conflicting forces.
• A minimal model with informed (nest-directed) and uninformed (motion-aligned) ants quantitatively matches constrained experimental data.
• There is a size-dependent order–disorder transition: small groups show random-walk-like behavior, while larger groups exhibit persistent oscillations or complete rotations.
• The oscillator period scales linearly with tether length and increases with system size, with a predicted transition boundary to rotation.
• Experiments with loads of different sizes confirm model predictions, including a regime where complete rotations occur for larger objects.