Robots that learn to evaluate models of collective behavior
2026-04-08 • Robotics
Robotics
AI summaryⓘ
The authors created a setup where a robotic fish controlled by computer models interacts with real fish to test how well these models predict fish behavior. They trained different models, including simple rule-based ones and a neural network, in a simulator and then used them to control the robot in real life. By comparing how the real fish responded to the robot versus the simulation, they measured which model behaved more like real fish. Their results show that the neural network model matched real fish behavior better than the simpler models. This method helps evaluate animal behavior models in a more interactive and realistic way.
Reinforcement learningBiomimetic roboticsFish behavior modelingConvolutional neural networksSim-to-real transferClosed-loop interactionBehavioral metricsWasserstein distanceCollective motionBio-inspired robotics
Authors
Mathis Hocke, Andreas Gerken, David Bierbach, Jens Krause, Tim Landgraf
Abstract
Understanding and modeling animal behavior is essential for studying collective motion, decision-making, and bio-inspired robotics. Yet, evaluating the accuracy of behavioral models still often relies on offline comparisons to static trajectory statistics. Here we introduce a reinforcement-learning-based framework that uses a biomimetic robotic fish (RoboFish) to evaluate computational models of live fish behavior through closed-loop interaction. We trained policies in simulation using four distinct fish models-a simple constant-follow baseline, two rule-based models, and a biologically grounded convolutional neural network model-and transferred these policies to the real RoboFish setup, where they interacted with live fish. Policies were trained to guide a simulated fish to goal locations, enabling us to quantify how the response of real fish differs from the simulated fish's response. We evaluate the fish models by quantifying the sim-to-real gaps, defined as the Wasserstein distance between simulated and real distributions of behavioral metrics such as goal-reaching performance, inter-individual distances, wall interactions, and alignment. The neural network-based fish model exhibited the smallest gap across goal-reaching performance and most other metrics, indicating higher behavioral fidelity than conventional rule-based models under this benchmark. More importantly, this separation shows that the proposed evaluation can quantitatively distinguish candidate models under matched closed-loop conditions. Our work demonstrates how learning-based robotic experiments can uncover deficiencies in behavioral models and provides a general framework for evaluating animal behavior models through embodied interaction.