ARS | Publication request: Data-driven identification of group dynamics for motion prediction and control

Submitted to: Journal of Field Robotics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: April 2, 2008
Publication Date: June 1, 2008
Citation: Schwager, M., Deweiler, C., Vasilescu, L., Anderson, D.M., Rus, D. 2008. Data-driven identification of group dynamics for motion prediction and control. Journal of Field Robotics. 25:305-324.

Interpretive Summary: This paper addresses a minimalist approach (Least Squares method) to modeling group behavior of free-ranging cows using 1 HZ global position system (GPS) data obtained on two different size cow herds. The groups were each monitored within a 466 ha arid rangeland paddock during February (n=3) and July (n=10) 2007 using electronic equipment developed by the Massachessusets Institute of Technology (MIT) Computer Science and Artificial Intelligence Laboratory (CSAIL) and an equipment harness designed and built by the US Department of Agriculture-Agricultural Research Service-Jornada Experimental Range (USDA-ARS-JER). The electronics box measured 21.5 cm x 12.0 cm x 5.5 cm and weighed approximately 1 kg. The box’s top was covered with solar panels and wireless antennas that were held in the upright position on the cow’s head using a prototype equipment platform in which the animal's ears prevented the harness from rotating. Biologically the data collected suggests a cow in a small herd may interact differently with peers compared to when she is a member of a larger herd of similar peers. Modeling these data suggest that a minimalist approach to modeling group interactions using only position data can lead to a meaningful dynamical model of individual members of the group as well as revealing characteristics of the group especially where no detailed environmental information is available. Future application of this information coupled with altering the animal's movement using sensory stimuli may allow for the development of a mathematical model that could assist in the control of free-ranging cattle.

Technical Abstract: A distributed model structure for representing groups of coupled dynamic agents is proposed, and the Least Squares method is used for fitting model parameters based on measured position data. The difference equation model embodies a minimalist approach, only incorporating factors essential to the movement and interaction of physical bodies. The model combines effects from an agent’s inertia, interactions between agents, and interactions between each agent and its environment. GPS tracking data were collected in field experiments from a group of three cows and a group of ten cows over the course of several days using custom-designed, head mounted sensor boxes. These data are used with the Least Squares method to fit the model to the cow groups. The modeling technique is shown to capture overall characteristics of the group as well as attributes of individual group members. Applications to livestock management are described, and the potential for surveillance, prediction, and control of various kinds of groups of dynamical agents are suggested.