Biocosm: A Simulation Environment for Designing Energy-Constrained Ethological Sensors

Long-duration ethological sensing increasingly depends on miniaturized, battery-constrained devices that must decide when to observe, transmit, store, compute, or sleep. These policy decisions shape not only device lifetime, but also which biological events are ultimately available for inference. This manuscript presents B iocosm , a measurement-process simulation framework for energy-constrained ethological instrumentation. B iocosm separates the latent biological world from the observation model, sensing policy, energy model, and analysis layer, allowing researchers to evaluate how alternative sensing strategies transform true behavioral interactions into observed data. Although motivated by wireless proximity logging, the framework is hardware-agnostic: beaconing, listening, sensing, storage, and computation are represented as parameterized state/action costs rather than as fixed properties of one device architecture. This structure supports reproducible design-space exploration, policy comparison, and sensitivity analysis before field deployment. B iocosm is grounded in prior work on animal proximity loggers, low-power wireless discovery and energy modeling, agent-based ethological simulation, and digital twin-inspired methods in medicine and neurotechnology. Framing measurement as a coupled world–observation–policy system makes instrumentation choices explicit and quantifiable.