Overview

The goal of this full-day workshop is to bring together leading researchers in the various areas of robotics to address the role of stochasticity in natural and artificial systems and establish a scientific and disciplinary foundation for robotics. Stochasticity has been central to numerous robotics problems: state and parameter estimation from noisy data, decision-making based on partial or incomplete information, and modeling of dynamic and unpredictable environments. In biology, there are many examples of systems composed of a multitude of stochastically interacting components whose collective behavior are cohesive, robust, and even versatile. In fact, recent advances in systems biology have shown that the stochastic nature of cellular behaviors and molecular dynamics is crucial to understanding their collective behaviors and regulatory mechanisms. Researchers in swarm robotics and other areas are taking inspiration from such examples in nature and developing new stochastic modeling, analysis, and design techniques. This workshop will bring together researchers from diverse backgrounds to address stochasticity in various areas in robotics. We believe by focusing on common themes across various disciplines rather than specific problems at hand will facilitate interdisciplinary studies and lead to important cross-disciplinary discoveries and breakthroughs.

Robots and the environments they operate in are inherently noisy. Consequently, the ability to model and analyze stochasticity is crucial to the development of effective robotic systems. In recent years, the robotics community has explored new system architectures, algorithms, communication and control protocols to deal with the stochastic nature of our physical world. However, the conventional belief where uncertainty is considered an undesirable feature of the system remains.

Similar to robotic systems, biological systems are noisy systems that built from ensembles of random thermodynamic processes. Unlike engineered systems, however, biological systems are highly regulated systems that are both robust and adaptable. While behaviors of single biological entities may be highly stochastic and difficult to predict, collective behaviors of an ensemble of these systems can result in coherent and well-defined high-level functionality. Examples include behaviors of insect and bacteria swarms, cell populations that sub-divide and specialize, and phenotypic multiplicity of multi-cellular organisms. These examples suggest that there exist some mechanism through which stochasticity can be exploited to increase the robustness and adaptability of today’s engineered systems. To understand such a mechanism is an ambitious engineering challenge and requires the exploration of various bio-inspired design approaches.

This workshop seeks to continue the exploration of emerging theories and methodologies that utilize stochasticity for various robotic applications. The objective is to bring together researchers from the different branches in robotics and develop new techniques to leverage stochasticity for the development of more robust and adaptable robotic systems, thus charting new directions for robotics research. The overarching aim of this workshop is to identify and explore common themes in the areas of robotics, cellular and evolutionary biology, neuroscience, and control and put into perspective the diverse research activities that address stochasticity in natural and artificial systems.

---