Stanislav Funiak

Distributed Localization of Modular Robot Ensembles

Internal localization, the problem of estimating relative pose for each module of a modular robot, is a prerequisite for many shape control, locomotion, and actuation algorithms. In this paper, we propose a robust hierarchical approach that uses normalized cut to identify dense sub-regions with small mutual localization error, then progressively merges those sub-regions to localize the entire ensemble. Our method works well in both two and three dimensions, and requires neither exact measurements nor rigid inter-module connectors. Most of the computations in our method can be distributed effectively. The result is a robust algorithm that scales to large ensembles. We evaluate our algorithm in two- and three-dimensional simulations of scenarios with up to 10,000 modules.

Multi-Modal Particle Filtering for Hybrid Systems with Autonomous Mode Transitions

Abstract Model-based diagnosis of embedded systems relies on the ability to estimate their hybrid state from noisy observations. This task is especially challenging for systems with many state variables and autonomous transitions. We propose a fair sampling algorithm that combines Rao-Blackwellised particle filters with a multi-modal Gaussian representation. In order to handle autonomous transitions, we let the continuous state estimates contribute to the proposal distribution in the particle filter. The algorithm outperforms purely simulational particle filters and provides unification of particle filters with hybrid hidden Markov model (HMM) observers.

Claytronics: highly scalable communications, sensing, and actuation networks

We propose a demonstration of extremely scalable modular robotics algorithms developed as part of the Claytronics Project (http://www-2.cs.cmu.edu/~claytronics/), as well as a demonstration of proof-of-concept prototypes. Our effort envisions multi-million-module robot ensembles able to morph into three-dimensional scenes, eventually with sufficient fidelity so as to convince a human observer the scenes are real. Although this work is potentially revolutionary in the sense that it holds out the possibility of radically altering the relationship between computation, humans, and the physical world, many of the research questions involved are similar in flavor to more mainstream systems research, albeit larger in scale. For instance, as in sensor networks, each robot will incorporate sensing, computation, and communications components. However, unlike most sensor networks each robot will also include mechanisms for actuation and motion. Many of the key challenges in this project involve coordination and communication of sensing and actuation across such large ensembles of independent units.

A combined stochastic and greedy hybrid estimation capability for concurrent hybrid models with autonomous mode transitions

Probabilistic hybrid discrete/continuous models, such as Concurrent Probabilistic Hybrid Automata (CPHA) are convenient tools for modeling complex robotic systems. In this paper, we present a novel method for estimating the hybrid state of CPHA that achieves robustness by balancing greedy and stochastic search. To accomplish this, we (1) develop an efficient stochastic sampling approach for CPHA based on Rao-Blackwellised Particle Filtering, (2) perform an empirical comparison of the greedy and stochastic approaches to hybrid estimation and (3) propose a strategy for mixing stochastic and greedy search. The resulting method handles nonlinear dynamics, concurrently operating components and autonomous mode transitions. We demonstrate the robustness of the mixed method empirically.