Automatic Gait Generation in Modular Robots: to Oscillate or to Rotate? that is the question

Modular robots offer the possibility to design robots with a high diversity of shapes and functionalities. This nice feature also brings an important challenge: namely how to design efficient locomotion gaits for arbitrary robot structures with many degrees of freedom. In this paper, we present a framework that allows one to explore and identify highly different gaits for a given arbitraryshaped modular robot. For this, we use simulated robots made of several Roombots modules that have three rotational joints each. These modules have as interesting feature that they can produce both oscillatory movements (i.e. periodic movements around a rest position) and rotational movements (i.e. with continuously increasing angle), leading to very rich locomotion patterns. Here we ask ourselves which types of movements — purely oscillatory, purely rotational, or a combination of both— lead to the fastest gaits. To address this question we designed a control architecture based on a distributed system of coupled phase oscillators that can produce synchronized rotations and oscillations in many degrees of freedom. We also designed a specific optimization algorithm that can automatically design hybrid controllers, i.e. controllers that use oscillations in some joints and rotations in others, for fast gaits. The proposed framework is verified through multiple simulations for several robot morphologies. The results show that (i) the question whether it is better to oscillate or to rotate depends on the morphology of the robot, and that in general it is best to do both, (ii) the optimization framework can successfully generate hybrid controllers that outperform purely oscillatory and purely rotational ones, and (iii) the resulting gaits are fast, innovative, and would have been hard to design by hand

Learning to move in modular robots using central pattern generators and online optimization

Auke Ijspeert, Jérôme Maye

This article addresses the problem of how modular robotics systems, i.e. systems composed of multiple modules that can be configured into different robotic structures, can learn to locomote. In particular, we tackle the problems of online learning, that is, learning while moving, and the problem of dealing with unknown arbitrary robotic structures. We propose a framework for learning locomotion controllers based on two components: a central pattern generator (CPG) and a gradient-free optimization algorithm: Powell’s method. The CPG is implemented as a system of coupled nonlinear oscillators in our YaMoR modular robotic system, with one oscillator per module. The nonlinear oscillators are coupled together across modules using Bluetooth communication to obtain specific gaits, i.e. synchronized patterns of oscillations among modules. Online learning is done by running the Powell optimization algorithm in parallel to the CPG model, with the speed of locomotion being the criterion to be optimized. Interesting aspects of the optimization are: it is carried out online, it does not require stopping or resetting the robots, and it is fast. We present results showing the interesting properties of this framework for a modular robotic system. In particular, our CPG model can readily be implemented in a distributed system, it is cheap computationally, it exhibits limit cycle behavior (temporary perturbations are rapidly forgotten), it produces smooth trajectories even when control parameters are abruptly changed, and it is robust against imperfect communication among modules. We also present results of learning to move with three different robot structures. Interesting locomotion modes are obtained after running the optimization for less than 60 minutes.

2008

Synthesis, modeling, and experimental validation of distributed robotic search

James Pugh

The work of this thesis centers around the research subject of distributed robotic search. Within the field of distributed robotic systems, a task of particular interest is attempting to locate one or more targets in a possibly unknown environment. While numerous studies have proposed and analyzed methods to accomplish this, there is currently a lack of a strong foundation of tools and techniques that can be used to facilitate the development and evaluation of different approaches to distributed robotic search. In this work, we aim to provide such a foundation through tools, methods, models, and analysis of experimentation. An often overlooked aspect of the distributed robotic research process is the development and analysis of tools and modules to be used with robotic systems. These may include plug-ins for realistic robotic simulators, software/hardware systems to track multiple mobile robots in real-time, extension boards for robotic platforms, and the robots themselves. Along with other tools developed and used for this work, we focus particularly on the development, characterization, and validation of a fast, accurate on-board system for relative positioning and communication between robots, a capability which is critical for effective distributed robotic search. Designing individual robot controllers to generate a specific group behavior is a difficult and often counter-intuitive process. A possible alternative to hand-crafting distributed search controllers is automatic synthesis using machine-learning techniques. We explore the effectiveness of using a noise-resistant version of the Particle Swarm Optimization algorithm to optimize the weights of an embedded Artificial Neural Network, allowing the robot to learn obstacle avoidance behavior (a common benchmark for robotic learning techniques); we find that this technique appears to offer superior performance as compared to the canonical approach of using Genetic Algorithms for this type of learning. A method for faster learning using distributed evaluation in a robot team is tested and is found to offer comparable performance using only a fraction of the original learning time. This technique can be used for fast, effective learning and adaptation in a distributed robotic system performing search. The process of designing and analyzing algorithms for distributed robotic systems can be greatly facilitated if models are available to describe the dynamics of the algorithm at an abstract level. Inspired by previous examples in the distributed robotic field, we work to design a model of robotic search that captures the system at different levels of abstraction, ranging from accurately recreating the details of individual robots to describing the entire system as an indivisible whole. To capture the entire search process, we model both the exploration phase, where robots cover an environment in an effort to detect traces of targets, and the localization phase, where robots use target emission sensing to navigate towards the target. The utility of our models is demonstrated by using them to develop an effective technique for the declaration phase of search, where robots decide that a target has been accurately localized and announce its position. In distributed robotics research, it is important that techniques developed with abstracted simulations and models are ultimately validated using real robotic platforms in order to verify their correctness. In that spirit, we run systematic sound search experiments using teams of up to ten real robots. These experiments utilize the tools developed throughout the research process, demonstrate the utility of our learning technique for fast search adaptation, and serve to validate our models of distributed robotic localization. In addition, they allow us to analyze and better understand the subtle dynamics of the search process, providing information which should be useful for any future work on distributed robotic search.

EPFL2008

Design, Modeling and Optimization of Stochastic Reactive Distributed Robotic Systems

Grégory Mermoud

This dissertation describes a complete methodological framework for designing, modeling and optimizing a specific class of distributed systems whose dynamics result from the multiple, stochastic interactions of their constitutive components. These components can be robots endowed with very minimal capabilities, or even simpler entities such as insects, bacteria, particles, or molecules. We refer to such components as Smart Minimal Particles (SMPs). One of the main difficulties facing the modeling of SMPs is the potential complexity and richness of their dynamics. On the one hand, one needs detailed models that account for the physico-chemical properties of the lower-level components (e.g., shape, material, surface chemistry, charge, etc.), which, in turn, determine the nature and the magnitude of their interactions. On the other hand, one is also interested in models that can yield accurate numerical predictions of macroscopic quantities, and investigate formally their dependence on the system’s design and control parameters. These competing requirements motivate a combination of models at multiple levels of abstraction, as advocated by the Multi-Level Modeling Methodology (MLMM), which was introduced in prior works. The MLMM enables the fulfillment of both requirements in a very efficient way by incrementally building up models at increasing levels of abstraction in order to capture the relevant features of the system. This thesis extends and consolidates the MLMM along several axes. In a first step, we provide a theoretical consolidation of the MLMM. We propose a thorough classification of the different models of SMPs, and we discuss their underlying assumptions and simplifications. We shed light on the fundamental impact of embodiment and spatiality on models’ accuracy, and we define the conditions under which the macro-deterministic approximation is valid. These theoretical considerations are experimentally supported by five case studies of aggregation and Self-Assembly (SA) at different scales. The five case studies utilize three types of components: (i) miniature wheeled robots (Alice, 2 cm in size) endowed with limited computation, sensing, actuation, and communication capabilities, (ii) water-floating passive modules (Lily, 3 cm in size) endowed with four permanent magnets for mutual latching, and (iii) micro-fabricated cylinders (about 100 μm in diameter, studied in realistic simulation only) that can achieve SA in liquids. In a second step, we introduce the core contribution of this thesis, that is, a systematic and generic methodology for bridging the gap between real, physical systems and computationally efficient models at multiple abstraction levels. In particular, we describe the M3 computational framework, which enables the automated construction of models of SMPs. Our approach consists in observing (or simulating realistically) a system of interest, and building a hierarchical suite of models based on the observations (i.e., trajectories) collected during these experiments (or simulations). Internally, the framework first builds up a microscopic representation of the system based on these observations and on a list of interactions of interest specified by the user. This representation, called the Canonical Microscopic Model (CMM), is a formal and generic description of SMPs, and it serves as a blueprint for the construction of a macroscopic model, specified using the Chemical Reaction Network (CRN) formalism. The rates of the CRN are finally calibrated using a Maximum Likelihood Estimation (MLE) scheme. We validate the M3 framework on each of the three platforms discussed earlier, thereby illustrating its relevance both as a modeling and as an analysis tool. Finally, we discuss the role of multi-level modeling when designing and optimizing SMPs. In particular, we show that top-down model-based design of multi-robot systems is generally not amenable to efficient implementations when dealing with very resource-constrained robots. Instead, faithful and computationally efficient models built incrementally from the bottom up prove to be an essential tool for designing such systems. We further corroborate this claim by applying our automated modeling framework to the real-time control of the stochastic SA of Lily modules. Our scientific contribution is therefore three-fold. First, we provide a solid experimental and theoretical consolidation of the MLMM, which has been the subject of intense research efforts for the last decade. Second, we propose, for the first time, an approach to generate models at high abstraction level in a completely automated fashion, based solely on observations of the system of interest. Third, we provide deep insights into the modeling and the design of SMPs, with a specific emphasis on self-assembling systems ranging from the centimeter scale down to the micrometer scale.