Swarmanoid project in swarm robotics Swarm robotics @ IDSIA


Recent work

Swarm robotics is the study of robotic systems consisting of a large group of relatively small and simple robots that interact and cooperate with each other in order to jointly solve tasks that are outside their own individual capabilities. Swarm robotic systems typically exhibit interesting properties such as high degrees of parallelism, redundancy, and robustness. They are also highly adaptive to changes in the environment, and show good scalability to increased problem and/or swarm size.

At IDSIA, we have been involved in two EU-funded projects related to swarm robotics: Swarm-bots (2001-2005) and Swarmanoid (2006-2010). Swarm-bots was concerned with the design, implementation and control of the s-bots: a swarm of small robots moving on a combination of tracks and wheels that can self-organize and self-assemble. Swarmanoid goes a step further, aiming at the development and control of a heterogeneous swarm consisting of three different types of robots: foot-bots, which move over the ground and have capabilities that are similar to those of the s-bots, eye-bots, which fly and have the capability to attach to the ceiling, and hand-bots, which have arms and grippers to manipulate objects and are able to climb in the vertical space using a rope.

Foot-bots forming a dynamic chain for navigation Foot-bot Foot-bots carrying hand-bot Eye-bots magnetically attached to the ceiling


The following video showing the concept and the results of the Swarmanoid project won the best video award at AAAI-11. See also the article in New Scientist.

More recently, two new projects have been approved. One is the NCCR on robotics: a large initiative funded by the Swiss National Science Foundation, involving many partners all around Switzerland. IDSIA participates to the Project cluster 4, on Distributed robotics. A summary of our research activities is shown in the following figure (poster presented at the 1st NCCR Robotics Symposium, Zurich, 16-17 Jun 2011).

Another project is SWARMIX, a Sinergia project, also funded by the Swiss NSF, in colaboration with the CSG group of Bernhard Plattner at ETHZ, the LIS lab of Dario Floreano at EPFL, and the Department of Ethology, of Adam Miklosi, at Eotvos Lorand University of Budapest. The project considers a mixed swarm of humans, dogs, and robots for cooperative search and rescue.

Copyright of www.swarmix.org


Recent work

  • Human-Swarm interaction through distributed cooperative gesture recognition

    In the context of the NCCR Robotics project we are focusing on symbiotic peer-to-peer interaction and cooperation between humans and robot swarms. As a first step, we considered human-swarm interaction, and selected the use of hand gestures to let a human communicate with a swarm of mobile robots. The purpose is to let a human communicating commands to be executed by the swarm (e.g., split in two groups). Hand gestures are a powerful and intuitive way to communicate, and do not require the use of additional devices. However, real-time vision-based recognition of hand gestures is a challenging task for the single robot, due to the limited processing power and field of view of robots that we use, the foot-bots. In this work, we investigated how to exploit robot mobility, swarm spatial distribution, and multi-hop wireless communications, to let the robots in the swarm: (i) implement a distributed and cooperative sensing of hand gestures based on a statistical classifier, and (ii) robustly reach a consensus about a gesture. The video below describes the system that we have implemented. A live implementation and demostration was also given at AAMAS 2012.

  • Distributed spatial clustering in robotic swarms

    The aim of this work is to endow robots in a swarm with awareness of their relative position with respect to the rest of the swarm. Such spatial awareness can be used to support spatially differentiated task allocation (e.g., split the swarm in different, spatially close, groups, and let each group engage in a different task, such as exploring different regions of an environment), or for pattern formation. The task we first focus on is to assign the robots of the swarm to two different classes, C0 and C1, in such a way that the two classes are spatially segregated: the robots in class C0 are found on one side of the swarm, and the robots in class C1 on the other side of the swarm. To solve this problem, we designed and implemented a distributed algorithm that is robust, scalable, efficient, works in a decentralized way, and has limited requirements in terms of available sensor or actuators. The algorithm only uses local, low-bandwidth communications. The algorithm combines elements from different sets of approaches to similar problems: algorithms for solving minimum bisection problems, algorithms for swarm robotics aggregation, and distributed algorithms for load balancing and distributed consensus filter. The videos below describes the system implemented on a swarm of 15 foot-bot robots. They make use of a quite unreliable line-of-sight wireless system (100 bytes/s) with a range of 1.5m. Videos are played at a speed twice faster than real.

  • Communication assisted navigation in robotic swarms

    Here we build on earlier work on the use of delay tolerant network communication to support robot navigation (see below). We present a communication based navigation algorithm for robotic swarms. It lets robots guide each other's navigation by exchanging messages containing navigation information through the wireless network formed among the swarm. We study the use of this algorithm in two different scenarios. In the first scenario, the swarm guides a single robot to a target, while in the second, all robots of the swarm navigate back and forth between two targets. In both cases, the algorithm provides efficient navigation, while being robust to failures of robots in the swarm. Moreover, we show that in the latter case, the system lets the swarm self-organize into a robust dynamic structure. This self-organization further improves navigation efficiency, and is able to find shortest paths in cluttered environments. We test our system both in simulation and on real robots. The tests with the real robots were performed by F. Ducatelle and G. Di Caro. We used the foot-bot robots, developed at the EPFL LSRO lab by M. Bonani, S. Magnenat, P. Retornaz and F. Mondada.

  • Self-organized cooperation in a heterogeneous robotic swarm

    We study self-organized cooperation in a heterogeneous robotic swarm consisting of two sub-swarms. The robots of each sub-swarm play distinct roles based on their different characteristics. We investigate how the swarm as a whole can solve complex tasks through a self-organized process based on local interactions between the sub-swarms. We focus on an indoor navigation task, in which we use a swarm of wheeled robots, called foot-bots, and a swarm of flying robots that can attach to the ceiling, called eye-bots. Foot-bots have to move back and forth between a source and a target location. Eye-bots are deployed in stationary positions against the ceiling, with the goal of guiding foot-bots. We study how the combined system can find efficient paths through a cluttered environment in a distributed way. The key component of our approach is a process of mutual adaptation, in which foot-bots execute instructions given by eye-bots, and eye-bots observe the behavior of foot-bots to adapt the instructions they give. The system is based on pheromone mediated navigation of ant colonies, as eye-bots function as stigmergic markers for foot-bots. Through simulation, we show that the system finds feasible paths in cluttered environments, converges onto the shortest of two paths, and spreads over different paths in case of congestion.

  • Mobile stigmergic markers for navigation in a heterogeneous robotic swarm

    We study self-organized navigation in a heterogeneous robotic swarm consisting of two types of robots: small wheeled robots, called foot-bots, and flying robots that can attach to the ceiling, called eye-bots. The task of foot-bots is to navigate back and forth between a source and a target location. The eye-bots are placed in a chain on the ceiling, connecting source and target using infrared communication. Their task is to guide foot-bots, by giving local directional instructions. The problem we address is how the positions of eye-bots and the directional instructions they give can be adapted, so that they indicate a path that is efficient for foot-bot navigation, also in the presence of obstacles. We propose an approach of mutual adaptation between foot-bots and eye-bots. Our solution is inspired by pheromone based navigation of ants, as eye-bots serve as mobile stigmergic markers for foot-bot navigation.

  • Supporting Navigation in Multi-Robot Systems through Network Communication

    We study a problem of navigation in networked multi-robot systems. The robots are deployed in a confined area, where they move around and solve tasks. They communicate with each other through an infrared communication device, so that an ad hoc network is formed among them. Due to the limited range and line of sight nature of the infrared communication, this network has intermittent connectivity. The question we address is how a particular robot can use this network to find a target location that is indicated by another robot (e.g., the other robot has identified a task to be serviced by the searching robot). All other robots are involved in tasks of their own, and do not change their movements to help the searching robot find its destination. However, they do offer support by forwarding messages over the network. We propose a new algorithm based on routing in ad hoc and delay tolerant networks that can run on the network formed between the robots and provide navigation information to the searching robot. We evaluate the validity of our approach both in simulation and through an implementation on a group of 16 e-puck robots. In previous work, we faced the same navigation problem using a network routing approach, establishing and using routing paths to gather navigation information for he robots.

  • Distributed motion planning for ground objects using a network of robotic ceiling cameras

    We study a distributed approach to path planning. We focus on holonomic kinematic motion in cluttered 2D areas. The problem consists in defining the precise sequence of roto-translations of a rigid object of arbitrary shape that has to be transported from an initial to a final location through a large, cluttered environment. Our planning system is implemented as a swarm of flying robots that are initially deployed in the environment and take static positions at the ceiling. Each robot is equipped with a camera and only sees a portion of the area below. Each robot acts as a local planner: it calculates the part of the path relative to the area it sees, and exchanges information with its neighbors through a wireless connection. This way, the robot swarm realizes a cooperative distributed calculation of the path. The path is communicated to ground robots, which move the object. We introduce a number of strategies to improve the system's performance in terms of scalability, resource efficiency, and robustness to alignment errors in the robot camera network. We report extensive simulation results that show the validity of our approach, considering a variety of object shapes and environments. We also validated the proposed approach on a set of experiments in a real setup. The holonomic object moving on the ground is implemented through a set of 2 non-holonomic robots, the e-pucks, interconnected by a rigid structure. In this way, they form an object with a relatively large shape, which is able to rotate and move in any direction. The size of the moving area is 33 m2. The multi-robot system on the ceiling is implemented with a set of 4 cameras connected to different computers. Each camera is controlled by an independent process, which cooperates and communicates with the other processes, locally plans the path, and then directs the navigation of the e-puck system through the ground area under its local field of view. The videos below show an example of path planning and movement execution (the camera logo image shows the camera which is currenly in charge to drive the robots).

  • Humans and robots sharing space: safe, human-like robot navigation

    We aim to let humans and robots sharing the same physical spaces with minimal mutual interference between each other. The most basic scenario in this respect regards multi-robot navigation in environments populated (also) by human beings. In order to achieve our objective, we want robots being able to move in the environment in a way which is safe (for both robots and humans), effective (moving trajectories should be smooth and close to shortest paths, given the spatial displacement of obstacles and other human/robot entities), and socially acceptable (humans should not perceive robots as potentially dangerous and/or as unpredictable entities). In order to achieve these goals, we started from the work of G. Theraulaz and his collaborators about the indentification of the rules that determine the walking behaviour of pedestrian social groups:

    Based on the models presented in the papers, we derived an adapted an algorithm for multi-robot robot navigation that has the characteristics that we want. The algorithm was studied both in simulation and implemented on our foot-bot robots (in spite of their poor vision and processing capabilities). In the videos below we show the behavior in some selected test scenarios including only robots. Soon we will make tests also including humans walking around. In the videos, each robot selects a target destination (based on color bands), lights up its beacon LED with the same color to indicate where it is going, searches the environment for the target using its on-board camera, and then moves towards target (keeping using the camera for target visual recognition). Once the target is reached, a new target is selected and the process goes on. Robots need to find their optimized way (in terms of traveling time and traveled distance) to each target while avoiding to bump into each other. In order to adaptively define their local mobility, robots locally exchange with each other their relative positions and instantaneous moving directions (using foot-boots' infrared-based range and bearing system that allows the robots to know and exchange this information through a line-of-sight wireless channel with a bandwidth of only 100 bytes/s). This information is used by the navigation model precisely let the robot move in a way which is similar to the way a human would move in the same situation. The first video is a short mix of all the other videos. Each video is based on a different scenario regarding target positioning. The last video shows a simulation with 20 foot-bots, implemented using the ARGoS multi-robot simulator.

  • A robot for a friend, aid to learning using pictogram-based communications

    We designed a system to let a child with communicative and/or motor disabilities interact with an e-puck mobile robot. The main purpose is to facilitate the acquisition of basic experiences that are necessary for child's pedagogical development.
    The work has been inspired by Simon Papert's microworlds and the frameworks of costructivist pedagogy and alternative and augmentative communication. The playground lets a child make decisions, collect experiences and expand her/his communication skills through the interaction with a robot that has the ability, like him/her, to understand pictogram-based communications.
    The demonstration video contains the model of a small grid-based town. A person gives orders to the robot by composing with pictograms a message on the wall. The robot reads it and executes the task. The language grammar is very simple but can still be used to provide different learning experiences.
    This work has been carried out by Jérôme Guzzi with the collaboration (and funding) of Gabriele Scascighini of CID/FIPPD (Centro Informatica Disabilità / Fondazione Informatica per la Promozione della Persona Disabile).

Full list of publications in swarm robotics (and related topics)

Page maintained by Gianni A. Di Caro