Lyuba Alboul

A Robot Swarm assisting a Human Fire Fighter

Emergencies in industrial warehouses are a major concern for fire-fighters. The large dimensions, together with the development of dense smoke that drastically reduces visibility, represent major challenges. The GUARDIANS robot swarm is designed to assist fire-fighters in searching a large warehouse. In this paper we discuss the technology developed for a swarm of robots assisting fire-fighters. We explain the swarming algorithms that provide the functionality by which the robots react to and follow humans while no communication is required. Next we discuss the wireless communication system, which is a so-called mobile ad-hoc network. The communication network provides also the means to locate the robots and humans. Thus, the robot swarm is able to provide guidance information to the humans. Together with the fire-fighters we explored how the robot swarm should feed information back to the human fire-fighter. We have designed and experimented with interfaces for presenting swarm-based information to human beings.

Mixed human-robot team navigation in the GUARDIANS project

A theoretical framework for generating navigation behaviour patterns in mixed human-robot groups in complex environments is proposed. This framework represents an essential part in the development of a multiple robot-human system for assisting fire-fighters in search and rescue operations in the GUARDIANS project. In order to produce the desired behaviours an artificial potential field method has been developed. We distinguish a three classes of agents: robots, humans and obstacles, and apply different potential functions to them. Depending on the situation, we switch from one function to another; this allows to generate desired behaviour patterns as well as to avoid certain local minima. Typical behaviour patterns are singled out and their stability is discussed. Stability analysis is based on geometric considerations, that permits to avoid bulky computations and provide graphic demonstrations of convergence. The proposed framework can be used in other robotic applications where a group of heterogenous agents is deployed.

Optimising triangulated polyhedral surfaces with self-intersections

We discuss an optimisation procedure for triangulated polyhedral surfaces (referred to as (2-3)D triangulations) which allows us to process self–intersecting surfaces. As an optimality criterion we use minimisation of total absolute extrinsic curvature (MTAEC) and as a local transformation – a diagonal flip, defined in a proper way for (2-3)D triangulations. This diagonal flip is a natural generalisation of the diagonal flip operation in 2D, known as Lawson’s procedure. The difference is that the diagonal flip operation in (2-3)D triangulations may produce self-intersections. We analyze the optimisation procedure for (2-3)D closed triangulations, taking into account possible self–intersections. This analysis provides a general insight on the structure of triangulations, allows to characterise the types of self–intersections, as well as the conditions for global convergence of the algorithm. It provides also a new view on the concept of optimisation on the whole and is useful in the analysis of global and local convergence for other optimisation algorithms. At the end we present an efficient implementation of the optimality procedure for (2-3)D triangulations of the data, situated in the convex position, and conjecture possible results of this procedure for non–convex data.

On flips in polyhedral surfaces

Let V be a finite point set in 3-space, and let S(V) be the set of triangulated polyhedral surfaces homeomorphic to a sphere and with vertex set V. Let abc and cbd be two adjacent triangles belonging to a surface SeS(V); the flip of the edge bc would replace these two triangles by the triangles abd and adc. The flip operation is only considered when it does not produce a self-intersecting surface. In this paper we show that given two surfaces S1,S2eS(V), it is possible that there is no sequence of flips transforming S1 into S2, even in the case that V consists of points in convex position.

Challenges of the Multi-robot Team in the GUARDIANS Project

The Guardians multi-robot team is being designed to be deployed in a large warehouse to assist firefighters in the event or danger of a fire. The large size of the environment together with development of dense smoke that drastically reduces visibility, represent major challenges in search and rescue operations. The GUARDIANS robots act alongside a firefighter and should provide, among others, the following tasks: to guide or accompany the firefighters on the site while indicating possible obstacles and locations of danger and maintaining communications links. In order to fulfill the aforementioned tasks the robots need to be able to exert certain behaviours. Among the basic behaviours are capabilities to unite in a group - generate a formation - and navigate on the site while keeping this formation. The basic control model used to generate these behaviours is based on the so-called social potential field framework, which we adapt to fulfill specific tasks of the Guardians scenario. All of the tasks can be achieved without central control, and some of the tasks can be performed even without explicit communication among the robots. We discuss advantages and shortcomings of our model and present samples of implementation on ERA-MOBI robots, commonly referred to as Erratics.

Experiments in cooperative human multi-robot navigation

In this paper, we consider the problem of a group of autonomous mobile robots and a human moving coordinately in a real-world implementation. The group moves throughout a dynamic and unstructured environment. The key problem to be solved is the inclusion of a human in a real multi-robot system and consequently the multiple robot motion coordination. We present a set of performance metrics (system efficiency and percentage of time in formation) and a novel flexible formation definition whereby a formation control strategy both in simulation and in real-world experiments of a human multi-robot system is presented. The formation control proposed is stable and effective by means of its uniform dispersion, cohesion and flexibility.

Towards Autonomous Robotic Systems

Robot with non-back-drivable actuators will appear stiff when in contact with the environment and human. This scenario is unsafe for the Human-Robot Interaction (HRI). In order to guarantee safety in HRI, the robot will be made “soft” such that a compliant control can be introduced. Apart from utilizing the proper mechanism design, the back drivability actuators can be achieved by a suitable choice of control. In particular, in this paper, a PID control is employed to achieve an active compliance control. The reference impedance model characteristics are exploited for which the system allows us to introduce a virtual mass-spring-damper system to adjust the compliant control level. The performance of the PID control will be tested on the RED Hand in the simulation. The results are recorded and analyzed for the thumb finger. The results show that the PID controller is capable of controlling the motion and position of the RED Hand. In addition, the compliance behavior for the RED Hand can be suitably adjusted based on the required compliant level.

An approach to multi-robot site exploration based on principles of self-organisation

This paper presents a novel approach to multi-robot site exploration and map building considering the robot team as a self-organising system. The approach has been developed within the framework of the project GUARDIANS. The Map Building process represents not a separate activity, but an inherent byproduct of self-organisation. The system consists of an (heterogeneous) robot swarm, a mobile ad-hoc network and an (evolving) topological map of the environment. The proposed map building approach takes advantage of a cooperating robot team (as opposed to a single robot) allowing accurate deployment and localisation in a structured, yet dynamic manner. A topological graph representation of the environment is formed, from which an initial metric representation is elicitable as edges are assigned lengths. This reasonable sketch of the environment can be further developed to a full metric map and be used as the basis of building ad-hoc mobile wireless communication and sensor networks. The presented algorithms also take into consideration sensor limitation and are tested on a group of Khepera III robots, specially upgraded to fulfil the needs of our approach.

Polyhedral gauss maps and curvature characterisation of triangle meshes

We design a set of algorithms to construct and visualise unambiguous Gauss maps for a large class of triangulated polyhedral surfaces, including surfaces of non-convex objects and even non-manifold surfaces. The resulting Gauss map describes the surface by distinguishing its domains of positive and negative curvature, referred to as curvature domains. These domains are often only implicitly present in a polyhedral surface and cannot be revealed by means of the angle deficit. We call the collection of curvature domains of a surface the Gauss map signature. Using the concept of the Gauss map signature, we highlight why the angle deficit is sufficient neither to estimate the Gaussian curvature of the underlying smooth surface nor to capture the curvature information of a polyhedral surface. The Gauss map signature provides shape recognition and curvature characterisation of a triangle mesh and can be used further for optimising the mesh or for developing subdivision schemes.

A system for reconstruction from point clouds in 3D: Simplification and mesh representation

In this paper we present a complete system for acquisition of fused (textured) point clouds in 3D, from a Laser Range Finder (LRF) and a CCD camera. Furthermore, we describe an approach to build and process the resulting models, including their pre-processing and mesh simplification. This approach allows manipulating the resulting data structure into consistent geometric representations, which can be further adapted based on user requirements. The advantage of our system is that of low computational cost, ease of use and accuracy in the representation of the environment, even without prior data smoothing.