Several geometric criteria to fit a polygonal closed curve to discrete two-dimensional data are c... more Several geometric criteria to fit a polygonal closed curve to discrete two-dimensional data are considered and analysed. Most of these criteria are related to the concept of curvature, for example, one criterion is minimisation of total absolute curvature. On the basis of these criteria algorithms to construct a polygonal curve which is optimal with respect to a specific criterion, are designed.
ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regu... more ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regular or scattered). The first problem which we need to confront, is to make a choice of a suitable triangulation. It is well-known that a triangulation strongly influences the quality of a resulting surface. The most typical questions concerning a triangulation are the following ones: Is it possible to determine an appropriate triangulation? How does the shape of a surface influence the choice of triangulations? The triangulation itself can be considered as the simplest C 0 -approximation, or in other words, a polyhedral approximation. We discuss advantages and disadvantages of polyhedral approximations, obtained from possible triangulations, from a geometric point of view. We use the notion of manifolds with polyhedral metrics (or, simply, polyhedra), i.e., metric spaces, every point of which has a neighbourhood isometric to some neighbourhood of a two-dimensional cone vertex. Making use of the polyhedral analogues of well-known concepts of classical differential geometry, such as the Gaussian curvature, Mean curvature etc., we propose a pure geometric approach for choosing the “best” (in a certain sense) triangulation of a given 3-D dataset.
We describe here a rigorous and accurate model for the simulation of three-dimensional deformable... more We describe here a rigorous and accurate model for the simulation of three-dimensional deformable particles (DPs). The method is very versatile, easily simulating various types of deformable particles such as vesicles, capsules, and biological cells. Each DP is resolved explicitly and advects within the surrounding Newtonian fluid. The DPs have a preferred rest shape (e.g., spherical for vesicles, or biconcave for red blood cells). The model uses a classic hybrid system: an Eulerian approach is used for the Navier-Stokes solver (the lattice Boltzmann method) and a Lagrangian approach for the evolution of the DP mesh. Coupling is accomplished through the lattice Boltzmann velocity field, which transmits force to the membranes of the DPs. The novelty of this method resides in its ability (by design) to simulate a large number of DPs within the bounds of current computational limitations: our simple and efficient approach is to (i) use the lattice Boltzmann method because of its acknowledged efficiency at low Reynolds number and its ease of parallelization, and (ii) model the DP dynamics using a coarse mesh (approximately 500 nodes) and a spring model constraining (if necessary) local area, total area, cell volume, local curvature, and local primary stresses. We show that this approach is comparable to the more common-yet numerically expensive-approach of membrane potential function, through a series of quantitative comparisons. To demonstrate the capabilities of the model, we simulate the flow of 200 densely packed red blood cells-a computationally challenging task. The model is very efficient, requiring of the order of minutes for a single DP in a 50 μm×40 μm×40 μm simulation domain and only hours for 200 DPs in 80 μm×30 μm×30 μm. Moreover, the model is highly scalable and efficient compared to other models of blood cells in flow, making it an ideal and unique tool for studying blood flow in microvessels or vesicle or capsule flow (or a mixture of different particles). In addition to directly predicting fluid dynamics in complex suspension in any geometry, the model allows determination of accurate, empirical rules which may improve existing macroscopic, continuum models.
ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regu... more ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regular or scattered). The first problem which we need to confront, is to make a choice of a suitable triangulation. It is well-known that a triangulation strongly influences the quality of a resulting surface. The most typical questions concerning a triangulation are the following ones: Is it possible to determine an appropriate triangulation? How does the shape of a surface influence the choice of triangulations? The triangulation itself can be considered as the simplest C 0 -approximation, or in other words, a polyhedral approximation. We discuss advantages and disadvantages of polyhedral approximations, obtained from possible triangulations, from a geometric point of view. We use the notion of manifolds with polyhedral metrics (or, simply, polyhedra), i.e., metric spaces, every point of which has a neighbourhood isometric to some neighbourhood of a two-dimensional cone vertex. Making use of the polyhedral analogues of well-known concepts of classical differential geometry, such as the Gaussian curvature, Mean curvature etc., we propose a pure geometric approach for choosing the “best” (in a certain sense) triangulation of a given 3-D dataset.
ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regu... more ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regular or scattered). The first problem which we need to confront, is to make a choice of a suitable triangulation. It is well-known that a triangulation strongly influences the quality of a resulting surface. The most typical questions concerning a triangulation are the following ones: Is it possible to determine an appropriate triangulation? How does the shape of a surface influence the choice of triangulations? The triangulation itself can be considered as the simplest C 0 -approximation, or in other words, a polyhedral approximation. We discuss advantages and disadvantages of polyhedral approximations, obtained from possible triangulations, from a geometric point of view. We use the notion of manifolds with polyhedral metrics (or, simply, polyhedra), i.e., metric spaces, every point of which has a neighbourhood isometric to some neighbourhood of a two-dimensional cone vertex. Making use of the polyhedral analogues of well-known concepts of classical differential geometry, such as the Gaussian curvature, Mean curvature etc., we propose a pure geometric approach for choosing the “best” (in a certain sense) triangulation of a given 3-D dataset.
Journal of Computational and Applied Mathematics, 2000
When reconstructing a surface from irregularly spaced data we need to decide how to identify a go... more When reconstructing a surface from irregularly spaced data we need to decide how to identify a good triangulation. As a measure of quality we consider various differential geometrical properties, namely integral absolute Gaussian curvature, integral absolute mean curvature and area. A comparison is made with data-dependent triangulation methods that exist in the literature.
When reconstructing a surface from irregularly spaced data we need to decide how to identify a go... more When reconstructing a surface from irregularly spaced data we need to decide how to identify a good triangulation. As a measure of quality we consider various differential geometrical properties, namely integral absolute Gaussian curvature, integral absolute mean curvature and area. A comparison is made with data-dependent triangulation methods that exist in the literature.
Emergencies in industrial warehouses are a major concern for fire-fighters. The large dimensions,... more 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.
In this paper, we consider the problem of a group of autonomous mobile robots and a human moving ... more 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.
A novel graph-based approach to search in unknown environments is presented. A virtual geometric ... more A novel graph-based approach to search in unknown environments is presented. A virtual geometric structure is imposed on the environment represented in computer memory by a graph. Algorithms use this representation to coordinate a team of robots (or entities). Local discovery of environmental features cause dynamic expansion of the graph resulting in global exploration of the unknown environment. The algorithm is shown to have O(k · n H ) time complexity, where n H is the number of vertices of the discovered environment and 1 ≤ k ≤ n H . A maximum bound on the length of the resulting walk Ω is given.
The Guardians multi-robot team is being designed to be deployed in a large warehouse to assist fi... more 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.
An algorithm for improving the quality of an initial triangulation on a fixed set of vertices is ... more An algorithm for improving the quality of an initial triangulation on a fixed set of vertices is suggested. The edge flip operation is performed consecutively, aiming to minimise the discrete Willmore energy over a triangulated surface (or mesh). The Willmore energy of a surface is a function of Gaussian and mean curvature, and measures local deviation from a sphere. Virtual points are introduced in the triangulation to overcome the local invariance of Willmore energy under edge flips. Some experimental results are given.
This paper presents a novel approach to multi-robot site exploration and map building considering... more 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 by-product 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.
The research reported in this paper is part of a project investigating distributed control archit... more The research reported in this paper is part of a project investigating distributed control architectures for groups of autonomous robots. The focus of this paper is on modelling interaction patterns occurring in robot group behaviour. As such, we do not focus on defining control architectures for individual robots, instead, we focus on individual behaviour patterns to develop a formal theory of group behaviour resulting from multiple robot interaction. The problem underlying the research is that concepts relating to group behaviour have to be imposed upon the robots and the understanding of these patterns will lead to more efficient control and the realisation of cooperative tasks that would not be possible otherwise. The paper considers a balanced conflict in a system of three robots where action comes to a halt and a slightly deviating conflict in which action is continued in a predictable pattern. We prove how the group behaves in both types of conflicts. We introduce practical constraints of robot design such as limitations of a sensory system and discuss simulations of both types of conflicts incorporating these constraints. Having proved the behaviour of the group starting from these conflicts, the conflicts might be used to test and calibrate robots; we discuss the constraints to meet.
The GUARDIANS multi-robot team is to be deployed in a large warehouse in smoke. The team is to as... more The GUARDIANS multi-robot team is to be deployed in a large warehouse in smoke. The team is to assist firefighters search the warehouse in the event or danger of a fire. The large dimensions of the environment together with development of smoke which drastically reduces visibility, represent major challenges for search and rescue operations. The GUARDIANS robots guide and accompany the firefighters on site whilst indicating possible obstacles and the locations of danger and maintaining communications links.
A theoretical framework for generating navigation behaviour patterns in mixed human-robot groups ... more 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.
Several geometric criteria to fit a polygonal closed curve to discrete two-dimensional data are c... more Several geometric criteria to fit a polygonal closed curve to discrete two-dimensional data are considered and analysed. Most of these criteria are related to the concept of curvature, for example, one criterion is minimisation of total absolute curvature. On the basis of these criteria algorithms to construct a polygonal curve which is optimal with respect to a specific criterion, are designed.
ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regu... more ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regular or scattered). The first problem which we need to confront, is to make a choice of a suitable triangulation. It is well-known that a triangulation strongly influences the quality of a resulting surface. The most typical questions concerning a triangulation are the following ones: Is it possible to determine an appropriate triangulation? How does the shape of a surface influence the choice of triangulations? The triangulation itself can be considered as the simplest C 0 -approximation, or in other words, a polyhedral approximation. We discuss advantages and disadvantages of polyhedral approximations, obtained from possible triangulations, from a geometric point of view. We use the notion of manifolds with polyhedral metrics (or, simply, polyhedra), i.e., metric spaces, every point of which has a neighbourhood isometric to some neighbourhood of a two-dimensional cone vertex. Making use of the polyhedral analogues of well-known concepts of classical differential geometry, such as the Gaussian curvature, Mean curvature etc., we propose a pure geometric approach for choosing the “best” (in a certain sense) triangulation of a given 3-D dataset.
We describe here a rigorous and accurate model for the simulation of three-dimensional deformable... more We describe here a rigorous and accurate model for the simulation of three-dimensional deformable particles (DPs). The method is very versatile, easily simulating various types of deformable particles such as vesicles, capsules, and biological cells. Each DP is resolved explicitly and advects within the surrounding Newtonian fluid. The DPs have a preferred rest shape (e.g., spherical for vesicles, or biconcave for red blood cells). The model uses a classic hybrid system: an Eulerian approach is used for the Navier-Stokes solver (the lattice Boltzmann method) and a Lagrangian approach for the evolution of the DP mesh. Coupling is accomplished through the lattice Boltzmann velocity field, which transmits force to the membranes of the DPs. The novelty of this method resides in its ability (by design) to simulate a large number of DPs within the bounds of current computational limitations: our simple and efficient approach is to (i) use the lattice Boltzmann method because of its acknowledged efficiency at low Reynolds number and its ease of parallelization, and (ii) model the DP dynamics using a coarse mesh (approximately 500 nodes) and a spring model constraining (if necessary) local area, total area, cell volume, local curvature, and local primary stresses. We show that this approach is comparable to the more common-yet numerically expensive-approach of membrane potential function, through a series of quantitative comparisons. To demonstrate the capabilities of the model, we simulate the flow of 200 densely packed red blood cells-a computationally challenging task. The model is very efficient, requiring of the order of minutes for a single DP in a 50 μm×40 μm×40 μm simulation domain and only hours for 200 DPs in 80 μm×30 μm×30 μm. Moreover, the model is highly scalable and efficient compared to other models of blood cells in flow, making it an ideal and unique tool for studying blood flow in microvessels or vesicle or capsule flow (or a mixture of different particles). In addition to directly predicting fluid dynamics in complex suspension in any geometry, the model allows determination of accurate, empirical rules which may improve existing macroscopic, continuum models.
ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regu... more ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regular or scattered). The first problem which we need to confront, is to make a choice of a suitable triangulation. It is well-known that a triangulation strongly influences the quality of a resulting surface. The most typical questions concerning a triangulation are the following ones: Is it possible to determine an appropriate triangulation? How does the shape of a surface influence the choice of triangulations? The triangulation itself can be considered as the simplest C 0 -approximation, or in other words, a polyhedral approximation. We discuss advantages and disadvantages of polyhedral approximations, obtained from possible triangulations, from a geometric point of view. We use the notion of manifolds with polyhedral metrics (or, simply, polyhedra), i.e., metric spaces, every point of which has a neighbourhood isometric to some neighbourhood of a two-dimensional cone vertex. Making use of the polyhedral analogues of well-known concepts of classical differential geometry, such as the Gaussian curvature, Mean curvature etc., we propose a pure geometric approach for choosing the “best” (in a certain sense) triangulation of a given 3-D dataset.
ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regu... more ABSTRACT This paper deals with surface modelling based on a given three-dimensional dataset (regular or scattered). The first problem which we need to confront, is to make a choice of a suitable triangulation. It is well-known that a triangulation strongly influences the quality of a resulting surface. The most typical questions concerning a triangulation are the following ones: Is it possible to determine an appropriate triangulation? How does the shape of a surface influence the choice of triangulations? The triangulation itself can be considered as the simplest C 0 -approximation, or in other words, a polyhedral approximation. We discuss advantages and disadvantages of polyhedral approximations, obtained from possible triangulations, from a geometric point of view. We use the notion of manifolds with polyhedral metrics (or, simply, polyhedra), i.e., metric spaces, every point of which has a neighbourhood isometric to some neighbourhood of a two-dimensional cone vertex. Making use of the polyhedral analogues of well-known concepts of classical differential geometry, such as the Gaussian curvature, Mean curvature etc., we propose a pure geometric approach for choosing the “best” (in a certain sense) triangulation of a given 3-D dataset.
Journal of Computational and Applied Mathematics, 2000
When reconstructing a surface from irregularly spaced data we need to decide how to identify a go... more When reconstructing a surface from irregularly spaced data we need to decide how to identify a good triangulation. As a measure of quality we consider various differential geometrical properties, namely integral absolute Gaussian curvature, integral absolute mean curvature and area. A comparison is made with data-dependent triangulation methods that exist in the literature.
When reconstructing a surface from irregularly spaced data we need to decide how to identify a go... more When reconstructing a surface from irregularly spaced data we need to decide how to identify a good triangulation. As a measure of quality we consider various differential geometrical properties, namely integral absolute Gaussian curvature, integral absolute mean curvature and area. A comparison is made with data-dependent triangulation methods that exist in the literature.
Emergencies in industrial warehouses are a major concern for fire-fighters. The large dimensions,... more 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.
In this paper, we consider the problem of a group of autonomous mobile robots and a human moving ... more 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.
A novel graph-based approach to search in unknown environments is presented. A virtual geometric ... more A novel graph-based approach to search in unknown environments is presented. A virtual geometric structure is imposed on the environment represented in computer memory by a graph. Algorithms use this representation to coordinate a team of robots (or entities). Local discovery of environmental features cause dynamic expansion of the graph resulting in global exploration of the unknown environment. The algorithm is shown to have O(k · n H ) time complexity, where n H is the number of vertices of the discovered environment and 1 ≤ k ≤ n H . A maximum bound on the length of the resulting walk Ω is given.
The Guardians multi-robot team is being designed to be deployed in a large warehouse to assist fi... more 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.
An algorithm for improving the quality of an initial triangulation on a fixed set of vertices is ... more An algorithm for improving the quality of an initial triangulation on a fixed set of vertices is suggested. The edge flip operation is performed consecutively, aiming to minimise the discrete Willmore energy over a triangulated surface (or mesh). The Willmore energy of a surface is a function of Gaussian and mean curvature, and measures local deviation from a sphere. Virtual points are introduced in the triangulation to overcome the local invariance of Willmore energy under edge flips. Some experimental results are given.
This paper presents a novel approach to multi-robot site exploration and map building considering... more 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 by-product 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.
The research reported in this paper is part of a project investigating distributed control archit... more The research reported in this paper is part of a project investigating distributed control architectures for groups of autonomous robots. The focus of this paper is on modelling interaction patterns occurring in robot group behaviour. As such, we do not focus on defining control architectures for individual robots, instead, we focus on individual behaviour patterns to develop a formal theory of group behaviour resulting from multiple robot interaction. The problem underlying the research is that concepts relating to group behaviour have to be imposed upon the robots and the understanding of these patterns will lead to more efficient control and the realisation of cooperative tasks that would not be possible otherwise. The paper considers a balanced conflict in a system of three robots where action comes to a halt and a slightly deviating conflict in which action is continued in a predictable pattern. We prove how the group behaves in both types of conflicts. We introduce practical constraints of robot design such as limitations of a sensory system and discuss simulations of both types of conflicts incorporating these constraints. Having proved the behaviour of the group starting from these conflicts, the conflicts might be used to test and calibrate robots; we discuss the constraints to meet.
The GUARDIANS multi-robot team is to be deployed in a large warehouse in smoke. The team is to as... more The GUARDIANS multi-robot team is to be deployed in a large warehouse in smoke. The team is to assist firefighters search the warehouse in the event or danger of a fire. The large dimensions of the environment together with development of smoke which drastically reduces visibility, represent major challenges for search and rescue operations. The GUARDIANS robots guide and accompany the firefighters on site whilst indicating possible obstacles and the locations of danger and maintaining communications links.
A theoretical framework for generating navigation behaviour patterns in mixed human-robot groups ... more 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.
Uploads
Papers by Lyuba Alboul