Search results

The purpose of this paper is to review some of the various worldwide projects to develop and apply innovative swarm-type robots to many challenging applications.

An in-depth review of published information and interviews with researchers and developers of swarm robot technology were conducted.

Swarm robots continue to be developed to match an ever-increasing number of interesting and innovative applications.

Readers may be very surprised at the tasks that autonomous swarm robots can address and the developments that are underway to further extend the abilities of swarm robots.

This paper is a review of a wide range of the latest swarm robot developments, innovations and applications.

The purpose of this paper is to propose a distributed unmanned aerial vehicle (UAV) swarm control method to ensure safety and obstacle avoidance during swarm flight and realize effective guidance.

This paper proposes a distributed UAV swarm control framework with limited interaction. UAVs in the swarm realize the selection of limited interactive neighbors according to the random line of sight and limited field of view. The designed interaction force and obstacle avoidance mechanism are combined to ensure the safety of UAVs and avoid collisions and obstacles. Informed UAVs are deployed to guide the swarm to move in the desired direction.

The proposed distributed swarm control framework achieves high safety of swarm motion and the participation of informed UAVs is conducive to the guidance of the UAV swarm. Simulation results demonstrate the feasibility and effectiveness of the proposed approach.

The UAV swarm control method developed in this paper can be applied to the practice of UAV swarm control.

A distributed UAV swarm control method is proposed to ensure the effective control of the consistency and safety of swarm motion.

A simultaneous solution to the localization and mapping problem of a graph‐like environment by a swarm of robots requires solutions to task coordination and map merging. The purpose of this paper is to examine the performance of two different map‐merging strategies.

Building a representation of the environment is a key problem in robotics where the problem is known as simultaneous localization and mapping (SLAM). When large groups of robots operate within the environment, the SLAM problem becomes complicated by issues related to coordination of the elements of the swarm and integration of the environmental representations obtained by individual swarm elements. This paper considers these issues within the formalism of a group of simulated robots operating within a graph‐like environment. Starting at a common node, the swarm partitions the unknown edges of the known graph and explores the graph for a pre‐arranged period. The swarm elements then meet at a particular time and location to integrate their partial world models. This process is repeated until the entire world has been mapped. A correctness proof of the algorithm is presented, and different coordination strategies are compared via simulation.

The paper demonstrates that a swarm of identical robots, each equipped with its own marker, and capable of simple sensing and action abilities, can explore and map an unknown graph‐like environment. Moreover, experimental results show that exploration with multiple robots can provide an improvement in exploration effort over a single robot and that this improvement does not scale linearly with the size of the swarm.

The paper represents efforts toward exploration and mapping in a graph‐like world with robot swarms. The paper suggests several extensions and variations including the development of adaptive partitioning and rendezvous schedule strategies to further improve both overall swarm efficiency and individual robot utilization during exploration.

The novelty associated with this paper is the formal extension of the single robot graph‐like exploration of Dudek et al. to robot swarms. The paper here examines fundamental limits to multiple robot SLAM and does this within a topological framework. Results obtained within this topological formalism can be readily transferred to the more traditional metric representation.

The purpose of this paper is to present a neuro‐fuzzy system with a reinforcement learning algorithm (RL) for adaptive swarm behaviors acquisition. The basic idea is that each individual (agent) has the same internal model and the same learning procedure, and the adaptive behaviors are acquired only by the reward or punishment from the environment. The formation of the swarm is also designed by RL, e.g. temporal difference (TD)‐error learning algorithm, and it may bring out a faster exploration procedure comparing with the case of individual learning.

The internal model of each individual composes a part of input states classification by a fuzzy net, and a part of optimal behavior learning network which adopting a kind of RL methodology named actor‐critic method. The membership functions and fuzzy rules in the fuzzy net are adaptively formed online by the change of environment states observed in the trials of agent's behaviors. The weights of connections between the fuzzy net and the action‐value functions of actor which provides a stochastic policy of action selection, and critic which provides an evaluation to state transmission, are modified by TD‐error.

Simulation experiments of the proposed system with several goal‐directed navigation problems are accomplished and the results show that swarms are successfully formed and optimized routes are found by swarm learning faster than the case of individual learning.

Two techniques, i.e. fuzzy identification system and RL algorithm, are fused into an internal model of the individuals for swarm formation and adaptive behavior acquisition. The proposed model may be applied to multi‐agent systems, swarm robotics, metaheuristic optimization, and so on.

The purpose of this paper is to investigate how a viable system, the honey bee swarm, gathers meaningful information about potential new nest sites in its problematic environment.

This investigation uses a cybernetic model of a self-organising information network to analyse the findings from the last 60 years published research on swarm behaviour.

Nest site scouts used a modified foraging network to carry out a very thorough survey of the swarm’s problematic environment, providing the swarm with a considerable diversity of potential nest sites for consideration. The swarm utilised a number of randomly recruited groups of scouts to obtain numerous independent opinions about potential nest sites, each privately evaluated, publicly reported and repeatedly tested by new recruits. Independent evaluation of site quality was balanced by interdependent reporting of site location. Noise was reduced by integration over a large number of individual scouts and over a period of time. The swarm was therefore able to reduce potential sources of bias, distortion and noise, providing it with comparatively reliable information for decision making.

Information gathering by a honey bee swarm has not previously been modelled as a self-organising information network. The findings may be of value to human decision-making groups.

The purpose of this paper is to develop an effective and efficient approach to exploit meta‐heuristic in particle swarm optimization (PSO) for the job shop scheduling problem (JSP), a class of NP‐hard optimization problems. The approach is to be built on a PSO with multiple independent swarms. PSO was inspired by bird flocking and animal social behaviors. The particles operate collectively like a swarm that flies through the hyperdimensional space to search for possible optimal solutions. The behavior of the particles is influenced by their tendency to learn from their personal past experience and from the success of their peers to adjust their flying speed and direction. Research in fusing the multiple‐swarm concept into PSO is well‐established in solving single objective optimization problems and multimodal problems.

This study examines the optimization of the JSP via a search space division scheme and use of the meta‐heuristic method of PSO by assigning each machine in a JSP an independent swarm of particles. The use of multiple swarms in PSO is motivated by the idea of “divide and conquer” to reduce the computational complexity incurred through solving a NP‐hard combinatorial optimization problem. The resulted design, JSP/PSO algorithm, fully exploits the computing power presented by the multiple‐swarm PSO.

Simulation experiments show that the proposed JSP/PSO algorithm can effectively solve the JSP problems from small to median size. If certain mechanism of information sharing between swarms can be incorporated, it is believed that the new design could offer even more computing power to tackle the large‐sized problems.

The proposed JSP/PSO algorithm is effective in solving JSPs. The proposed algorithm shows considerable promise when searching the space of non‐delay schedules. It demands relatively lower number of function evaluations compared to other state‐of‐the‐art. The drawback to the JSP/PSO is that the GT scheduling adopted is too computationally expensive. Future works will address this concern.

The purpose of this paper is to carry out a detailed investigation of the mechanisms operating during decision making by the honey bee swarm, which is now considered to be one of the best examples of collective decision making outside the human domain.

This investigation is based on a review of the last 60 years’ published literature about swarm behaviour. It introduces a different perspective to the work by utilising a cybernetic model of a self-organising information network to analyse the findings of this body of research.

Scout bees evaluating potential nest sites accumulated support for their site by differential net recruitment, so the total scout numbers present at each site was a good measure of the total evidence in favour of the site and hence the relative probability of choosing it as the swarm’s new home. The accumulation of evidence continued at a number of alternative nest site locations until a critical quorum threshold was sensed at one of them. The first alternative to reach the threshold was chosen as the preferred nest site. Quorum scouts then prepared the swarm for departure and steered it to its new home.

Swarm decision making has not been modelled as a self-organising information network before. This novel approach reveals how a combination of network modifications, self-amplification, self-attenuation, cross-inhibition, integration and quorum mechanisms together contribute towards accurate group decision making.

The purpose of this paper is to describe a multi‐robot solution to the problem of chemical source localization, in which a team of inexpensive, simple vehicles with short‐range, low‐power sensing, communication, and processing capabilities trace a chemical plume to its source emitter

The source localization problem is analyzed using computational fluid dynamics simulations of airborne chemical plumes. The analysis is divided into two parts consisting of two large experiments each: the first part focuses on the issues of collaborative control, and the second part demonstrates how task performance is affected by the number of collaborating robots. Each experiment tests a key aspect of the problem, e.g. effects of obstacles, and defines performance metrics that help capture important characteristics of each solution.

The new empirical simulations confirmed previous theoretical predictions: a physics‐based approach is more effective than the biologically inspired methods in meeting the objectives of the plume‐tracing mission. This gain in performance is consistent across a variety of plume and environmental conditions. This work shows that high success rate can be achieved by robots using strictly local information and a fully decentralized, fault‐tolerant, and reactive control algorithm.

This is the first paper to compare a physics‐based approach against the leading alternatives for chemical plume tracing under a wide variety of fluid conditions and performance metrics. This is also the first presentation of the algorithms showing the specific mechanisms employed to achieve superior performance, including the underlying fluid and other physics principles and their numerical implementation, and the mechanisms that allow the practitioner to duplicate the outstanding performance of this approach under conditions of many robots navigating through obstacle‐dense environments.

The purpose of this paper is to demonstrate a novel use of a generalized Lennard‐Jones (LJ) force law in Physicomimetics, combined with offline evolutionary learning, for the control of swarms of robots moving through obstacle fields towards a goal. The paper then extends the paradigm to demonstrate the utility of a real‐time online adaptive approach named distributed agent evolution with dynamic adaptation to local unexpected scenarios (DAEDALUS).

To achieve the best performance, the parameters of the force law used in the Physicomimetics approach are optimized, using an evolutionary algorithm (EA) (offline learning). A weighted fitness function is utilized consisting of three components: a penalty for collisions, lack of swarm cohesion, and robots not reaching the goal. Each robot of the swarm is then given a slightly mutated copy of the optimized force law rule set found with offline learning and the robots are introduced to a more difficult environment. The online learning framework (DAEDALUS) is used for swarm adaptation in this more difficult environment.

The novel use of the generalized LJ force law combined with an EA surpasses the prior state‐of‐the‐art in the control of swarms of robots moving through obstacle fields. In addition, the DAEDALUS framework allows the swarms of robots to not only learn and share behavioral rules in changing environments (in real time), but also to learn the proper amount of behavioral exploration that is appropriate.

There are significant issues that arise with respect to “wall following methods” and “local minimum trap” problems. “Local minimum trap” problems have been observed in this paper, but this issue is not addressed in detail. The intention is to explore other approaches to develop more robust adaptive algorithms for online learning. It is believed that the learning of the proper amount of behavioral exploration can be accelerated.

In order to provide meaningful comparisons, this paper provides a more complete set of metrics than prior papers in this area. The paper examines the number of collisions between robots and obstacles, the distribution in time of the number of robots that reach the goal, and the connectivity of the formation as it moves.

This paper addresses the difficult task of moving a large number of robots in formation through a large number of obstacles. The important real‐world constraint of “obstructed perception” is modeled. The obstacle density is approximately three times the norm in the literature. The paper shows how concepts from population genetics can be used with swarms of agents to provide fast online adaptive learning in these challenging environments. In addition, this paper also presents a more complete set of metrics of performance.

Collaboration is a common phenomenon in human society. The best way of collaborations can make the group achieve the best interests. Because of the low cost and high repeatability of simulation, it is a good method to explore the best way of collaborations by means of simulation. The traditional simulation is difficult to adapt to the crowd intelligence network simulation, so the crowd collaborations simulation is proposed.

In this paper, the atomic swarm intelligence unit and collective swarm intelligence unit are proposed to represent the behavior of individuals and groups in physical space and the interaction between them.

To explore the best collaboration mode of the group, a framework of crowd collaborations simulation is proposed, which decomposes the big goal into the small goals by constructing the cooperation chain and analyzes the cooperation results and feeds them back to the next simulation.

Two kinds of swarm intelligence units are used to represent the simulated individuals in the group, and the pattern is used to represent individual behavior. It is suitable for the simulation of collaboration problems in various types and situations.