Papers by Krishnanand Kaipa
Neural Networks, Jan 1, 2010
Infants exploit the perception that others are &a... more Infants exploit the perception that others are 'like me' to bootstrap social cognition (Meltzoff, 2007a). This paper demonstrates how the above theory can be instantiated in a social robot that uses itself as a model to recognize structural similarities with other robots; this thereby enables the student to distinguish between appropriate and inappropriate teachers. This is accomplished by the student robot first performing self-discovery, a phase in which it uses actuation-perception relationships to infer its own structure. Second, the student models a candidate teacher using a vision-based active learning approach to create an approximate physical simulation of the teacher. Third, the student determines that the teacher is structurally similar (but not necessarily visually similar) to itself if it can find a neural controller that allows its self model (created in the first phase) to reproduce the perceived motion of the teacher model (created in the second phase). Fourth, the student uses the neural controller (created in the third phase) to move, resulting in imitation of the teacher. Results with a physical student robot and two physical robot teachers demonstrate the effectiveness of this approach. The generalizability of the proposed model allows it to be used over variations in the demonstrator: The student robot would still be able to imitate teachers of different sizes and at different distances from itself, as well as different positions in its field of view, because change in the interrelations of the teacher's body parts are used for imitation, rather than absolute geometric properties.
Bookmarks Related papers MentionsView impact
Swarm Intelligence, Jan 1, 2009
This paper presents glowworm swarm optimization (GSO), a novel algorithm for the simultaneous com... more This paper presents glowworm swarm optimization (GSO), a novel algorithm for the simultaneous computation of multiple optima of multimodal functions. The algorithm shares a few features with some better known swarm intelligence based optimization algorithms, such as ant colony optimization and particle swarm optimization, but with several significant differences. The agents in GSO are thought of as glowworms that carry a luminescence quantity called luciferin along with them. The glowworms encode the fitness of their current locations, evaluated using the objective function, into a luciferin value that they broadcast to their neighbors. The glowworm identifies its neighbors and computes its movements by exploiting an adaptive neighborhood, which is bounded above by its sensor range. Each glowworm selects, using a probabilistic mechanism, a neighbor that has a luciferin value higher than its own and moves toward it. These movements—based only on local information and selective neighbor interactions—enable the swarm of glowworms to partition into disjoint subgroups that converge on multiple optima of a given multimodal function. We provide some theoretical results related to the luciferin update mechanism in order to prove the bounded nature and convergence of luciferin levels of the glowworms. Experimental results demonstrate the efficacy of the proposed glowworm based algorithm in capturing multiple optima of a series of standard multimodal test functions and more complex ones, such as stair-case and multiple-plateau functions. We also report the results of tests in higher dimensional spaces with a large number of peaks. We address the parameter selection problem by conducting experiments to show that only two parameters need to be selected by the user. Finally, we provide some comparisons of GSO with PSO and an experimental comparison with Niche-PSO, a PSO variant that is designed for the simultaneous computation of multiple optima.
Bookmarks Related papers MentionsView impact
Proceedings of the 11th …, Jan 1, 2009
Bookmarks Related papers MentionsView impact
We present the theoretical foundations for the multiple rendezvous problem involving design of lo... more We present the theoretical foundations for the multiple rendezvous problem involving design of local control strategies that enable groups of visibility-limited mobile agents to split into subgroups, exhibit simultaneous taxis behavior towards, and eventually rendezvous at, multiple unknown locations of interest. The theoretical results are proved under certain restricted set of assumptions. The algorithm used to solve the above problem is based on a glowworm swarm optimization (GSO) technique, developed earlier, that finds multiple optima of multimodal objective functions. The significant difference between our work and most earlier approaches to agreement problems is the use of a virtual local-decision domain by the agents in order to compute their movements. The range of the virtual domain is adaptive in nature and is bounded above by the maximum sensor/visibility range of the agent. We introduce a new decision domain update rule that enhances the rate of convergence by a factor of approximately two. We use some illustrative simulations to support the algorithmic correctness and theoretical findings of the paper
Bookmarks Related papers MentionsView impact
Filtration & Separation, Jan 1, 2005
Abstract This paper presents a glowworm swarm based algorithm that finds solutions to optimizatio... more Abstract This paper presents a glowworm swarm based algorithm that finds solutions to optimization of multiple optima continuous functions. The algorithm is a variant of a well-known ant-colony optimization (ACO) technique, but with several significant modifications. Similar to how each moving region in the ACO technique is associated with a pheromone value, the agents in our algorithm carry a luminescence quantity along with them. Agents are thought of as glowworms that emit a light whose intensity is proportional to the associated ...
Bookmarks Related papers MentionsView impact
Ant Colony Optimization …, Jan 1, 2006
Bookmarks Related papers MentionsView impact
Abstract This paper presents a glowworm metaphor based distributed algorithm that enables a colle... more Abstract This paper presents a glowworm metaphor based distributed algorithm that enables a collection of minimalist mobile robots to split into subgroups, exhibit simultaneous taxis-behavior towards, and rendezvous at multiple radiation sources such as nuclear/hazardous chemical spills and fire-origins in a fire calamity. The algorithm is based on a glowworm swarm optimization (GSO) technique that finds multiple optima of multimodal functions. The algorithm is in the same spirit as the ant-colony optimization (ACO) algorithms, but with ...
Bookmarks Related papers MentionsView impact
Bookmarks Related papers MentionsView impact
Robotics and Autonomous Systems, Jan 1, 2005
In this paper we address the problem of synthesizing simple rules and local interactions at the i... more In this paper we address the problem of synthesizing simple rules and local interactions at the individual level so that pre-specified complex behavior emerges at the group level of a collection of autonomous mobile agents. Usually, the emergent collective behavior is used to perform certain spatial group-tasks. Specifically, we consider self-assembling of a group of mobile robots into grid, line, and wedge patterns. We introduce the notion of local-templates in which each mobile agent – capable of simple forward/backward movements and a clock-wise/counter clock-wise spin – actively encodes distinctive information into multiple non-overlapping sectorial regions of the surrounding environment in order to form pose-specific virtual links with similar minimalist agents in a local neighborhood. The resulting local patterns around each agent lead to the desired global formation. In order to take mobile robots closer to their biological counterparts, there is a need to further simplify the manner in which they currently perceive their surroundings, communicate with their neighbors, and compute their actions. We have built a robotic platform consisting of four wheeled-mobile robots that are christened as Kinbots. They are similar in principle to Braitenberg Vehicles and use simple perception/interaction/actuation techniques to achieve individual vehicle complexity and produce effective group behavior through cooperation. To validate the proposed approach, we demonstrate a column-formation task in computer simulations and physical experiments. We illustrate an experiment which is representative of various prominent stages in a group-formation task such as formation-achievement, maintenance, and response of formation movement to the presence of obstacles.
Bookmarks Related papers MentionsView impact
Bookmarks Related papers MentionsView impact
Bookmarks Related papers MentionsView impact
International Journal of Computational Intelligence Studies, Jan 1, 2009
This paper presents an exposition of a new method of swarm intelligence based algorithm for optim... more This paper presents an exposition of a new method of swarm intelligence based algorithm for optimising multi-modal functions. The main objective of using this method is to ensure capture of all local maxima of the function. The application of this method is in the area of multiple signal source location or identification of odour sources and hazardous spills. The method is based upon a dynamic decision domain for each agent in the swarm that decides its direction of movement by the strength of the signal picked up from its neighbours. This ...
Bookmarks Related papers MentionsView impact
This chapter presents glowworm swarm optimization (GSO), a novel swarm intelligence algorithm, wh... more This chapter presents glowworm swarm optimization (GSO), a novel swarm intelligence algorithm, which was recently proposed for simultaneous capture of multiple optima of multimodal functions. In particular, GSO prescribes individual-level rules that cause a swarm of agents deployed in a signal medium to automatically partition into subswarms that converge on the multiple sources of the signal profile. The sources could represent multiple optima in a numerical optimization problem or physical quantities like sound, light, or heat in a realistic robotic source localization task. We present the basic GSO model and use a numerical example to characterize the group-level phases of the algorithm that gives an insight into how GSO explicitly addresses the issue of achievement/maintenance of swarm diversity. We briefly summarize the results from the application of GSO to the following three problems−multimodal function optimization, signal source localization, and pursuit of mobile signal sources.
Bookmarks Related papers MentionsView impact
Ubiquitous computing based environments may be defined as human surroundings that are furnished w... more Ubiquitous computing based environments may be defined as human surroundings that are furnished with a network of intelligent computing devices, which could be either stationary or mobile (or an assortment of both), in order to service certain human- generated or needed tasks. Mark Weiser introduced ubiquitous computing, in its current form, in 1988 at the Computer Science Lab at Xerox PARC and wrote some of the earliest papers on ubiquitous computing (Weiser 1999).Ubiquitous computing based environments have several applications to the industry like environmental monitoring (Kim et al. 2007), ubiquitous factory environments (Jabbar et al. 2007), and self-sensing spaces (El-Zabadani et al. 2007). Kim et al. (Kim et al. 2007) develop a framework that uses ubiquitous sensor networks for atmospheric environment monitoring. Jabbar et al.(Jabbar et al. 2007) present methods that integrate latest technologies like RFID, PDA, and Wi-Fi in order to transform a nuclear power plant into an ubiquitous factory environment where effective data communication among local area operators and control room and minimization of work duration and errors in wake of safety requirements are achieved. El-Zabadani et al. (El-Zabadani et al. 2007) propose a novel approach to mapping and sensing smart spaces in which a mobile platform equipped with on-board RFID modules identifies and locates RFID-tags that are embedded in the carpet in the form of a grid.
Bookmarks Related papers MentionsView impact
We address the problem of multiple signal source localization where robotic swarms are used to lo... more We address the problem of multiple signal source localization where robotic swarms are used to locate multiple signal sources like light, sound, heat, leaks in pressurized systems, hazardous plumes/aerosols resulting from nuclear or chemical spills, fire-origins in forest fires, hazardous chemical discharge in water bodies, oil spills, deep-sea hydrothermal vent plumes, etc. In particular, we present a multi-robot system that implements a modified version of the glowworm swarm optimization (GSO) algorithm, which is originally developed to solve multimodal function optimization problems, for this purpose. The GSO algorithm uses a leapfrogging behavior for the basic search capability and an adaptive decision range that enables the agents to partition into disjoint subgroups, simultaneously taxis towards, and rendezvous at, multiple source locations of interest. Transition of agent behaviors from simulation to real-robot-implementation needs modifications to certain algorithmic aspects mainly because of the point-agent model of the basic GSO algorithm and the physical dimensions and dynamics of a real robot. We briefly describe the basic GSO algorithm and the modifications incorporated into the algorithm in order to make it suitable for a robotic implementation. Realization of each sensing-decision-action cycle of the GSO algorithm requires the robots to perform subtasks such as identification and localization of neighbors, selection of a leader among current neighbors, updating of the associated luciferin and local-decision range, and making a step-movement towards the selected leader. Experiments in this regard validate each robot’s capability to perform the above basic algorithmic primitives. Real-robot-experiments are conducted in the context of light source localization in order to validate the GSO approach to localization of signal sources. These experiments constitute a first step toward implementation in multiple robots for detection of multiple sources.
Bookmarks Related papers MentionsView impact
This chapter will deal with the problem of searching higher dimensional spaces using glowworm swa... more This chapter will deal with the problem of searching higher dimensional spaces using glowworm swarm optimization (GSO), a novel swarm intelligence algorithm, which was recently proposed for simultaneous capture of multiple optima of multimodal functions. Tests are performed on a set of three benchmark functions and the average peak-capture fraction is used as an index to analyze GSO’s performance as a function of dimension number. Results reported from tests conducted up to a maximum of eight dimensions show the efficacy of GSO in capturing multiple peaks in high dimensions. With an ability to search for local peaks of a function (which is the measure of fitness) in high dimensions, GSO can be applied to identification of multiple data clusters, satisfying some measure of fitness defined on the data, in high dimensional databases.
Bookmarks Related papers MentionsView impact
Bookmarks Related papers MentionsView impact
Multiagent and Grid Systems, Jan 1, 2006
This paper presents multimodal function optimization, using a nature-inspired glowworm swarm opti... more This paper presents multimodal function optimization, using a nature-inspired glowworm swarm optimization (GSO) algorithm, with applications to collective robotics. GSO is similar to ACO and PSO but with important differences. A key feature of the algorithm is the use of an adaptive local-decision domain, which is used effectively to detect the multiple optimum locations of the multimodal function. Agents in the GSO algorithm have a finite sensor range which defines a hard limit on the local-decision domain used to compute their movements. ...
Bookmarks Related papers MentionsView impact
Robotics and Autonomous Systems, Jan 1, 2008
We present theoretical foundations for a variation of the multi-agent rendezvous problem involvin... more We present theoretical foundations for a variation of the multi-agent rendezvous problem involving design of local control strategies that enable agent swarms, with hard-limited sensing ranges, to split into disjoint subgroups, exhibit simultaneous taxis behavior toward, and eventually rendezvous at, multiple unknown locations of interest. The algorithm used to solve the above problem is based on a glowworm swarm optimization (GSO) technique, developed earlier, that finds multiple optima of multi-modal objective functions. We characterize the various phases of the algorithm that help us to develop a theoretical framework required for analysis. In particular, we show through simulations that the implementation of the GSO algorithm at the individual agent level gives rise to two major phases at the group level–splitting of the agent-swarm into subgroups and local convergence of agents in each subgroup to the peak locations. We provide local convergence results under certain restricted set of assumptions, leading to a simplified model of the algorithm, making it amenable to analysis, while still reflecting most of the features of the original algorithm. In particular, we find an upper bound on the time taken by the agents to converge to an isolated leader and on the time taken by the agents to converge to one of the leaders with non-isolated and non-overlapping neighborhoods. Finally, we show that agents under the influence of multiple leaders with overlapping neighborhoods asymptotically converge to one of the leaders. We present some illustrative simulations to support the theoretical findings of the paper.
Bookmarks Related papers MentionsView impact
Uploads
Papers by Krishnanand Kaipa