A Fast and Unified Method to Find a Minimum-Jerk Robot Joint Trajectory Using Particle Swarm Optimization

In robot trajectory planning, finding the minimum-jerk joint trajectory is a crucial issue in robotics because most robots are asked to perform a smooth trajectory. Jerk, the third derivative of joint position of a trajectory, influences how smoothly and efficiently a robot moves. Thus, the minimum-jerk joint trajectory makes the robot control algorithm simple and robust. To find the minimum-jerk joint trajectory, it has been formulated as an optimization problem constrained by joint inter-knot parameters including initial joint displacement and velocity, intermediate joint displacement, and final joint displacement and velocity. In this paper, we propose a fast and unified approach based on particle swarm optimization (PSO) with K-means clustering to solve the near-optimal solution of a minimum-jerk joint trajectory. This work differs from previous work in its fast computation and unified methodology. Computer simulations were conducted and showed the competent performance of our approach on a six degree-of-freedom robot manipulator.     

Optimal replica placement in hierarchical Data Grids with locality assurance

In this paper, we address three issues concerning data replica placement in hierarchical Data Grids that can be presented as tree structures. The first is how to ensure load balance among replicas. To achieve this, we propose a placement algorithm that finds the optimal locations for replicas so that their workload is balanced. The second issue is how to minimize the number of replicas. To solve this problem, we propose an algorithm that determines the minimum number of replicas required when the maximum workload capacity of each replica server is known. Finally, we address the issue of service quality by proposing a new model in which each request must be given a quality-of-service guarantee. We describe new algorithms that ensure both workload balance and quality of service simultaneously.     

Application-independent evaluation of speaker detection

We propose and motivate an alternative to the traditional error-based or cost-based evaluation metrics for the goodness of speaker detection performance. The metric that we propose is an information-theoretic one, which measures the effective amount of information that the speaker detector delivers to the user. We show that this metric is appropriate for the evaluation of what we call application-independent detectors, which output soft decisions in the form of log-likelihood-ratios, rather than hard decisions. The proposed metric is constructed via analysis and generalization of cost-based evaluation metrics. This construction forms an interpretation of this metric as an expected cost, or as a total error-rate, over a range of different application-types. We further show how the metric can be decomposed into a discrimination and a calibration component. We conclude with an experimental demonstration of the proposed technique to evaluate three speaker detection systems submitted to the NIST 2004 Speaker Recognition Evaluation.     

Active SLAM and Exploration with Particle Filters Using Kullback-Leibler Divergence

Autonomous exploration under uncertain robot location requires the robot to use active strategies to trade-off between the contrasting tasks of exploring the unknown scenario and satisfying given constraints on the admissible uncertainty in map estimation. The corresponding problem, namely active SLAM (Simultaneous Localization and Mapping) and exploration, has received a large attention from the robotic community for its relevance in mobile robotics applications. In this work we tackle the problem of active SLAM and exploration with Rao-Blackwellized Particle Filters. We propose an application of Kullback-Leibler divergence for the purpose of evaluating the particle-based SLAM posterior approximation. This metric is then applied in the definition of the expected information from a policy, which allows the robot to autonomously decide between exploration and place revisiting actions (i.e., loop closing). Extensive tests are performed in typical indoor and office environments and on well-known benchmarking scenarios belonging to SLAM literature, with the purpose of comparing the proposed approach with the state-of-the-art techniques and to evaluate the maturity of truly autonomous navigation systems based on particle filtering.     

Multi-granularity and metric spatial reasoning

Composition reasoning is a basic reasoning task in qualitative spatial reasoning (QSR). It is an important qualitative method for robot navigation, node localization in wireless sensor networks and other fields. The previous composition reasoning works dedicated in single granularity framework. Multi-granularity spatial relation is not rare in real world, and some qualitative spatial relation models are multi-granularity models, such as RCC, STAR_m, CDC_m and OPRA_m. Although multi-granularity composition reasoning is very useful in many applications, it has not been systematically studied before. A special case of multi-granularity composition reasoning, referred to as metric spatial reasoning, is also discussed here. The general frameworks and basic theories for multi-granularity and metric spatial reasoning are put forward here. Furthermore, we redefine the spatial relation models for distance, topology and direction under the proposed multi-granularity and metric frameworks. We add metric representation for the OPRA_m. The multi-granularity and metric reasoning tasks are studied for these four models for the first time. Finally we perform some experiments on OPRA_m with encouraging results to verify our theories. Multi-granularity and metric spatial reasoning tasks are new problems in QSR and quite different from the previous works. Our works can be potentially applied in robot navigation, wireless sensor networks and other applications.     

Virtual reality as a communication process

In this work, we consider immersive Virtual Reality (VR) as a communication process between humans, mediated by computer systems, which uses interaction, visualization, and other sensory stimuli to convey information. From this viewpoint, it is relevant to understand how VR can solve a given communication problem, what is therefore the expressive power of VR system, i.e., its ability in establishing the communication, what are the guidelines to design an effective system, and what are the more relevant models of VR applications. Firstly, we try to clarify the notion of reality in Virtual Reality systems and conclude that reality is not an intrinsic characteristic of VR, rather the result of a conventional way of coding information. The purpose of coding is to lead the observer to the conclusion that the VR set is what is called in italian as verisimile (from Latin veri similis), i.e., ??similar-to-the-real-thing??. So the creation of an effective VR application is an artifice or an illusion. But in order to avoid an over-reliance on the creativity of the VR designer, we intend to identify a solid ground on which different kinds of VR solutions can be considered in terms of their ability to solve the desired communication objective. To this aim, we will rely on methods derived from rhetoric to semiotics.     

Mapping Simple Polygons: How Robots Benefit from Looking Back

We consider the problem of mapping an initially unknown polygon of size n with a simple robot that moves inside the polygon along straight lines between the vertices. The robot sees distant vertices in counter-clockwise order and is able to recognize the vertex among them which it came from in its last move, i.e. the robot can look back. Other than that the robot has no means of distinguishing distant vertices. We assume that an upper bound on n is known to the robot beforehand and show that it can always uniquely reconstruct the visibility graph of the polygon. Additionally, we show that multiple identical and deterministic robots can always solve the weak rendezvous problem in which the robots need to position themselves such that all of them are mutually visible to each other. Our results are tight in the sense that the strong rendezvous problem, where robots need to gather at a vertex, cannot be solved in general, and, without knowing a bound beforehand, not even n can be determined. In terms of mobile agents exploring a graph, our result implies that they can reconstruct any graph that is the visibility graph of a simple polygon. This is in contrast to the known result that the reconstruction of arbitrary graphs is impossible in general, even if n is known.     

Transferring knowledge across robots: A risk sensitive approach

One of the most impressive characteristics of human perception is its domain adaptation capability. Humans can recognize objects and places simply by transferring knowledge from their past experience. Inspired by that, current research in robotics is addressing a great challenge: building robots able to sense and interpret the surrounding world by reusing information previously collected, gathered by other robots or obtained from the web. But, how can a robot automatically understand what is useful among a large amount of information and perform knowledge transfer? In this paper we address the domain adaptation problem in the context of visual place recognition. We consider the scenario where a robot equipped with a monocular camera explores a new environment. In this situation traditional approaches based on supervised learning perform poorly, as no annotated data are provided in the new environment and the models learned from data collected in other places are inappropriate due to the large variability of visual information. To overcome these problems we introduce a novel transfer learning approach. With our algorithm the robot is given only some training data (annotated images collected in different environments by other robots) and is able to decide whether, and how much, this knowledge is useful in the current scenario. At the base of our approach there is a transfer risk measure which quantifies the similarity between the given and the new visual data. To improve the performance, we also extend our framework to take into account multiple visual cues. Our experiments on three publicly available datasets demonstrate the effectiveness of the proposed approach.     

Optimality and competitiveness of exploring polygons by mobile robots

A mobile robot, represented by a point moving along a polygonal line in the plane, has to explore an unknown polygon and return to the starting point. The robot has a sensing area which can be a circle or a square centered at the robot. This area shifts while the robot moves inside the polygon, and at each point of its trajectory the robot “sees” (explores) all points for which the segment between the robot and the point is contained in the polygon and in the sensing area. We focus on two tasks: exploring the entire polygon and exploring only its boundary. We consider several scenarios: both shapes of the sensing area and the Manhattan and the Euclidean metrics.We focus on two quality benchmarks for exploration performance: optimality (the length of the trajectory of the robot is equal to that of the optimal robot knowing the polygon) and competitiveness (the length of the trajectory of the robot is at most a constant multiple of that of the optimal robot knowing the polygon). Most of our results concern rectilinear polygons. We show that optimal exploration is possible in only one scenario, that of exploring the boundary by a robot with square sensing area, starting at the boundary and using the Manhattan metric. For this case we give an optimal exploration algorithm, and in all other scenarios we prove impossibility of optimal exploration. For competitiveness the situation is more optimistic: we show a competitive exploration algorithm for rectilinear polygons whenever the sensing area is a square, for both tasks, regardless of the metric and of the starting point. Finally, we show a competitive exploration algorithm for arbitrary convex polygons, for both shapes of the sensing area, regardless of the metric and of the starting point.     

Balancing task allocation in multi-robot systems using K-means clustering and auction based mechanisms

This paper aims to solve the balanced multi-robot task allocation problem. Multi-robot systems are becoming more and more significant in industrial, commercial and scientific applications. Effectively allocating tasks to multi-robots i.e. utilizing all robots in a cost effective manner becomes a tedious process. The current attempts made by the researchers concentrate only on minimizing the distance between the robots and the tasks, and not much importance is given to the balancing of work loads among robots. It is also found from the literature that the multi-robot system is analogous to Multiple Travelling Salesman Problem (MTSP). This paper attempts to develop mechanism to address the above two issues with objective of minimizing the distance travelled by ‘m’ robots and balancing the work load between ‘m’ robots equally. The proposed approach has two fold, first develops a mathematical model for balanced multi-robot task allocation problem, and secondly proposes a methodology to solve the model in three stages. Stage I groups the ‘N’ tasks into ‘n’ clusters of tasks using K-means clustering technique with the objective of minimizing the distance between the tasks, stage II calculates the travel cost of robot and clusters combination, stage III allocates the robot to the clusters in order to utilise all robot in a cost effective manner.     

Towards modelling complex robot training tasks through system identification

Previous research has shown that sensor–motor tasks in mobile robotics applications can be modelled automatically, using NARMAX system identification, where the sensory perception of the robot is mapped to the desired motor commands using non-linear polynomial functions, resulting in a tight coupling between sensing and acting — the robot responds directly to the sensor stimuli without having internal states or memory.However, competences such as for instance sequences of actions, where actions depend on each other, require memory and thus a representation of state. In these cases a simple direct link between sensory perception and the motor commands may not be enough to accomplish the desired tasks. The contribution of this paper to knowledge is to show how fundamental, simple NARMAX models of behaviour can be used in a bootstrapping process to generate complex behaviours that were so far beyond reach.We argue that as the complexity of the task increases, it is important to estimate the current state of the robot and integrate this information into the system identification process. To achieve this we propose a novel method which relates distinctive locations in the environment to the state of the robot, using an unsupervised clustering algorithm. Once we estimate the current state of the robot accurately, we combine the state information with the perception of the robot through a bootstrapping method to generate more complex robot tasks: We obtain a polynomial model which models the complex task as a function of predefined low level sensor–motor controllers and raw sensory data.The proposed method has been used to teach Scitos G5 mobile robots a number of complex tasks, such as advanced obstacle avoidance, or complex route learning.     

Algorithm for the discrete Weber’s problem with an accuracy estimate

We consider a relaxation of the quadratic assignment problem without the constraint on the number of objects assigned to a specific position. This problem is NP-hard in the general case. To solve the problem, we propose a polynomial algorithm with guaranteed posterior accuracy estimate; we distinguish a class of problems with special assignment cost functions where the algorithm is 2-approximate. We show that if the graph in question contains one simple loop, and the set of assignment positions is a metric space, the proposed algorithm is 2-approximate and guaranteed to be asymptotically exact. We conduct a computational experiment in order to analyze the algorithm’s errors and evaluate its accuracy.     

Autonomous and fast robot learning through motivation

Research on robot techniques that are fast, user-friendly, and require little application-specific knowledge by the user, is more and more encouraged in a society where the demand of home-care or domestic-service robots is increasing continuously. In this context we propose a methodology which combines reinforcement learning and genetic algorithms to teach a robot how to perform a task when only the specification of the main restrictions of the desired behaviour is provided. Through this combination, both paradigms must be merged in such a way that they influence each other to achieve a fast convergence towards a good robot-control policy, and reduce the random explorations the robot needs to carry out in order to find a solution.Another advantage of our proposal is that it is able to easily incorporate any kind of domain-dependent knowledge about the task. This is very useful for improving a robot controller, for applying a robot-controller to move a different robot-platform, or when we have certain “feelings” about how the task should be solved.The performance of our proposal is shown through its application to solve a common problem in mobile robotics.     

Pure reactive behavior learning using Case Based Reasoning for a vision based 4-legged robot

A traditional problem in robotics is adaptation of developed algorithms to different platforms and sensors, as each of them has its specifics and associated errors. Hierarchical control architectures deal with the problem through division of the system into layers, where deliberative processing is performed at high level and low level layers are in charge of dealing with reactive behaviors and adaptation to platform and sensor hardware. Specifically, approaches based on the Emergent Behavior Theory rely on building high level behaviors by combining simpler ones that provide intuitive reactive responses to sensory instance. This combination is controlled by higher layers in order to obtain more complex behaviors. Unfortunately, low level behaviors might be difficult to develop, specially when dealing with legged robots and sensors like video cameras, where resulting motion is heavily influenced by the robot kinematics and dynamics and sensory input is affected by external conditions, transformations, distortions, noise and motion itself (e.g. the camera bouncing problem).In this paper, we propose a new learning based method to solve most of these problems. It basically consists of creating a reactive behavior by supervisedly driving a robot for a time. During that time, its visual input is reactively associated to commands sent to the robot through a Case Based Reasoning (CBR) behavior builder. Thus, the robot learns what the person would do in its situation to achieve a certain goal. This approach has two advantages. First, humans are particularly good at adapting and taking into account the specifics of a given mobile after some use. Thus, kinematics and dynamics are absorbed into the casebase along with how the person thinks they should be dealt with by that particular robot. Similarly, commands are associated to the input sensor as is, so systematic errors in sensors and motors are also implicitly learnt in the casebase (camera bouncing, distorsions, noise …). Also, different reactive strategies to reach a simple goal can be programmed into the robot by showing, rather than by coding. This is particularly useful because some reactive behaviors are ill-fitted to equations. Naturally, CBR allows online adaptation to potential changes after supervised training, so the system is able to learn by itself when working autonomously too.The proposed system has been successfully tested in a 4-legged Aibo robot in a controlled environment. To prove that it is adequate to create low level layers for hybrid architectures, two different CBR reactive behaviors have been tested and combined into an emergent one. A deliberative layer could be used to extent the system to more complex environments.     

Using perspective taking to learn from ambiguous demonstrations

This paper addresses an important issue in learning from demonstrations that are provided by “naïve” human teachers—people who do not have expertise in the machine learning algorithms used by the robot. We therefore entertain the possibility that, whereas the average human user may provide sensible demonstrations from a human’s perspective, these same demonstrations may be insufficient, incomplete, ambiguous, or otherwise “flawed” from the perspective of the training set needed by the learning algorithm to generalize properly. To address this issue, we present a system where the robot is modeled as a socially engaged and socially cognitive learner. We illustrate the merits of this approach through an example where the robot is able to correctly learn from “flawed” demonstrations by taking the visual perspective of the human instructor to clarify potential ambiguities.     

Good recognition is non-metric

Recognition is the fundamental task of visual cognition, yet how to formalize the general recognition problem for computer vision remains an open issue. The problem is sometimes reduced to the simplest case of recognizing matching pairs, often structured to allow for metric constraints. However, visual recognition is broader than just pair-matching: what we learn and how we learn it has important implications for effective algorithms. In this review paper, we reconsider the assumption of recognition as a pair-matching test, and introduce a new formal definition that captures the broader context of the problem. Through a meta-analysis and an experimental assessment of the top algorithms on popular data sets, we gain a sense of how often metric properties are violated by recognition algorithms. By studying these violations, useful insights come to light: we make the case for local distances and systems that leverage outside information to solve the general recognition problem.     

On fractional faces of the metric polytope

The integrality recognition problem is considered on a sequence M _{n, k} of nested relaxations of a Boolean quadric polytope, including the rooted semimetric M _n and metric M _{n, 3} polytopes. The constraints of the metric polytope cut off all faces of the rooted semimetric polytope that contain only fractional vertices. This makes it possible to solve the integrality recognition problem on M _n in polynomial time. To solve the integrality recognition problem on the metric polytope, we consider the possibility of cutting off all fractional faces of M _{n, 3} by a certain relaxation M _{n, k} . The coordinates of points of the metric polytope are represented in homogeneous form as a three-dimensional block matrix. We show that in studying the question of cutting off the fractional faces of the metric polytope, it is sufficient to consider only constraints in the form of triangle inequalities.     

A topological characterization of random sequences

The set of random sequences is large in the sense of measure, but small in the sense of category. This is the case when we regard the set of infinite sequences over a finite alphabet as a subset of the usual Cantor space. In this note we will show that the above result depends on the topology chosen. To this end we will use a relativization of the Cantor topology, the U^δ-topology introduced by Staiger [RAIRO Inform. Théor. 21 (1987) 147-173]. This topology is also metric, but the distance between two sequences does not depend on their longest common prefix (Cantor metric), but on the number of their common prefixes in a given language U. The resulting space is complete, but not always compact. We will show how to derive a computable set U from a universal Martin-Löf test such that the set of non-random sequences is nowhere dense in the Uδ-topology. As a byproduct we obtain a topological characterization of the set of random sequences. We also show that the Law of Large Numbers, which fails with respect to the usual topology, is true for the Uδ-topology.     

Learning a Mahalanobis distance metric for data clustering and classification

Distance metric is a key issue in many machine learning algorithms. This paper considers a general problem of learning from pairwise constraints in the form of must-links and cannot-links. As one kind of side information, a must-link indicates the pair of the two data points must be in a same class, while a cannot-link indicates that the two data points must be in two different classes. Given must-link and cannot-link information, our goal is to learn a Mahalanobis distance metric. Under this metric, we hope the distances of point pairs in must-links are as small as possible and those of point pairs in cannot-links are as large as possible. This task is formulated as a constrained optimization problem, in which the global optimum can be obtained effectively and efficiently. Finally, some applications in data clustering, interactive natural image segmentation and face pose estimation are given in this paper. Experimental results illustrate the effectiveness of our algorithm.     

A neural framework for adaptive robot control

This paper investigates how dynamics in recurrent neural networks can be used to solve some specific mobile robot problems such as motion control and behavior generation. We have designed an adaptive motion control approach based on a novel recurrent neural network, called Echo state networks. The advantage is that no knowledge about the dynamic model is required, and no synaptic weight changing is needed in presence of time varying parameters in the robot. To generate the robot behavior over time, we adopted a biologically inspired approach called neural fields. Due to its dynamical properties, a neural field produces only one localized peak that indicates the optimum movement direction, which navigates a mobile robot to its goal in an unknown environment without any collisions with static or moving obstacles.