Integrating planning and control for single-bodied wheeled mobile robots

This paper presents an approach to couple path planning and control for mobile robot navigation in a hybrid control framework. We build upon an existing hybrid control approach called sequential composition, in which a set of feedback control policies are prescribed on well-defined domains contained in the robot’s free space. Each control policy drives the robot to a goal set, which lies in the domain of a subsequent policy. Control policies are deployed into the free state space so that when composed among one another, the overall action of the set of control policies drives the robot to perform a task, such as moving from a start to a goal location or patrolling a perimeter. A planner determines the sequence of control policies to be invoked. When control policies defined in this framework respect the low-level dynamics and kinematics of the system, this formal approach guarantees that high-level tasks are either accomplished by a given set of policies, or verifies that the tasks are not achievable with the given policies.     

Asymmetric trajectory generation and impedance control for running of biped robots

An online asymmetric trajectory generation method for biped robots is proposed to maintain dynamical postural stability and increase energy autonomy, based on the running stability criterion defined in phases. In a support phase, an asymmetric trajectories for the hip and swing leg of the biped robots is obtained from an approximated running model with two springless legs and a spring-loaded inverted pendulum model so that the zero moment point should exist inside the safety boundary of a supporting foot, and the supporting leg should absorb large reaction forces, take off and fly through the air. The biped robot is under-actuated with six degrees of under-actuation during flight. The trajectory generation strategies for the hip and both legs in a flight phase use the approximated running model and non-holonomic constraints based on the linear and angular momenta at the mass center. Next, we present an impedance control with a force modulation strategy to guarantee a stable landing on the ground and simultaneously track the desired trajectories where the desired impedance at the hip link and both legs is specified. A series of computer simulations for two different types of biped robots show that the proposed running trajectory and impedance control method satisfy the two conditions for running stability and make the biped robot more robust to variations in the desired running speed, gait transitions between walking and running, and parametric modeling errors. We have examined the feasibility of this method with running experiments on a 12-DOF biped robot without arms. The biped robot could run successfully with average forward speed of about 0.3359 [m/s]. Electronic Supplementary Material  The online version of this article () contains supplementary material, which is available to authorized users.

Model-based control architecture for attentive robots in rescue scenarios

In this paper we present a control architecture for an autonomous rescue robot specialized in victim finding in an unknown and unstructured environment. The reference domain for rescue robots is the rescue-world arenas purposefully arranged for the Robocup competitions. The main task of a rescue mobile robot is to explore the environment and report to the rescue-operators the map of visited areas annotated with its finding. In this context all the attentional activities play a major role in decision processes: salient elements in the environment yield utilities and objectives. A model-based executive controller is proposed to coordinate, integrate, and monitor the distributed decisions and initiatives emerging from the modules involved in the control loop. We show how this architecture integrates the reactive model-based control of a rescue mission, with an attentive perceptual activity processing the sensor and visual stimuli. The architecture has been implemented and tested in real-world experiments by comparing the performances of metric exploration and attentive exploration. The results obtained demonstrate that the attentive behavior significantly focus the exploration time in salient areas enhancing the overall victim finding effectiveness.

Design and control of a heavy material handling manipulator for agricultural robots

In this paper, we propose a manipulation system for agricultural robots that handle heavy materials. The structural systems of a mobile platform and a manipulator are selected and designed after proposing new knowledge about agricultural robots. Also, the control systems for these structural systems are designed in the presence of parametric perturbation and uncertainty while avoiding conservative results. The validity of both the structural and control systems is confirmed by conducting watermelon harvesting experiments in an open field. Furthermore, an explicit design procedure is confirmed for both the structural and control systems and three key design tools are clarified.

A recursive, numerically stable, and efficient simulation algorithm for serial robots with flexible links

A methodology for the formulation of dynamic equations of motion of a serial flexible-link manipulator using the decoupled natural orthogonal complement (DeNOC) matrices, introduced elsewhere for rigid bodies, is presented in this paper. First, the Euler Lagrange (EL) equations of motion of the system are written. Then using the equivalence of EL and Newton–Euler (NE) equations, and the DeNOC matrices associated with the velocity constraints of the connecting bodies, the analytical and recursive expressions for the matrices and vectors appearing in the independent dynamic equations of motion are obtained. The analytical expressions allow one to obtain a recursive forward dynamics algorithm not only for rigid body manipulators, as reported earlier, but also for the flexible body manipulators. The proposed simulation algorithm for the flexible link robots is shown to be computationally more efficient and numerically more stable than other algorithms present in the literature. Simulations, using the proposed algorithm, for a two link arm with each link flexible and a Space Shuttle Remote Manipulator System (SSRMS) are presented. Numerical stability aspects of the algorithms are investigated using various criteria, namely, the zero eigenvalue phenomenon, energy drift method, etc. Numerical example of a SSRMS is taken up to show the efficiency and stability of the proposed algorithm. Physical interpretations of many terms associated with dynamic equations of flexible links, namely, the mass matrix of a composite flexible body, inertia wrench of a flexible link, etc. are also presented. The work has been carried out in the Dept. of Mechanical Engineering, Indian Institute of Technology Delhi, New Delhi 110016, India.     

Classification-based wheel slip detection and detector fusion for mobile robots on outdoor terrain

This paper introduces a signal-recognition based approach for detecting autonomous mobile robot immobilization on outdoor terrain. The technique utilizes a support vector machine classifier to form class boundaries in a feature space composed of statistics related to inertial and (optional) wheel speed measurements. The proposed algorithm is validated using experimental data collected with an autonomous robot operating in an outdoor environment. Additionally, two detector fusion techniques are proposed to combine the outputs of multiple immobilization detectors. One technique is proposed to minimize false immobilization detections. A second technique is proposed to increase overall detection accuracy while maintaining rapid detector response. The two fusion techniques are demonstrated experimentally using the detection algorithm proposed in this work and a dynamic model-based algorithm. It is shown that the proposed techniques can be used to rapidly and robustly detect mobile robot immobilization in outdoor environments, even in the absence of absolute position information.

Interacting and Annealing Particle Filters: Mathematics and a Recipe for Applications

Interacting and annealing are two powerful strategies that are applied in different areas of stochastic modelling and data analysis. Interacting particle systems approximate a distribution of interest by a finite number of particles where the particles interact between the time steps. In computer vision, they are commonly known as particle filters. Simulated annealing, on the other hand, is a global optimization method derived from statistical mechanics. A recent heuristic approach to fuse these two techniques for motion capturing has become known as annealed particle filter. In order to analyze these techniques, we rigorously derive in this paper two algorithms with annealing properties based on the mathematical theory of interacting particle systems. Convergence results and sufficient parameter restrictions enable us to point out limitations of the annealed particle filter. Moreover, we evaluate the impact of the parameters on the performance in various experiments, including the tracking of articulated bodies from noisy measurements. Our results provide a general guidance on suitable parameter choices for different applications.

Self-assembly and self-repair of arbitrary shapes by a swarm of reactive robots: algorithms and simulations

Self-assembly of active, robotic agents, rather than of passive agents such as molecules, is an emerging research field that is attracting increasing attention. Active self-assembly techniques are especially attractive at very small spatial scales, where alternative construction methods are unavailable or have severe limitations. Building nanostructures by using swarms of very simple nanorobots is a promising approach for manufacturing nanoscale devices and systems. The method described in this paper allows a group of simple, physically identical, identically programmed and reactive (i.e., stateless) agents to construct and repair polygonal approximations to arbitrary structures in the plane. The distributed algorithms presented here are tolerant of robot failures and of externally-induced disturbances. The structures are self-healing, and self-replicating in a weak sense. Their components can be re-used once the structures are no longer needed. A specification of vertices at relative positions, and the edges between them, is translated by a compiler into reactive rules for assembly agents. These rules lead to the construction and repair of the specified shape. Simulation results are presented, which validate the proposed algorithms.     

Generating B-spline curves with points, normals and curvature constraints: a constructive approach

This paper presents a constructive method for generating a uniform cubic B-spline curve interpolating a set of data points simultaneously controlled by normal and curvature constraints. By comparison, currently published methods have addressed one or two of those constraints (point, normal or cross-curvature interpolation), but not all three constraints simultaneously with C2 continuity. Combining these constraints provides better control of the generated curve in particular for feature curves on free-form surfaces. Our approach is local and provides exact interpolation of these constraints.     

A cohesive modeling technique for theoretical and experimental estimation of damping in serial robots with rigid and flexible links

Dynamic model incorporating damping characteristics, namely joint damping and structural damping in flexible links, of the serial robots with rigid and flexible links is presented. A novel procedure, based on the unified approach of theoretical formulation and analysis of experimental data, is proposed for the estimation of damping coefficients. First, the dynamic model of a robotic system with rigid and flexible links is presented. Next, the modifications in the dynamic model due to the considerations of damping characteristics of joints and structural damping characteristics of the flexible links are presented. A systematic methodology based on analysis of data obtained from experiments is presented for estimation and determination of damping coefficients of rigid-flexible links. The determination of joint damping coefficients, is based on the logarithmic decay of the amplitude of the oscillations of robotic links, while the structural damping coefficients are estimated mainly using the modal analysis and the method of evolving spectra. The method of evolving spectra, based on the Fast Fourier Transform of the decay of the amplitude in structural vibrations of the robot links in progressive windows is used to estimate the structural damping ratios while the critical structural coefficients are determined using the modal analysis. The methodology is illustrated through a series of simple experiments on simple robotic systems. The experimental results are then compared with the simulation results incorporating the damping coefficients determined using the proposed procedure. The comparisons leads to the validation of the proposed dynamic modeling technique, modeling of the damping characteristics, and the method proposed for estimation of damping coefficients for rigid-flexible link robotic systems.     

Fault detection in autonomous robots based on fault injection and learning

In this paper, we study a new approach to fault detection for autonomous robots. Our hypothesis is that hardware faults change the flow of sensory data and the actions performed by the control program. By detecting these changes, the presence of faults can be inferred. In order to test our hypothesis, we collect data from three different tasks performed by real robots. During a number of training runs, we record sensory data from the robots while they are operating normally and after a fault has been injected. We use back-propagation neural networks to synthesize fault detection components based on the data collected in the training runs. We evaluate the performance of the trained fault detectors in terms of number of false positives and time it takes to detect a fault. The results show that good fault detectors can be obtained. We extend the set of possible faults and go on to show that a single fault detector can be trained to detect several faults in both a robot’s sensors and actuators. We show that fault detectors can be synthesized that are robust to variations in the task, and we show how a fault detector can be trained to allow one robot to detect faults that occur in another robot.

A methodology for the generation of planar models for multibody chain drives

To overcome the difficulty on building manually complex models of chain drives, this work proposes a comprehensive methodology to build multibody models of any general chain drive automatically from a minimal set of data. The proposed procedure also evaluates the initial positions and velocities of all components of the drive that are consistent with the kinematic joints or with the contact pairs used in the model. In this methodology, all links and sprockets are represented by rigid bodies connected to each other either by ideal or by clearance revolute joints. The clearance revolute joint contact is further extended to handle the contact between the chain rollers and the sprocket teeth exact profiles. A suitable cylindrical continuous contact law is applied to describe the interaction on all contact pairs. One of the complexities of the computational study of roller chain drives is the large number of bodies in the system and the dynamics of the successive engagement and disengagement of the rollers with the sprockets. Each time a roller engages or disengages with a sprocket tooth, the number of rigid bodies in contact changes. The search for the contact pairs is recognized as one of the most time consuming task in contact analysis. This work proposes a procedure to specify the contact pairs and their update during the dynamic analysis optimizing the computational efficiency of the contact search. The methodologies adopted result in a general computer program that is applied and demonstrated in a generic chain drive that can be used in industrial machines, vehicle engines or any other type of mechanical system.     

Lower Bounds for Kernelizations and Other Preprocessing Procedures

We first present a method to rule out the existence of parameter non-increasing polynomial kernelizations of parameterized problems under the hypothesis P≠NP. This method is applicable, for example, to the problem Sat parameterized by the number of variables of the input formula. Then we obtain further improvements of corresponding results in (Bodlaender et al. in Lecture Notes in Computer Science, vol. 5125, pp. 563–574, Springer, Berlin, 2008; Fortnow and Santhanam in Proceedings of the 40th ACM Symposium on the Theory of Computing (STOC’08), ACM, New York, pp. 133–142, 2008) by refining the central lemma of their proof method, a lemma due to Fortnow and Santhanam. In particular, assuming that the polynomial hierarchy does not collapse to its third level, we show that every parameterized problem with a “linear OR” and with NP-hard underlying classical problem does not have polynomial self-reductions that assign to every instance x with parameter k an instance y with |y|=k ^O(1)⋅|x|^1−ε (here ε is any given real number greater than zero). We give various applications of these results. On the structural side we prove several results clarifying the relationship between the different notions of preprocessing procedures, namely the various notions of kernelizations, self-reductions and compressions.     

Branch and Recharge: Exact Algorithms for Generalized Domination

In this paper we present branching algorithms for infinite classes of problems.     

Feature-subspace aggregating: ensembles for stable and unstable learners

This paper introduces a new ensemble approach, Feature-Subspace Aggregating (Feating), which builds local models instead of global models. Feating is a generic ensemble approach that can enhance the predictive performance of both stable and unstable learners. In contrast, most existing ensemble approaches can improve the predictive performance of unstable learners only. Our analysis shows that the new approach reduces the execution time to generate a model in an ensemble through an increased level of localisation in Feating. Our empirical evaluation shows that Feating performs significantly better than Boosting, Random Subspace and Bagging in terms of predictive accuracy, when a stable learner SVM is used as the base learner. The speed up achieved by Feating makes feasible SVM ensembles that would otherwise be infeasible for large data sets. When SVM is the preferred base learner, we show that Feating SVM performs better than Boosting decision trees and Random Forests. We further demonstrate that Feating also substantially reduces the error of another stable learner, k-nearest neighbour, and an unstable learner, decision tree.     

Discretization for naive-Bayes learning: managing discretization bias and variance

Quantitative attributes are usually discretized in Naive-Bayes learning. We establish simple conditions under which discretization is equivalent to use of the true probability density function during naive-Bayes learning. The use of different discretization techniques can be expected to affect the classification bias and variance of generated naive-Bayes classifiers, effects we name discretization bias and variance. We argue that by properly managing discretization bias and variance, we can effectively reduce naive-Bayes classification error. In particular, we supply insights into managing discretization bias and variance by adjusting the number of intervals and the number of training instances contained in each interval. We accordingly propose proportional discretization and fixed frequency discretization, two efficient unsupervised discretization methods that are able to effectively manage discretization bias and variance. We evaluate our new techniques against four key discretization methods for naive-Bayes classifiers. The experimental results support our theoretical analyses by showing that with statistically significant frequency, naive-Bayes classifiers trained on data discretized by our new methods are able to achieve lower classification error than those trained on data discretized by current established discretization methods.     

Local and Nonlocal Discrete Regularization on Weighted Graphs for Image and Mesh Processing

We propose a discrete regularization framework on weighted graphs of arbitrary topology, which unifies local and nonlocal processing of images, meshes, and more generally discrete data. The approach considers the problem as a variational one, which consists in minimizing a weighted sum of two energy terms: a regularization one that uses the discrete p-Dirichlet form, and an approximation one. The proposed model is parametrized by the degree p of regularity, by the graph structure and by the weight function. The minimization solution leads to a family of simple linear and nonlinear processing methods. In particular, this family includes the exact expression or the discrete version of several neighborhood filters, such as the bilateral and the nonlocal means filter. In the context of images, local and nonlocal regularizations, based on the total variation models, are the continuous analog of the proposed model. Indirectly and naturally, it provides a discrete extension of these regularization methods for any discrete data or functions.