Evaluative techniques offer a tremendous potential for online controller design. However, when the optimization space is large and the performance metric is noisy, the overall adaptation process becomes extremely time consuming. Distributing the adaptation process reduces the required time and increases robustness to failure of individual agents. In this paper, we analyze the role of the four algorithmic parameters that determine the total evaluation time in a distributed implementation of a Particle Swarm Optimization (PSO) algorithm. For an obstacle avoidance case study using up to eight robots, we explore in simulation the lower boundaries of these parameters and propose a set of empirical guidelines for choosing their values. We then apply these guidelines to a real robot implementation and show that it is feasible to optimize 24 control parameters per robot within 2 h, a limited amount of time determined by the robots' battery life. We also show that a hybrid simulate-and-transfer approach coupled with a noise-resistant PSO algorithm can be used to further reduce experimental time as compared to a pure real-robot implementation.

This paper proposes an efficient implementation of a force-closure test for frictional three-finger grasps. The implementation is based on a condition that transforms force-closure testing into the problem of convex hull intersection in projective space. The proposed implementation further reduces the problem into the problem of computing whether a line segment intersects a convex hull of at most four points. Implementation results are presented along with a thorough performance analysis and comparison with several existing methods. The results are also verified with arbitrary precision floating point computation. This provides comparison of qualitative error resulting from floating point roundoff. The result shows that the proposed implementation outperforms other methods in terms of speed and precision.

In this paper, a vision-based scheme for the autonomous hovering of a miniature quad-rotor is developed. Cameras are used to estimate the position and the translational velocity of the vehicle. The dynamic model of the miniature quad-rotor is developed using the Newton–Euler approach. A nonlinear controller based on a separated saturation control strategy for a miniature quad-rotor is presented. To validate the theoretical results, an embedded control system for the miniature quad-rotor has been developed. Thus, the analytic results are supported by experimental tests. Experimental results have validated the proposed control strategy.

An assistive device for upper limb support was developed and evaluated in terms of usability, user satisfaction and motor performance on six end-users affected by neuro-motor disorders (three spinal cord injury; one multiple sclerosis; two Friedreich's ataxia). The system consisted of a lightweight 3-degrees-of-freedom robotic exoskeleton arm for weight relief, equipped with electromagnetic brakes. Users could autonomously control the brakes using a USB-button or residual electromyogram activations. The system functionally supported all of the potential users in performing reaching and drinking tasks. For three of them, time, smoothness, straightness and repeatability were also comparable to healthy subjects. An overall high level of usability (system usability score, median value of 90/100) and user satisfaction (Tele-healthcare Satisfaction Questionnaire - Wearable Technology, median value of 104/120) were obtained for all subjects.

This work aims to propose an innovative mechanism of human–robot collaboration (HRC) for mobile service robots in the application of elderly and disabled assistance. Previous studies on HRC mechanism usually focused on integrating decision-making intelligence of human beings by qualitative judgment and reasoning intelligence of robots by quantitative calculation. Instead, novelties of the proposed methodology include (1) constructing an HRC framework by taking reference from the Adaptive Control of Thought – Rational (ACT-R) human cognitive architecture; (2) establishing semantic webs of cognitive reasoning through human–robot interaction (HRI) and HRC to plan and implement complex tasks; and (3) realizing human–robot intelligence fusion by mutual encouragement, connect, and integration of modules of human, robot, perception, HRI, and HRC in the ACT-R architecture. Its technical feasibility is validated by some selected experiments within a “pouring” scenario. Further, although this study is oriented to mobile service robots, the modularized design of hardware and software makes its extensive use feasible in other types of service robots like smart rehabilitation beds, wheelchairs, and cleaning equipments.

Rapid and efficient dynamic stability control has been one of the important motivations in legged robot research, especially for legged robots running at high speed and/or on rough terrain. This paper presents a feasible control strategy, named Hybrid Feedback Control (HFC), for running systems based on the spring-loaded inverted pendulum principle (SLIP). The HFC strategy, which comprises two modules (i.e., touchdown angle control and energy compensation), predicts and regulates touchdown angle of the current cycle and need-to-complement energy input of the next cycle through hybrid feedback of flying apex state. This strategy can significantly reduce the computational complexity and enable the system to quickly converge to its control target, meeting the requirements of real-time control. Simulation experiments on various terrains were conducted to verify the adaptability of our HFC strategy. Results of these simulation experiments show that the approach herein can realize the periodical stability control of SLIP systems on different terrain conditions quickly and effectively.

Our goal is to introduce a more appropriate method of representing, generalising and comparing gaits; particularly, walking gait. Human walking gaits are a result of complex, interdependent factors that include variations resulting from embodiments, environment and tasks, making techniques that use average template frameworks suboptimal for systematic analysis or corrective interventions. The proposed work aims to devise methodologies for being able to represent gaits and gait transitions such that optimal policies that eliminate the inter-personal variations from tasks and embodiments may be recovered. Our approach is built upon (i) work in the domain of nullspace policy recovery and (ii) previous work in generalisation for point-to-point movements. The problem is formalised using a walking-phase model, and the nullspace learning method is used to generalise a consistent policy from multiple observations with rich variations. Once recovered, the underlying policies (mapped to different gait phases) can serve as reference guideline to quantify and identify pathological gaits while being robust against interpersonal and task variations. To validate our methods, we have demonstrated robustness of our method with simulated sagittal two-link gait data with multiple ground truth constraints and policies. Pathological gait identification was then tested on real-world human gait data with induced gait abnormality, with the proposed method showing significant robustness to variations in speed and embodiment compared to template-based methods. Future work will extend this to kinetic features and higher dimensional features.

The use of large groups of robots in the execution of complex tasks has received much attention in recent years. Generally called robotic swarms, these systems employ a large number of simple agents to perform different types of tasks. A basic requirement for most robotic swarms is the ability for safe navigation in shared environments. Particularly, two desired behaviors are to keep robots close to their kin and to avoid merging with distinct groups. These are respectively called cohesion and segregation, which are observed in several biological systems. In this paper, we investigate two different approaches that allow swarms of robots to navigate in a cohesive fashion while being segregated from other groups of agents. Our first approach is based on artificial potential fields and hierarchical abstractions. However, this method has one drawback: It needs a central entity which is able to communicate with all robots. To cope with this problem, we introduce a distributed mechanism that combines hierarchical abstractions, flocking behaviors, and an efficient collision avoidance mechanism. We perform simulated and real experiments to study the feasibility and effectiveness of our methods. Results show that both approaches ensure cohesion and segregation during swarm navigation.

This paper deals with the problem of the practical tracking control of an experimental car-like system called the Robucar. The car-like Robucar is a four-wheeled car in a single steering mode. Based on a kinematic model of the car-like Robucar, a practical tracking controller is designed using the second-order sliding mode control of the super twisting algorithm. Hence, the output tracking of the desired trajectory is achieved, and the tracking errors vanish asymptotically. Experimental tests on the car-like Robucar are presented for simple and real-time nonholonomic trajectories, and comparative results with the conventional sliding controller demonstrate the applicability and efficiency of the proposed controller.

A non-overconstrained three-DOF parallel orientation mechanism that is kinematically equivalent to the Agile Eye is presented in this paper. The output link (end-effector) of the mechanism is connected to the base by one spherical joint and by another three identical legs. Each leg comprises of, in turns from base, a revolute joint, a universal joint, and three prismatic joints. The three lower revolute joints are active joints, while all other joints are passive ones. Based on a special configuration, some three projective angles of the end-effector coordinates are fully decoupled with respect to the input actuated joints, that is, by actuating any revolute joint the end-effector rotates in such a way that the corresponding projective angle changes with the same angular displacement. The fully decoupled motion is analyzed geometrically and proved theoretically. Besides, the inverse and direct kinematics solutions of the mechanism are provided based on the geometric reasoning and theoretical proof.

A robot mapping procedure using a modified speeded-up robust feature (SURF) is proposed for building persistent maps with visual landmarks in robot simultaneous localization and mapping (SLAM). SURFs are scale-invariant features that automatically recover the scale and orientation of image features in different scenes. However, the SURF method is not originally designed for applications in dynamic environments. The repeatability of the detected SURFs will be reduced owing to the dynamic effect. This study investigated and modified SURF algorithms to improve robustness in representing visual landmarks in robot SLAM systems. Many modifications of the SURF algorithms are proposed in this study including the orientation representation of features, the vector dimension of feature description, and the number of detected features in an image. The concept of sparse representation is also used to describe the environmental map and to reduce the computational complexity when using extended Kalman filter (EKF) for state estimation. Effective procedures of data association and map management for SURFs in SLAM are also designed to improve accuracy in robot state estimation. Experimental works were performed on an actual system with binocular vision sensors to validate the feasibility and effectiveness of the proposed algorithms. The experimental examples include the evaluation of state estimation using EKF SLAM and the implementation of indoor SLAM. In the experiments, the performance of the modified SURF algorithms was compared with the original SURF algorithms. The experimental results confirm that the modified SURF provides better repeatability and better robustness for representing the landmarks in visual SLAM systems.

The virtual decomposition control (VDC) is an efficient tool suitable to deal with the full-dynamics-based control problem of complex robots. However, the regressor-based adaptive control used by VDC to control every subsystem and to estimate the unknown parameters demands specific knowledge about the system physics. Therefore, in this paper, we focus on reorganizing the equation of the VDC for a serial chain manipulator using the adaptive function approximation technique (FAT) without needing specific system physics. The dynamic matrices of the dynamic equation of every subsystem (e.g. link and joint) are approximated by orthogonal functions due to the minimum approximation errors produced. The control, the virtual stability of every subsystem and the stability of the entire robotic system are proved in this work. Then the computational complexity of the FAT is compared with the regressor-based approach. Despite the apparent advantage of the FAT in avoiding the regressor matrix, its computational complexity can result in difficulties in the implementation because of the representation of the dynamic matrices of the link subsystem by two large sparse matrices. In effect, the FAT-based adaptive VDC requires further work for improving the representation of the dynamic matrices of the target subsystem. Two case studies are simulated by Matlab/Simulink: a 2-R manipulator and a 6-DOF planar biped robot for verification purposes.

This study deals with a self-reconfiguration problem of hexagonal-shaped modules from an arbitrary initial configuration to a straight chain. Modules are modeled as the same-sized rigid bodies. Two categories of modules with different functionalities are used. One category comprises two powerful modules, which are expected to play the role of terminal modules in a goal configuration. The other category comprises several ordinary modules, which are expected to fill in the middle portion in a goal configuration. A distributed control strategy, inspired by the idea of curve shortening, is developed for each module to act cooperatively to attain a goal configuration. It is verified that under the proposed strategy, modules eventually converge to a straight chain.

We consider the problem of dynamic reconfiguration by modular self-reconfigurable robots (MSRs) in the presence of uncertainty in their motion and the environment. Specifically, we consider the situation where the MSR is unable to continue its motion in its current configuration and needs to identify a new configuration among the existing modules, which would be the most configuration suitable for performing the robot's assigned task under the current circumstances. To address this problem, we propose a new data structure called an uncertain coalition structure graph (UCSG) that accommodates uncertainty in the MSR's motion and the environment, using a framework from cooperative game theory called the coalition structure graph. We then propose a new search algorithm called searchUCSG that intelligently prunes nodes from the UCSG using a modified branch-and-bound technique. We have shown analytically that our algorithm is anytime, that is, if it terminates arbitrarily, it returns the best solution found thus far, which is guaranteed to be within a constant bound from the optimal solution. We have verified the performance of our algorithm experimentally in simulation and shown that it is able to find a solution that is within the worst bound of 80% of the optimal solution while exploring only half of the nodes in the UCSG. Our algorithm also takes lesser computation time than the existing algorithms (that do not model uncertainty) for solving similar problems. Finally, to verify the operation of our algorithm, we have implemented it to partition a set of mobile e-puck robots into clusters and shown how different number of robots and different robot motion uncertainty parameters affect the formed clusters.

Biped robots have gained much attention for decades. A variety of researches have been conducted to make them able to assist or even substitute for humans in performing special tasks. In addition, studying biped robots is important in order to understand human locomotion and to develop and improve control strategies for prosthetic and orthotic limbs. This paper discusses the main challenges encountered in the design of biped robots, such as modeling, stability and their walking patterns. The subject is difficult to deal with because the biped mechanism intervenes with mechanics, control, electronics and artificial intelligence. In this paper, we collect and introduce a systematic discussion of modeling, walking pattern generators and stability for a biped robot.

In this paper an optimal path planning method based on a new evolutionary algorithm is presented for higher order robotic systems. It is a combination of immune system and wavelet mutation. By increasing the system's dimensions, the complexity of algorithm grows linearly. The obtained results have been compared with other optimal path producing algorithms, and its excellence in terms of optimality has been proved. Strengths of this method are simplicity in large-scale path planning, being free of most of the common deadlocks in usual method, and ability to obtain more optimized results than other similar methods. The effectiveness of this approach on simulation case studies for a three-link planar robot and 5 degrees of freedom mobile manipulators as well as an experiment for a mobile robot called K-joniour is shown.

This study presents novel robotic therapy control algorithms for upper-limb rehabilitation, using newly developed passive and progressive resistance therapy modes. A fuzzy-logic based proportional-integral-derivative (PID) position control strategy, integrating a patient's biomechanical feedback into the control loop, is proposed for passive movements. This allows the robot to smoothly stretch the impaired limb through increasingly rigorous training trajectories. A fuzzy adaptive impedance force controller is addressed in the progressive resistance muscle strength training and the adaptive resistive force is generated according to the impaired limb's muscle strength recovery level, characterized by the online estimated impaired limb's bio-damping and bio-stiffness. The proposed methods are verified with a custom constructed therapeutic robot system featuring a Barrett WAM™ compliant manipulator. Twenty-four recruited stroke subjects were randomly allocated in experimental and control groups and enrolled in a 20-week rehabilitation training program. Preliminary results show that the proposed therapy control strategies can not only improve the impaired limb's joint range of motion but also enhance its muscle strength.

In order to obtain a more human-like walking and less energy consumption, a it foot rotation phase is considered in the single support phase of a 3D biped robot, in which the stance heel lifts from the ground and the stance foot rotates about the toe. Since there is no actuation at the toe, a walking phase of the robot is composed of a fully actuated phase and an under-actuated phase. The objective of this paper is to present an asymptotically stable walking controller that integrates these two phases. To get around the under-actuation issue, a strict monotonic parameter of the robot is used to describe the reference trajectory instead of using the time parameter. The overall control law consists of a zero moment point (ZMP) controller, a swing ankle rotation controller and a partial joint angles controller. The ZMP controller guarantees that the ZMP follows the desired ZMP. The swing ankle rotation controller assures a flat-foot impact at the end of the swinging phase. Each of these controllers creates two constraints on joint accelerations. In order to determine all the desired joint accelerations from the control law, a partial joint angles controller is implemented. A word “partial” emphasizes the fact that not all the joint angles can be controlled. The outputs controlled by a partial joint angles controller are defined as a linear combination of all the joint angles. The most important question addressed in this paper is how this linear combination can be defined in order to ensure walking stability. The stability of the walking gait under closed-loop control is evaluated with the linearization of the restricted Poincaré map of the hybrid zero dynamics. Finally, simulation results validate the effectiveness of the control law even in presence of initial errors and modelling errors.

Singularity-free workspace is a very important criterion for the design of manipulators, especially for parallel manipulators which are well known for their limited workspace and complex singularities. This paper studies geometric parameters and dexterity measures that affect the size of a singularity-free joint space and proposes methods for the development of 6-DOF Stewart–Gough parallel manipulators that have better singularity-free joint space. With a local dexterity measure as the objective function, a systematic method is employed to search for the design with a maximal singularity-free joint space. The related workspaces are also investigated. It is shown that the workspace is not proportional to the size of the joint space and that manipulators with a larger singularity-free workspace usually have relatively poor dexterity.

It is important to develop a robotic orthosis or exoskeleton that can provide back-drivable and good assistive performances with lightweight structures for overground gait rehabilitation of stroke patients. In this paper, we describe a robotic knee device with a five-bar linkage to allow low-impedance voluntary knee motion within a specified rotation range during the swing phase, and to assist knee extension during the stance phase. The device can provide free motion through the five-bar linkage with 2-degree-of-freedom (DOF) actuation via the patient's shank using a linear actuator, and can assist knee extension at any controlled knee angle while bearing weight via a geared five-bar linkage with 1 DOF actuation of the linear actuator. The kinematic transition between the two modes can be implemented by contact with a circular structure and a linear link, and the resultant range of motion can be determined by the linear actuator. The kinematic weight of the device was optimized using the simple genetic algorithm to reduce the mass. The optimization cost function was based on the sum of the total link lengths and the actuator power. The optimization results reduced the total link length and motor power by 47% and 43%, respectively, compared to the initial design. We expect that the device will facilitate rehabilitation of stroke patients by allowing safe and free overground walking while providing support for stumbling.