PU- and P*U*-equivalent parallel mechanisms (PMs) are critical families of PMs with one translational and two rotational (1T2R) degree of freedoms and have always been a research hotspot among lower mobility PMs. However, researches on these two families of PMs remain to be inadequate, and existing types are few and uncomprehensive. In this study, first, general wrench systems of PU-equivalent are derived based on virtual-chain approach and screw theory, revealing its constraint characteristics under general configuration. Cause of one parasitic motion is put forward and general wrench systems of P*U*-equivalent PMs with one parasitic motion are obtained. In addition, constraint analysis has been carried out to figure out constraint characteristics of P*U*-equivalent PMs with three parasitic motions. Second, branch chains are divided by generating constraints, and their structures are synthesized based on the presented rules. Then, the process for type synthesis of PU-equivalent PMs and P*U*-equivalent PMs and a series of novel 1T2R PMs are attained based on this. Finally, a novel PU-equivalent PM, 2PRU-PRUPc, and a novel P*U*-equivalent PM with one parasitic motion, 2PRU-PUU PM, are analyzed, demonstrating the effectiveness of the proposed type synthesis method.

In this paper, a sliding mode control using a control point concept is proposed for an under-actuated quadrotor. The proposed controller controls the position of the control point, a displaced point from the quadrotor’s geometric center, and the yaw angle. This method solves singularity issues in control matrix inversion and enables the utilization of the multi-input multi-output equation to derive the control inputs. The sliding surface is designed to control four outputs while stabilizing roll and pitch angles. Simulation and experimental results show the effectiveness and robustness of the proposed controller in the tracking of a trajectory under parametric uncertainties.

NASA has developed a free-flying camera known as AERCam to assist Space Shuttle and Space Station operations. The first AERCam vehicle was controlled from within the Space Shuttle using a hand-held controller. We have developed a system for controlling AERCam remotely from the ground, which involves significant communication delays. We have tested its use with NASA's AERCam dynamics simulation in place of the actual vehicle. Our ground system uses a predictive display1 that is based on a similar simulation of Shuttle and AERCam dynamics. For both physical and numeric reasons, the behavior of the remote vehicle may drift from what is expected by the predictive display system. We have developed a mechanism to periodically re-establish correspondence of the ground simulation with the remote vehicle using tracking data that reports AERCam's recent behavior. Our system also provides a basis for future development of more advanced autonomous ground control for space vehicles such as AERCam.

Cooperating multiple robots may experience slips between their end-effectors and the object when making no rigid grasps, as is the case of manipulating fingers. Consequently, these slips make the robotic system unsuccessful in executing a job properly and can even cause a failure of control. Contact-point slips can be avoided by defining multiple constraints on grasping forces and usable workspace, but these constraints limit the utilization of a robotic system. The researchers working on finger grasps and multiple robots cooperation with contact have always been assuming no slips at the contact points. However, in normal practice, one has to encounter this problem. Hence, this paper has taken this contact-point slip problem as its main topic and theme of work. Consequently, the slip phenomenon is analyzed thoroughly and then generalized for a multiple robot system manipulating a geometrically regular-shaped object. Different slip configurations are introduced and methods for their detection are proposed. This detection does not consider dedicated sensors neither for the object's posture nor for the slip detection. It, rather, utilizes only the finger-tip position information usually available from the kinematics based on very common sensors available with the manipulator. Utilizing the proposed methodology, it is possible to detect the end-point slips in most of the cases and realize an exact task execution and stability of control, with a minimum number of sensors.

The paper proposes a novel method to detect graspable handles for picking objects from a confined and cluttered space, such as the bins of a rack in a retail warehouse. The proposed method combines color and depth curvature information to create a Gaussian mixture model that can segment the target object from its background and imposes the geometrical constraints of a two-finger gripper to localize the graspable regions. This helps in overcoming the limitations of a poorly trained deep network object detector and provides a simple and efficient method for grasp pose detection that does not require a priori knowledge about object geometry and can be implemented online with near real-time performance. The efficacy of the proposed approach is demonstrated through simulation as well as real-world experiment.

The goal of few-shot semantic segmentation is to learn a segmentation model that can segment novel classes in queries when only a few annotated support examples are available. Due to large intra-class variations, the building of accurate semantic correlation remains a challenging job. Current methods typically use 4D kernels to learn the semantic correlation of feature maps. However, they still face the challenge of reducing the consumption of computation and memory while keeping the availability of correlations mined by their methods. In this paper, we propose the adaptively mining correlation network (AMCNet) to alleviate the aforementioned issues. The key points of AMCNet are the proposed adaptive separable 4D kernel and the learnable pyramid correlation module, which form the basic block for correlation encoder and provide a learnable concatenation operation over pyramid correlation tensors, respectively. Experiments on the PASCAL VOC 2012 dataset show that our AMCNet surpasses the state-of-the-art method by $0.7\%$ and $2.2\%$ on 1-shot and 5-shot segmentation scenarios, respectively.

Teleoperated grasping requires the abilities to follow the intended trajectory from the user and autonomously search for a suitable pre-grasp pose relative to the object of interest. Challenges include dealing with uncertainty due to the noise of the teleoperator, human elements and calibration errors in the sensors. To address these challenges, an effective and robust algorithm is introduced to assist grasping during teleoperation. Although without premature object contact or regrasping strategies, the algorithm enables the robot to perform online adjustments to reach a pre-grasp pose before final grasping. We use three infrared (IR) sensors that are mounted on the robot hand, and design an algorithm that controls the robot hand to grasp objects using the information from the sensors' readings and the interface component. Finally, a series of experiments demonstrate that the system is robust when grasping a wide range of objects and tracking slow-moving mobile objects. Empirical data from a five-subject user study allows us to tune the relative contributions from the IR sensors and the interface component so as to achieve a balance of grasp assistance and teleoperation.

A method that has been proposed for the design of a control system of a muti-link arm is the optimum approach applied to the system linearized around its nominal trajectory, i.e. the dynamic trajectory that the manipulator should follow. This method works, in as much as the initial deviation from the reference trajectory is small enough. When such is not the case, it is necessary to take into account the quadratic term in the approximation, and the paper proposes a small parameter technique to solve the corresponding optimum control problem which otherwise involves cumbersome calculations.

Part I of this paper was concerned with kinematic workspaces of walking machines, while this paper addresses the static workspaces of a walking machine and their graphical representation. The results of static analysis are presented; the static workspace constraints are established; an algorithm for investigation of static workspaces is presented; and the position static workspaces are analysed and graphically represented for an example walking machine design.

Hyper-redundant robots have good prospects for applications in confined space due to their high flexibility and slim body size. However, the super-redundant structure brings great challenges for its inverse kinematics with shape constraints. Unfortunately, traditional Jacobian pseudo-inverse-based inverse kinematics method and forward and backward reaching inverse kinematics (FABRIK) method are difficult to constrain the arm shape and realize trajectory tracking in confined spaces. To solve this problem, we propose a shape-controllable FABRIK method to satisfy the given path and shape constraints. Firstly, the kinematic model of the hyper-redundant robot is established, and the canonical FABRIK method is introduced. Based on the preliminary works, the single-layer improved FABRIK method is developed to solve the position and pointing inverse kinematics considering path environment and joint angle constraints instead of two-layer geometric iterations. For tracking the desired end roll angles, the polygonal virtual arm is designed. The real arm roll angle is achieved by controlling its winding on the virtual arm. In this way, the shape can be controlled. Finally, we compare the proposed method with other three approaches by simulations. Results show that the proposed method is more efficient and the arm shape is controllable.

This paper investigates the influence of adding an upper body to a bipedal robot on its stable walking behavior. The robot’s parts are mutually interconnected through an actuator network system. Therefore, the movement pattern of the upper body depends on the type of interactions created with other limbs. Throughout the experiments, various interactions among the different body parts were tested. The results showed that a robot with a motionless upper body exhibited unstable walking behavior. However, once the same upper body was involved and interacted properly, with other body parts, its movement significantly helped to stabilize the behavior of the robot.

We describe a sensor integration system which we have designed and implemented for a large fully-integrated flexible assembly machine. The machine was developed by a consortium to explore the design and implementation issues involved. The sensor integration system was designed for execution on a multiple transputer architecture, and co-ordinates all the sensory information in the machine. It uses the concept of virtual sensing to provide sensory data at an appropriate level of abstraction to the machine supervisor, which controls execution of the assembly tasks.

This paper discusses the problem of impact with robotic systems. The original method for the solution of impact is presented. The main idea is the replacement of impact with a singularity and hence the approach is called the IVSA (Impact-Via-Singularity-Analysis) Method. This goal is achieved by considering the obstacle as a unilateral constraint and introducing the new set of generalized coordinates so as to incorporate the constraint in the dynamic model. Using the IVSA Method the impact is not described by algebraic equations but by a reduced set of differential equations resulting directly from the initial dynamic model. The integration of dynamic equations over the impact points is thus possible. A numerical example is presented.

Zero moment point (ZMP) is the most popular concept that is applied to stabilize the gait motion of a biped robot. This paper utilizes ZMP with the augmented-reality (AR) method to improve the stability of gait motion of a biped robot. The 3ds Max computer software package is used to build a virtual robot. Under an achieved joint angle data of solid robot to produce an animation of the robot's trajectory, the joint angle data are transmitted to the virtual robot to analyze the offset of the trunk. Furthermore, this investigation adopts AR to allow the user to make direct comparisons between the solid and virtual robot before and after the gait motion is corrected. The animated trajectories of the virtual robot are compared and the relevant data provide feedback to the solid robot to adjust the joint angle and further correct its posture. The experimental results reveal that the proposed scheme can improve gait motion, even when the biped robot is affected by an unexpected loading disturbance. As well as improving the stability of gait motion of a biped robot, the results of this study can also be used to teach the application of the proposed method in a robotics class.

Increased speed of inverse dynamics computation is essential for improving the characteristics of robot control systems. This is achieved by reducing the numerical complexity of the models and by introducing parallelism in model computation. In this paper customized symbolic models with a near minimum numerical complexity will be used as a basis for the examination of parallelism in inverse dynamic robot models. A scheduling algorithm for the distribution of computational load onto an arbitrary linear array of pipelined processors will be developed. The proposed algorithm is experimentally evaluated on a transputer network.

We consider the problem of distance estimation in the context of a mobile camera, as, for example, when the camera is attached to a robot arm. It is natural in this case to consider stereo motion, and we present here the theoretical basis for an axial stereo method based on cross-correlation using the Fast Fourier Transform, which is appropriate in the context considered here. Finally, we describe the state of implementation of the system in our laboratory.