This paper presents a method of planning a sub-optimal trajectory for a mobile manipulator subject to mechanical and control constraints. The path of the end-effector is defined as a curve that can be parameterised by any scaling parameter—the reference trajectory of a mobile platform is not needed. Constraints connected with the existence of mechanical limits for a given manipulator configuration, collision avoidance conditions and control constraints are considered. Nonholonomic constraints in a Pfaffian form are explicitly incorporated to the control algorithm. To avoid manipulator singularities, the motion of the robot is planned in order to maximise the manipulability measure.

In this paper, a rigid–flexible planar parallel manipulator (PPM) actuated by three linear ultrasonic motors for high-accuracy positioning is proposed. Based on the extended Hamilton's principle, a rigid–flexible dynamic model of the proposed PPM is developed utilizing exact boundary conditions. To derive an appropriate low-order dynamic model for the design of the controller, the assumed modes method is employed to discretize elastic motion. Then to investigate the interaction between the rigid and elastic motions, a proportional derivative feedback controller combined with a feed-forward-computed torque controller is developed to achieve motion tracking while attenuating the residual vibration. Then the controller is extended to incorporate an input shaper for the further suppression of residual vibration of flexible linkages. Computer simulations are presented as well as experimental results to verify the proposed dynamic model and controller. The input shaping method is verified to be effective in attenuating residual vibration in a highly coupled rigid–flexible PPM. The procedure employed for dynamic modeling and control analysis provides a valuable contribution into the vibration suppression of such a PPM.

Odometric error modelling for mobile robots is the basis of pose tracking. Without bounds the odometric accumulative error decreases localisation precision after long-range movement, which is often not capable of being compensated for in real time. Therefore, an efficient approach to odometric error modelling is proposed in regard to different drive type mobile robots. This method presents a hypothesis that the motion path approximates a circular arc. The approximate functional expressions between the control input of odometry and non-systematic error as well as systematic error derived from odometric error propagation law. Further an efficient algorithm of pose tracking is proposed for mobile robots, which is able to compensate for the non-systematic and systematic error in real time. These experiments denote that the odometric error modelling reduces the accumulative error of odometry efficiently and improves the specific localisation process significantly during autonomous navigation.

In this work, we investigate a quaternion-based formulation of 3D Simultaneous Localization and Mapping with Extended Kalman Filter (EKF-SLAM) using relative pose measurements. We introduce a discrete-time derivation that avoids the normalization problem that often arises when using unit quaternions in Kalman filter and we study its observability properties. The consistency of the estimation errors with the corresponding covariance matrices is also evaluated. The approach is further tested on real data from the Rawseeds dataset and it is applied within a delayed-state EKF architecture for estimating a dense 3D map of an unknown environment. The contribution is motivated by the possibility of abstracting multi-sensorial information in terms of relative pose measurements and for its straightforward extensions to the multi robot case.

Human rhythmic movement is generated by central pattern generators (CPGs), and their application to robot control has attracted interest of many scientists. But the coupling relationship between the central nervous system and the CPG network with external inputs is still not unveiled. According to biological experiment results, the CPG network is controlled by the neural system; in other words, the interaction between the central nervous system and the CPG network can control human movement effectively. This paper offers a complex human locomotion model, which illustrates the coupling relationship between the central nervous system and the CPG network with proprioception. Based on Matsuoka's CPG model (K. Matsuoka, Biol. Cybern. 52(6), 367–376 (1985)), the stability and robustness of the CPG network are analyzed with external inputs. In order to simulate the coupling relationship, the Radial Basis Function (RBF) neural network is used to simulate the cerebral cortex, and the Credit-Assignment Cerebellar Model Articulation Controller algorithm is employed to realize the locomotion mode conversion. A seven-link biped robot is chosen to simulate the walking gait. The main discoveries include: (1) the output of a new CPG network, which is stable and robust, can be treated as proprioception. Proprioception provides the central nervous system with the information about all joint angles; (2) analysis on a new locomotion model reveals that the cerebral cortex can modulate CPG parameters, leading to adjustment in walking gait.

Based on product of exponentials (POE) formula, three explicit error models are given in this paper for kinematic calibration of serial robot through measuring its end-effector positions. To obtain these error models, the tool frame should be chosen as reference frame at first, and then each position–error-related segment in the error models using pose measurement should be selected. And during kinematic parameter identification, all the errors in joint twists are identifiable, and the initial transformation errors and the joint zero-position errors can be identified conditionally. Namely, the initial transformation errors are identifiable if they do not contain orientation errors. And the joint zero-position errors are identifiable when a robot only consists of prismatic joints and the coordinates of its joint twists are linearly independent.

The effectiveness of this calibration method has been validated by simulations and experiments. The results show that: (1) the identification algorithms are robust and practical. (2) The method of position measurement is superior to that of pose measurement.

In this paper, we present a theoretical study on the control of a compass gait walker using energy regulation between steps. We use a return map to relate the mid-stance robot kinetic energy between steps with two control inputs, namely, foot placement and ankle push-off. We show that by regulating robot kinetic energy between steps using the two control inputs, we are able to (1) generate a wide range of walking speeds and stride lengths, including average human walking; (2) cancel the effect of external disturbance fully in a single step (dead-beat control); and (3) switch from one periodic gait to another in a single step. We hope that insights from this control methodology can help develop robust controllers for practical bipedal robots.

In order to resolve the redundancy of a wheeled mobile redundant manipulator comprising a two-wheel-drive mobile platform and a 6-degree-of-freedom manipulator, a physical-limits-constrained (PLC) minimum velocity norm (MVN) coordinating scheme (termed as PLC-MVN-C scheme) is proposed and investigated. Such a scheme can not only coordinate the mobile platform and the manipulator to fulfill the end-effector task and to achieve the desired optimal index (i.e., minimizing the norm of the rotational velocities of the wheels and the joint velocities of the manipulator) but also consider the physical limits of the robot (i.e., the joint-angle limits and joint-velocity limits of the manipulator as well as the rotational velocity limits of the wheels). The scheme is then reformulated as a quadratic program (QP) subject to equality and bound constraints, and is solved by a discrete QP solver, i.e., a numerical algorithm based on piecewise-linear projection equations (PLPE). Simulation results substantiate the efficacy and accuracy of such a PLC-MVN-C scheme and the corresponding discrete PLPE-based QP solver.

Accurate contour tracking is one of the main tasks in modern manufacturing processes. By considering coupling effects among multiple axes, this paper proposes a cross-coupled proportional-integral-derivative (PID) control developed in position domain, and the controller is applied to a multi-axis computer numerical control (CNC) machine for contour tracking performance improvement. Stability analysis is conducted for the developed position domain cross-coupled PID control using the Lyapunov method, and guidelines for the selection of control gains are provided. The contour tracking performance are improved compared to an equivalent time domain controller, since the reference axis in position domain control does not contribute any error to the overall contouring error of the system. Simulation results demonstrate the effectiveness of cross-coupled PID position domain control for both linear and circular contour tracking, and prove the robustness of the controller to deal with random disturbances. It also shows that position domain cross-coupled PID control provides better contour tracking performance over position domain PID control and the equivalent time domain PID control.

Parallel-jaw gripper finds wide applications in various industrial sectors. In this paper, we mainly focus on the problem of form closure caging grasps of polygons with a parallel-jaw gripper equipped with four fingers. The form closure caging grasp is helpful for the fingers placements and contact region selections of a pneumatic gripper, as it is less sensitive to fingers misplacements.

We firstly prove that there is always a path from a cage to a form closure grasp of the object that never breaks the cage, as long as the attractive region constructed in the configuration space has a local minimum. If such a minimum cannot be found, we further adjust the fingers arrangements to produce the form closure grasp. Meanwhile, we also develop an algorithm to compute the initial cage of the form closure grasp. Simulations of the grasping process witness the effectiveness of the above analysis results.