Experience-based imitation using RNNPB

Robot imitation is a useful and promising alternative to robot programming. Robot imitation involves two crucial issues. The first is how a robot can imitate a human whose physical structure and properties differ greatly from its own. The second is how the robot can generate various motions from finite programmable patterns (generalization). This paper describes a novel approach to robot imitation based on its own physical experiences. We considered the target task of moving an object on a table. For imitation, we focused on an active sensing process in which the robot acquires the relation between the object's motion and its own arm motion. For generalization, we applied the RNNPB (recurrent neural network with parametric bias) model to enable recognition/generation of imitation motions. The robot associates the arm motion which reproduces the observed object's motion presented by a human operator. Experimental results proved the generalization capability of our method, which enables the robot to imitate not only motion it has experienced, but also unknown motion through nonlinear combination of the experienced motions.     

Conceptual design and analysis of the 2T1R mechanism for a cooking robot

This paper deals with the design and analysis of a two-translation and one-rotation (2T1R) mechanism for a novel cooking robot. Firstly the motions involved in stir-fry, the most representative operation in the cooking processes used in Chinese cuisine, are analyzed in details. Then the featured motions are decomposed into four main movements that are used as a design base for a wok motion mechanism. Several three-degrees-of-freedom (DOF) parallel manipulators are considered. From these, a 2T1R mechanism is selected as an ideal candidate. A 4-DOF (2T1R+1T) cooking robot is constructed by combining the 2T1R parallel manipulator with a 1-DOF linear feed mechanism. It is shown that the combined 4-DOF robot can perform the required cooking operations, particularly the stir-fry. The analysis conducted on the proposed 2T1R parallel manipulator includes inverse kinematics, forward kinematics, the velocity analysis, the constant orientation workspace, and the total orientation workspace. A prototype of the cooking robot is developed. The experiments verify that the proposed cooking robot is suitable for performing the required operations.     

Pivoting motion control for a laparoscopic assistant robot and human clinical trials

This paper presents a compliant motion control method for the robotic assistant ERM (Endoscopic Robotic Manipulator), designed and developed by the authors for handling the camera in laparoscopic surgery. Since the robot has a passive wrist and it is not fixed to the operating table, the relative position between the robot camera holder and the insertion point is unknown. In this way, the proposed approach keeps the camera orientation according to the motion references in spite of this uncertainty and compensates for other unexpected disturbances about the relative robot–patient position. This system has been tested with live animals as well as in clinical trials on humans.     

Dynamic Rolling-Walk Motion by the Limb Mechanism Robot ASTERISK

New dynamic rolling-walk motion for a multi-legged robot with error compensation is proposed. The motion is realized by using the isotropic leg arrangement and the dynamic center of mass control inspired by bipedal robots. By using the preview control of the zero moment point (ZMP) with a cart-table model based on the bipedal robot's technique, the robot's center of mass trajectory is planned for the dynamic motion. The resolved momentum control for manipulating the multi-links robot as a single mass model is also implemented in the system to maintain the stability of the robot. In the new dynamic rolling-walk motion, the robot switches between the two-leg supporting phase and three-leg supporting phase to achieve dynamic motion with the preview control of the ZMP and resolved momentum control as dynamic motion controllers. The authors analyzed the motion and confirmed the feasibility in the Open Dynamics Engine before testing the motion with an actual robot. Due to the difficulties of controlling the ZMP during the two-leg supporting phase, the authors implemented error compensation by using a gyro sensor and compared the results.     

A one linear actuator hopping robot: modeling and control

This paper addresses the modeling and control design of a one linear actuator hopping robot. The robot consists of a body and a leg, which are in contact with a sufficiently wide horizontal ground surface; both are fixed rigidly. Force actuation affects the angle of the body by a force couple that arises due to the mass of the body, as well as the length of the leg; hence both the angular velocity of the body and height of a jump can be controlled by only one actuator. Since the aim of this study is to achieve continuous hopping motion while keeping the system as simple as possible, an ON-OFF actuator is employed. Hence, we consider utilizing the thrust timing of the actuator—when robot is in the stance phase—to control the gait. For better stability of the hopping motion, optimization of mechanical parameters was made possible by evaluating the numerically obtained transition map, which contains a transition from one standard position to the next. The system is considered as a discrete system, in which one cycle of motion is regarded as one sampling interval. Finally, a control system was designed in which, by simulation, the continuous hopping gait was realized.     

具有2DOF铰接式躯干的仿猎豹四足奔跑机器人

为了探索脊柱运动对腿运动的增强机理,设计了具有2自由度铰接式躯干的仿猎豹四足奔跑机器人。对带腾空相的跳跃（bound）步态奔跑运动的力学过程进行描述,采用阻尼型弹性负载倒立摆（D-SLIP）模型建立了四足机器人动力学模型。依据猎豹的奔跑运动模式,对四足机器人脊柱关节与腿关节的耦合运动进行了轨迹规划。提出一种改进的粒子群优化（PSO）算法,解决了机器人脊柱关节驱动机构尺寸和运动轨迹控制参数之间目标互斥的嵌套优化问题。对四足机器人跳跃奔跑运动进行动力学仿真,结果表明：脊柱与腿的协调运动可以增大奔跑步幅,使机器人产生腾空相,从而提高机器人的奔跑速度。     

Dynamic Analysis of a Nonholonomic Two-Wheeled Inverted Pendulum Robot

As a result of the increase in robots in various fields, the mechanical stability of specific robots has become an important subject of research. This study is concerned with the development of a two-wheeled inverted pendulum robot that can be applied to an intelligent, mobile home robot. This kind of robotic mechanism has an innately clumsy motion for stabilizing the robot's body posture. To analyze and execute this robotic mechanism, we investigated the exact dynamics of the mechanism with the aid of 3-DOF modeling. By using the governing equations of motion, we analyzed important issues in the dynamics of a situation with an inclined surface and also the effect of the turning motion on the stability of the robot. For the experiments, the mechanical robot was constructed with various sensors. Its application to a two-dimensional floor environment was confirmed by experiments on factors such as balancing, rectilinear motion, and spinning motion.     

Transporting an Object by a Passive Mobile Robot with Servo Brakes in Cooperation with a Human

In this paper, we introduce a passive mobile robot called PRP (Passive Robot Porter) to realize transportation of an object in cooperation with human, which is developed based on a concept of passive robotics. PRP consists of three omni-directional wheels with servo brakes and a controller. It can manipulate an object by controlling an external force/moment applied by a human based on the control of the servo brakes. We consider the characteristics of the servo brakes and control the brake torque of each wheel based on the brake force/moment constraint so that several motion functions of PRP are realized based on the applied force. This allows PRP to track a path which includes motion perpendicular to the pushing direction of the human without using servo motors. The impedance-based motion control is also realized with respect to the perpendicular to the pushing direction. These functions are implemented on PRP experimentally, and the experimental results illustrate the validity of PRP and its control method.     

Real-time arm motion imitation for human–robot tangible interface

In this paper we deal with a remote meeting system with tangible interface, in which a robot is used as tangible avatar instead of a remote meeting partner. For realizing such system, it is a critical issue how the robot imitates human motions with natural and exact. So, we suggested a new method that human arm motion is captured with a stereo vision system and transferred to the robotic avatar with real-time. For capturing 3D arm motions based on markerless method, we proposed a new metaball-based method which was designed in order to have some robust and efficient properties: a modified iso-surface equation of metaball for overcoming local minima and a downsizing method of 3D point cloud for improving time complexity. With our meeting system, we have implemented our new algorithm and run at approximately 12–16 Hz. Also, its accuracy in motion capturing could be acceptable for robot motion generation.     

Driving Mode Decision in the Obstacle Negotiation of a Variable Single-Tracked Robot

We propose a new algorithm for obstacle negotiation of a mobile robot known as a variable single-tracked robot (VSTR). The robot was originally designed for the purpose of rescue operations; hence, its ability to traverse various environments should be guaranteed. Previous experiments have already shown that the special structural characteristics of a VSTR enable it to easily climb indoor stairs. However, the VSTR has a critical drawback in that all of its functions are realized by an operator and the robot cannot detect obstacles or transform its driving mode by itself. For this reason, we have enhanced the system with the basis for autonomous navigation, i.e., the ability to negotiate obstacles, and we have introduced a methodology for practical use. Our experiments confirm that the driving mode decision, which is controlled by position-sensitive detector infrared sensors, gives decisive improvement. Moreover, the experimental results show that the robot can overcome several obstacles, thereby reflecting the benefits of the proposed algorithm.     

Virtual circle mapping for master–slave hand systems

The classical mapping methods were developed based on anthropomorphic robot hands. Normally, they are not applicable or cannot provide satisfactory performance if the robot hand is non-anthropomorphic. This paper presents a virtual circle mapping method dealing with the three-fingered non-anthropomorphic robot hand. The basic idea is to express the operator's motion by a virtual circle determined by the three fingertips. Four sets of parameters are used to describe the circle: circle radius, central angles, circle center and circle orientation. By transforming the four sets of parameters from the human frame to the robot frame, the information of relative positions between the fingertips is delivered. The robot fingertip positions are then computed according to the transformed parameters. The concept is introduced and implemented on a specific three-fingered non-anthropomorphic robot hand. The simulation results and the experiments have shown that the proposed method is able to obtain better workspace matching and, thus, the operator can tele-control the robot hand more intuitively.     

耦合型3自由度并联稳定平台机构及其运动特征

基于舰船摇荡运动的耦合特征,采用耦合型3-SRR/SRU 3自由度并联机构作为舰载稳定平台机构,实现了稳定平台三轴驱动、动平台五轴联动补偿的目标．运用螺旋理论分析了机构的自由度和运动特征, 并讨论了输入关节选取的合理性|采用坐标等效运动的方法,建立机构耦合运动约束方程,得到了位置解及工作空间．本文将耦合型3-SRR/SRU 3自由度并联机构用于舰载稳定平台五轴联动补偿,拓展了少自由度并联机构的应用领域．     

Gait motion for naturally curving variously shaped corners

Modeling the curving motion of humans in actual environment is rarely done because of the complexity and variability of the turning motion. In this study, various gait motions, including straight, round corner, and circular walks, were recorded and analyzed using factor analysis. As a result, we successfully extracted several factors that represent turning motions, such as long stride motion, turning motion led by the inner leg, and turning motion led by the outer leg. In particular, we found that the natural curving motion, which is a motion that results when turning around a round corner, is widely and continuously distributed on the factor space. Although several typical stepping strategies were reported by related studies, we found that the stepping motion changes between straight and turning gaits in the factor space during natural curving motions. Thus, the classification of curving motion into several typical distinct stepping patterns is probably insufficient to understand the natural curving motion. Furthermore, natural curving motions that comprise circular curving motions that were believed to represent typical curving motions was not validated. On the other hand, this result also suggests the possibility of generating curving motions for a physical assistant robot by combining straight gait and circler curving motion.     

Whole Body Motion Noise Cancellation of a Robot for Improved Automatic Speech Recognition

The motors of a robot produce ego-motion noise that degrades the quality of recorded sounds. This paper describes an architecture that enhances the capability of a robot to perform automatic speech recognition (ASR) even as the entire body of the robot moves. The architecture consists of three blocks: (i) a multichannel noise reduction block, consisting of microphone-array-based sound localization, geometric source separation and post-filtering, (ii) a single-channel template subtraction block and (iii) an ASR block. As the first step of our analysis strategy, we divided the whole-body motion noise problem into three subdomains of arm, leg and head motion noise, according to their intensity levels and spatial location. Subsequently, by following a synthesis-by-analysis approach, we determined the best method for suppressing each type of ego-motion noise. Finally, we proposed to utilize a control module in our ASR framework; this module was designed to make decisions based on instantaneously detected motions, allowing it to switch to the most appropriate method for the current type of noise. This proposed system resulted in improvements of up to 50 points in word correct rates compared with results obtained by single microphone recognition of arm, leg and head motions.     

Sensor system of a small biped entertainment robot

SDR-4X II is the latest prototype model of a small biped entertainment robot. It is the improved model of SDR-4X. In this paper we report on the sensing system of this robot, which is important and essential for a small biped entertainment robot which will be used in the home environment. One technology is the design of the motion sensing system, i.e. the inclination sensor system and the force sensor system which obtains the inclination of the trunk and the foot with force. Another technology is the real-world sensing system. One aspect is the touch sensing system. The robot is used in a normal home environment, so we should strongly consider the safety aspects for human. Another is the vision sensor system. The configuration and the distance image acquisition are explained. Next is the audio sensor system which obtains the sound and the voice information. The hardware system and the direction recognition are explained. These sensing systems are the key to making the biped robot walking and dynamic motion highly stable, and understanding the real-world around the robot.     

Internal Redundancy – the Way to Improve Robot Dynamics and Control Performances

The research follows the concept of variable geometry. The concept was originally introduced for general mechanical system in order to improve its dynamic behaviour [14]. Here, we apply the concept to robots. It can be shown that enhancement of robot dynamic performances is achieved. In this paper variation of geometry is considered equally as the motion in robot joints. Thus, a new set of degrees of freedom is introduced. This leads to redundancy – different internal motions are possible for the given external motion of the end-effector. However, there is an important difference from the usual notion of redundancy. Here, the additional joints do not influence the external motion and accordingly cannot improve the end-effector ability for maneuvering. For this reason the new notion is defined the internal redundancy.Although variation of geometry is treated equally with the motion in robot joints, there is still a difference in the aim of these new degrees of freedom. They should contribute to overcoming the limits of robot actuators, achieving better static compensation, etc. One might say that internal redundancy improves the robot dynamic capabilities. In this paper the mathematical formulation of kinematics and dynamics of robots with internal redundancy is carried out. A case study is presented in order to support the main idea.     

A generalized control framework of assistive controllers and its application to lower limb exoskeletons

Various control methods have been studied for the natural assistance of human motions by exoskeletal robots, i.e., wearable robots for assisting the human motions. For example, impedance control and compliance control are widely used for controlling interaction forces between a human and a robot. When an accurate measurement of the human muscular force is available (e.g., electromyography), a direct use of the estimated human joint torque is possible in the control of an assistive robot. The human motions in a daily living, however, are so complex that they are constituted by multiple phases, such as walking, sitting, and standing, where the walking can be further categorized into multiple sub-phases. Therefore, a single control method cannot be the best option for all the motion phases; a switch in the control algorithms may be necessary for assisting human movements in multiple motion phases. In this paper, a generalized control framework is proposed to incorporate the various assistive control methods in one general controller structure, which consists of Feedforward Disturbance Compensation Control, Reference Tracking Feedback Control, Reference Tracking Feedforward Control, Model-based Torque Control. The proposed control framework is designed taking into consideration of the linearity of each control algorithm, and thus it enables the continuous and smooth switching of assistive control algorithms, and makes it possible to analyze the stability of the overall control loop. The proposed method is implemented into a lower-limb exoskeleton robot and is verified by experimental results.     

一种五自由度视觉伺服机器人的跟踪控制研究

提出了一种基于图像的视觉伺服机器人系统的结构，以及该系统跟踪平面运动目标的算法. 该系统有5个自由度，包括一个3自由度的机器人和一个2自由度的手腕，手腕安装在机器人末端，而摄像机安装在手腕末端. 由于具有较小的运动惯量，手腕可快速旋转，以实现对运动目标的跟踪. 而在需要时，3自由度的机器人可将摄像机移动到适当的位置， 对目标进行更为仔细的观察. 该系统实现了对运动目标的跟踪. 此外，还提出了提高系统性能的方法，实验证实这些方法是有效的.     

A study on a brachiation controller for a multi-locomotion robot — realization of smooth,continuous brachiation

In this paper, we present a control method to realize smooth continuous brachiation. The target brachiation is basically divided into two actions: a swing action and a locomotion action. In order to realize the continuous brachiation effectively and smoothly, it is necessary to start the swing action as soon as the robot grasps the front target bar at the end of the locomotion action. The collision, which occurs at the moment the robot grips the target bar, affects the pendulum motion of the robot. The action of bending the elbow joint of the swinging arm is proposed in order to solve this gripping problem. The elbow-bending action enables the robot to decrease the impact forces and use the excess mechanical energy after the end of the locomotion phase. Thus, there is no loss of energy and waste of time during the subsequent swing phase. Experimental results show that the robot can successfully achieve smooth, continuous brachiation.     

An obstacle avoidance algorithm for robot manipulators based on decision-making force

Obstacle avoidance is a significant skill not only for mobile robots but also for robot manipulators working in unstructured environments. Various algorithms have been proposed to solve off-line planning and on-line adaption problems. However, it is still not able to ensure safety and flexibility in complex scenarios. In this paper, a novel obstacle avoidance algorithm is proposed to improve the robustness and flexibility. The method contains three components: A closed-loop control system is used to filter the preplanned trajectory and ensure the smoothness and stability of the robot motion; the dynamic repulsion field is adopted to fulfill the robot with primitive obstacle avoidance capability; to mimic human’s complex obstacle avoidance behavior and instant decision-making mechanism, a parametrized decision-making force is introduced to optimize all the feasible motions. The algorithms were implemented in planar and spatial robot manipulators. The comparative results show the robot can not only track the task trajectory smoothly but also avoid obstacles in different configurations.