Search

Model-free reinforcement learning methods have successfully been applied to practical applications such as decision-making problems in Atari games. However, these methods have inherent shortcomings, such as a high variance and low sample efficiency. To improve the policy performance and sample efficiency of model-free reinforcement learning, we propose proximal policy optimization with model-based methods (PPOMM), a fusion method of both model-based and model-free reinforcement learning. PPOMM not only considers the information of past experience but also the prediction information of the future state. PPOMM adds the information of the next state to the objective function of the proximal policy optimization (PPO) algorithm through a model-based method. This method uses two components to optimize the policy: the error of PPO and the error of model-based reinforcement learning. We use the latter to optimize a latent transition model and predict the information of the next state. For most games, this method outperforms the state-of-the-art PPO algorithm when we evaluate across 49 Atari games in the Arcade Learning Environment (ALE). The experimental results show that PPOMM performs better or the same as the original algorithm in 33 games.

In this letter, we present an underactuated soft exoskeleton which can sequentially provide knee extension and ankle plantarflexion assistance for each leg with only one motor. The aim of this work is to assist the motions that have chronological moments at lower-limb joints and minimize the mass carried by the wearer. The underactuated soft exoskeleton consists of a novel actuation mechanism: Continuum-Cable System (CCS), which allows the actuators to transmit assistive torques to knee joint when the knee is flexed and transmit assistive torques to ankle joint when the knee is fully extended. Design concepts and component descriptions were elaborated for this underactuated soft exoskeleton, including kinematics, actuation, etc. The performance of the underactuated soft exoskeleton was evaluated by a series of experiments. With effective assistance from the proposed exoskeleton, the power of the knee was reduced by 6.40%, compared to without exoskeleton condition. Meanwhile, the moment and power of the ankle also demonstrated significant reduction when comparing the exoskeleton powered condition to the without exoskeleton condition (reduced by 11.61% and 45.44%, respectively). This underactuated soft exoskeleton represents a feasible solution to apply sequential assistance to the knee and ankle, and the experimental results validated its effectiveness in assisting the motions which have chronological moments at each lower-limb joint.

Leg stiffness is considered a prevalent parameter used in data analysis of leg locomotion during different gaits, such as walking, running, and hopping. Quantification of the change in support leg stiffness during stair ascent and descent will enhance our understanding of complex stair climbing gait dynamics. The purpose of this study is to investigate a methodology to estimate leg stiffness during stair climbing and subdivide the stair climbing gait cycle. Leg stiffness was determined as the ratio of changes in ground reaction force in the direction of the support leg F

Environmental information plays an important role in deep reinforcement learning (DRL). However, many algorithms do not pay much attention to environmental information. In multi-agent reinforcement learning decision-making, because agents need to make decisions combined with the information of other agents in the environment, this makes the environmental information more important. To prove the importance of environmental information, we added environmental information to the algorithm. We evaluated many algorithms on a challenging set of StarCraft II micromanagement tasks. Compared with the original algorithm, the standard deviation (except for the VDN algorithm) was smaller than that of the original algorithm, which shows that our algorithm has better stability. The average score of our algorithm was higher than that of the original algorithm (except for VDN and COMA), which shows that our work significantly outperforms existing multi-agent RL methods.

Two-dimensional lung ultrasound (LUS) has widely emerged as a rapid and non-invasive imaging tool for the detection and diagnosis of coronavirus disease 2019 (COVID-19). However, image differences will be magnified due to changes in ultrasound imaging experience, such as ultrasound probe attitude control and force control, which will directly affect the diagnosis results. In addition, the risk of virus transmission between sonographer and patients is increased due to frequent physical contact. In this study, a fully automatic dual-probe ultrasound scanning robot for the acquisition of lung ultrasound images is proposed and developd. Furthermore, the trajectory was optimized based on the velocity look-ahead strategy, the stability of contact force of the system and the scanning efficiency were improved by 24.13% and 29.46%, respectively. And the control ability of the contact force of robotic automatic scanning was 34.14 times higher than that of traditional manual scanning, which significantly improves the smoothness of scanning. Importantly, there was no significant difference in image quality obtained by robotic automatic scanning and manual scanning. Furthermore, the scanning time for a single person is less than 4 minutes, which greatly improves the efficiency of screening triage of group COVID-19 diagnosis and suspected patients, and reduces the risk of virus exposure and spread.

During COVID-19, with a heavy sprayer filled with disinfectant, the risk of infection for epidemic prevention personnel has been increased by long-term environmental disinfection. In order to reduce the burden and save energy of human, this letter proposed a Wheel-Legged Robotic Limb (WRL) for the carriers. The mass of WRL is only 1.77 kg. The WRL has one rigid robotic limb located below the sprayer, which can provide active supporting force for the sprayer. The WRL adopts force closed-loop control method to ensure the system provide an expected supporting force. The system performance was evaluated including standing and walking at 5 km/h, under three experimental conditions included: 1) with a sprayer, 19.41 kg (SPRAYER), 2) with the powered WRL, 22.18 kg (WRL_ON), and 3) with the unpowered WRL, 22.18 kg (WRL_OFF). When the supporting force is set as 80 N, the experimental results show that the WRL_ON condition has reduced the vertical load force on the human, the vertical ground reaction force of human feet, and the metabolic power by 41.28%, 8.03%, and 17.46% during standing, and also reduced by 32.29%, 8.08% and 18.92% during walking, compared to SPRAYERcondition, respectively.

Predicting the locomotion intent of humans is important for controlling assistive robots. Previous studies have investigated assistive walking on structured terrains, but only a few studies have considered rough terrains. Human intent on rough terrains is more difficult to predict because there is a transition at every step. To predict the foot placements of humans on rough terrains, the present paper fuses sequential 3D gaze and the environmental context. The 3D gaze is assumed to be the intersection point of the line of sight as measured by an eye-tracker and the environmental point cloud as measured by an RGBD camera. The sequential 3D gaze and the environmental context are fused based on an RGBD SLAM algorithm. Then the segmented terrain that is closest to the center of sequential 3D gaze is regarded as the most possible foothold area at the next step. Six able-bodied subjects are invited to walk randomly on rough terrains. Their foot placements are labeled and compared with the predicted foot placements. Experimental results show that the proposed method can predict the foot placements of all subjects 0.5 step ahead. With environmental context and user-dependent time window, the distance error of predicting the foot placements can decrease to 0.086 m. Hence, gaze, environmental context, and time window are all important in predicting the human intent when navigating rough terrains.

In production, poor ergonomic environments not only lead to increased workloads and health hazards for employees but also tend to reduce efficiency and quality. Recently, the human-cyber-physical (HCP) system has been proposed and widely studies to meet human capabilities and limitations. However, most existing frameworks are still not adaptive enough to integrate humans into a smart production environment due to a lack of real-time individual human factors digitalization. This research proposes an HCP workstation model for comprehensive assembly resources digitalization and autonomous interaction by CPS enabling technologies. Based on this, an adaptive control system has been developed for proactive ergonomic risk mitigation. Computer vision is deployed for real-time individual ergonomic evaluation and a prototype has been set up for functional validation.

Fuzzy inference systems have been widely applied in robotic control. Previous studies proposed various methods to tune the fuzzy rules and the parameters of the membership functions (MFs). Training the systems with only supervised learning requires a large amount of input-output data, and the performance of the trained system is confined by that of the target system. Training the systems with only reinforcement learning (RL) does not require prior knowledge but is time-consuming, and the initialization of the system remains a problem. In this paper, a supervised-reinforced successive training framework is proposed for a multi-continuous-output fuzzy inference system (MCOFIS). The parameters of the fuzzy inference system are first tuned by a limited number of input-output data from an existing controller with supervised training and then are utilized to initialize the system in the reinforcement training stage. The proposed framework is applied in a robotic odor source searching task and the evaluation results demonstrate that the performance of the fuzzy inference system trained by the successive framework is superior to the systems trained by only supervised learning or RL. The system trained by the proposed framework can achieve around a 10% higher success rate compared to the systems trained by only supervised learning or RL.

Powered lower-limb prostheses with vision sensors are expected to restore amputees' mobility in various environments with supervised learning-based environmental recognition. Due to the sim-to-real gap, such as real-world unstructured terrains and the perspective and performance limitations of vision sensor, simulated data cannot meet the requirement for supervised learning. To mitigate this gap, this paper presents an unsupervised sim-to-real adaptation method to accurately classify five common real-world (level ground, stair ascent, stair descent, ramp ascent and ramp descent) and assist amputee's terrain-adaptive locomotion. In this study, augmented simulated environments are generated from a virtual camera perspective to better simulate the real world. Then, unsupervised domain adaptation is incorporated to train the proposed adaptation network consisting of a feature extractor and two classifiers is trained on simulated data and unlabeled real-world data to minimize domain shift between source domain (simulation) and target domain (real world). To interpret the classification mechanism visually, essential features of different terrains extracted by the network are visualized. The classification results in walking experiments indicate that the average accuracy on eight subjects reaches (98.06% ± 0.71 %) and (95.91% ± 1.09 %) in indoor and outdoor environments respectively, which is close to the result of supervised learning using both type of labeled data (98.37% and 97.05%). The promising results demonstrate that the proposed method is expected to realize accurate real-world environmental classification and successful sim-to-real transfer.

Gait phase-based control is a trending research topic for walking-aid robots, especially robotic lower-limb prostheses. Gait phase estimation is a challenge for gait phase-based control. Previous researches used the integration or the differential of the human's thigh angle to estimate the gait phase, but accumulative measurement errors and noises can affect the estimation results. In this paper, a more robust gait phase estimation method is proposed using a unified form of piecewise monotonic gait phase-thigh angle models for various locomotion modes. The gait phase is estimated from only the thigh angle, which is a stable variable and avoids phase drifting. A Kalman filter-based smoother is designed to further suppress the mutations of the estimated gait phase. Based on the proposed gait phase estimation method, a gait phase-based joint angle tracking controller is designed for a transfemoral prosthesis. The proposed gait estimation method, the gait phase smoother, and the controller are evaluated through offline analysis on walking data in various locomotion modes. And the real-time performance of the gait phase-based controller is validated in an experiment on the transfemoral prosthesis.

Predicting the next foot placement of humans during walking can help improve compliant interactions between humans and walking aid robots. Previous studies have focused on foot placement estimation with wearable inertial sensors after heel-strike, but few have predicted foot placements in advance during the early swing phase. In this study, a Bayesian inference-based foot placement prediction approach was proposed. Possible foot placements were modeled as a probability distribution grid map. With selected foot motion feature events detected sequentially in the early swing phase, the foot placement probability map could be updated iteratively using the feature models we built. The weighted center of the probability distribution was regarded as the predicted foot placement. Prediction errors were evaluated with collected walking data sets. When testing with the data from inertial measurement units, the prediction errors were (5.46 cm ± 10.89 cm, -0.83 cm ± 10.56 cm) for cross-velocity walking data and (-4.99 cm ± 12.31 cm, -11.27 cm ± 7.74 cm) for cross-subject-cross-velocity walking data. The results were comparable to previous works yet the prediction could be made earlier. For the subject who walked with more stable gaits, the prediction error can be further decreased. The proposed foot placement prediction approach can be utilized to help walking aid robots adjust their pose before each heel-strike event during walking, which will make human-robot interactions more compliant. This study is also expected to inspire additional probabilistic gait analysis works.

Wearable robotic systems have been widely studied in recent years, but it still remains a challenge to design a user-adaptive controller for wearable robotic systems to ensure personalized and accurate human-robot interaction. Accurate human motion classification and person identification are two premises helping design user-adaptive controllers for wearable robotic systems. In this paper, we proposed a multi-task learning method for human motion classification and person identification with a single neural network, which can serve as a solution to personalized human-robot interaction, and can also serve as a benchmark for the following studies in related fields. The multi-task learning neural network was trained and tested on a public human motion data set. The proposed method was capable to classify human motions and identify the person, with 99.13% and 96.51% accuracy, respectively. We also compared the proposed method with a benchmark single task learning method for human motion classification, the results showed that the performance of the multi-task learning method is more superior.

Previous work has proved that carrying an elastically suspended load can improve the biomechanics of human body and reduce the burden of human body compared with carrying a rigid load. However, how to design and implement an effective elastically suspended load has been less researched. In this paper, we firstly analyze the relationship between the elastic load force acting on the human body and the parameters (stiffness and damping) of the elastic system to guide the design. Then, our prototype which is an effective elastically suspended back frame and can bind loads of various shapes is realized. The mass of the whole prototype is 2.66 kg. Experiments were carried to test the effect of the system. Results show that an elastically suspended back frame with a load of 25.3 kg could reduce the amplitude of load by 30.2%, reduce maximal load force acting on the carrier by 16.6% compared with the rigid load with backpack. Compared with carrying the rigid load with backpack, the mechanical work is reduced by 56.5% and the maximal mechanical power is also reduced by 66.1%.

Upper-Limb prosthesis control is a huge challenge for high-level amputees or amputated patients with weak residual muscles signal. Previous researches achieved the control of prosthesis by foot electromyography (EMG). However, low adaptability and gesture classification accuracy due to muscle movement and device limits restrict the performance. Therefore, this paper proposes a flexible high-density wearable device based on convolutional neural network for foot gestures recognition. The flexible wearable device stretches with muscle movement and makes the recognition process more accurate and efficient. Nine classes of foot gestures that intuitively map the movements of prosthesis are classified by the convolutional neural network classifiers. This paper reaches an average classification accuracy of 93.98% for nine classes of foot gestures. High-accuracy recognition based on the flexible wearable device provides a possibility for the control of upper-limb prosthesis.

In order to offer appropriate and reliable assistance to users, wearable robotic devices usually detect human locomotion through multi-sensor fusion system. However, multi-sensor fusion system increased the complexity of the sensor system and the burden of wearing on users for ankle exoskeleton. To optimize the sensor system and recognize multi-gait, we present a multi-gait recognition algorithm for a soft ankle exoskeleton with two IMUs (Inertial Measurement Units) mounted on foot. Five gait features are extracted during swing phase, including mean vertical velocity, mean horizontal velocity, vertical displacement, horizontal displacement, and the inclination angle at foot contact. Then, these gait features are used as the input of BPNN (Back Propagation Neural Network) to recognize five common gait modes (level walking, stair ascent/descent, ramp ascent/descent). The proposed algorithm can provide an accurate automatic recognition result at the early beginning of each stance phase. The results of the experiment shown that the proposed algorithm can distinguish above gait modes with 99.0% success rates.

This paper is intended to solve the motor skills learning, representation and generalization problems in robot imitation learning. To this end, we present an Adapted Curvilinear Gaussian Mixture Model (AdC-GMM), which is a general extension of the GMM. The proposed model can encode data more compactly. More critically, it is inherently suitable for representing data with strong non-linearity. To infer the parameters of this model, a Cross Entropy Optimization (CEO) algorithm is proposed, where the cross entropy loss of the training data is minimized. Compared with the traditional Expectation Maximization (EM) algorithm, the CEO can automatically infer the optimal number of components. Finally, the generalized trajectories are retrieved by an Adapted Curvilinear Gaussian Mixture Regression (AdC-GMR) model. To encode observations from different frames, the sophisticated task parameterization (TP) technique is introduced. All above proposed algorithms are verified by comprehensive tasks. The CEO is evaluated by a hand writing task. Another goal-directed reaching task is used to evaluate the AdC-GMM and AdC-GMR algorithm. A novel hammer-over-a-nail task is designed to verify the task parameterization technique. Experimental results demonstrate the proposed CEO is superior to the EM in terms of encoding accuracy and the AdC-GMM can achieve more compact representation by reducing the number of components by up to 50%. In addition, the trajectory retrieved by the AdC-GMR is smoother and the approximation error is comparable to the Gaussian process regression (GPR) even far fewer parameters need to be estimated. Because of this, the AdC-GMR is much faster than the GPR. Finally, simulation experiments on the hammer-over-a-nail task demonstrates the proposed methods can be deployed and used in real-world applications.

Planar surfaces are prevalent components of man-made indoor scenes, and plane extraction plays a vital role in practical applications of computer vision and robotics, such as scene understanding, and mobile manipulation. Nowadays, most plane extraction methods are based on reconstruction of the scene. In this paper, plane representation is formulated in inverse-depth images. Based on this representation, we explored the potential to extract planes in images directly. A fast plane extraction approach, which employs the region growing algorithm in inverse-depth images, is presented. This approach consists of two main components: seeding, and region growing. In the seeding component, seeds are carefully selected locally in grid cells to improve exploration efficiency. After seeding, each seed begins to grow into a continuous plane in succession. Both greedy policy and a normal coherence check are employed to find boundaries accurately. During growth, neighbor coplanar planes are checked and merged to overcome the over-segmentation problem. Through experiments on public datasets and generated saw-tooth images, the proposed approach achieves 80.2% CDR (Correct Detection Rate) on the ABW SegComp Dataset, which has proven that it has comparable performance with the state-of-the-art. The proposed approach runs at 5 Hz on typical 680 × 480 images, which has shown its potential in real-time practical applications in computer vision and robotics with further improvement.

Stair ascending is laborious, which imposes very high load on human knee, leading to high risks of knee injuries. Therefore, in this paper, a biologically-inspired soft exosuit for knee extension assistance during stair ascent was designed and the assistant effect of proposed exosuit was evaluated. We placed Bowden cables in front of thigh and shank to mimic quadriceps muscle, generating knee extension moments by length contraction. A force-based position control was implemented to deliver a knee assistant moment profile, which was based on biological knee moment during normal stair ascent. Biomechanical parameters of three experimental conditions (without exosuit, with unpowered exosuit, with powered exosuit) were compared. Experimental results showed that with powered exosuit lower-limb joint kinematics exhibited minimal changes, while net (muscles plus cables) knee moment decreased by 10.92% and net knee power decreased by 30.1% as compared to without exosuit condition.

Based on the fact that push-off force is the main energy input of human walking, reducing the push-off force when carrying the load is helpful to improve the walking endurance. Previous work has proved that an elastically suspended backpack could change ground reaction forces (GRFs) of legs, and parameters of elastically suspended backpack and walking speeds could affect the push-off force. When the walking speed changes, people hope the elastically suspended backpack to work at their best state by adjusting parameters. In this paper, a simple model is proposed to predict the effect on push-off force. Based on the results of model analysis, an elastically suspended backpack with tunable and appropriate stiffness is designed, which could optimize push-off force by adjusting the stiffness. Two healthy subjects were recruited to attend the experiment using our backpack prototype and the GRFs data was compared. Each subject walked under the multiple conditions (load states: no load, fixed load, elastic load with different stiffness; walking speed: 4.0 km/h, 4.5 km/h, 5.0 km/h, 5.5 km/h). The results show that the elastically suspended backpack could effectively change the stiffness and the effect of push-off force could be improve greatly by changing the stiffness.

Human activities recognition (HAR) and human intent recognition (HIR) are important for medical diagnosis and human-robot interaction. HAR and HIR usually rely on the signals of some wearable sensors, such as inertial measurement unit (IMU), but these signals may be user-dependent, which degrades the performance of the recognition algorithm on new subjects. Traditional supervised learning methods require labeling signals and training specific classifiers for each new subject, which is burdensome. To deal with this problem, this paper proposes a novel non-adversarial cross-subject adaptation method called Gaussian-guided feature alignment (GFA). The proposed GFA metric quantifies the discrepancy between the labeled features of source subjects and the unlabeled features of target subjects so that minimizing the GFA metric leads to the alignment of the source and target features. The GFA metric is estimated by calculating the divergence between the feature distribution and Gaussian distribution, as well as the mean squared error of the mean and variance between source and target features. This paper analytically proves the effect of the GFA metric and validates its performance using three public human activity datasets. Experimental results show that the proposed GFA achieves 1% higher target classification accuracy and 0.5% lower variance than state-of-the-art methods in case of cross-subject validation. These results indicate that the proposed GFA is feasible for improving the generalization of the HAR and HIR.

An increasing number of robotic systems involving lots of robotic individuals are used to serve human, such as intelligent terminal, intelligent storage, intelligence factories, etc. It is a trend of robotics technology that robotics system will become huger with more individuals. In these systems, they form the robotic societies and need establish some computing rules and mechanisms to ensure the operation like all biological social systems. In this paper, a novel system architecture for swarm robotic system, including three layers: human–computer interaction layer, planning layer and execution layer, is put forward, which is effective for task-oriented swarm robotic system. Then, a hierarchical organizational model for the system is presented, which is used to establish management relationship between different layers and individuals. Because task-oriented characteristic is required, this paper elaborates task description knowledge to explain the relationship between tasks for task decomposition and task logic. In addition, a method of behavior generation based on proposition/transition Petri networks is designed, which would effectively assist the system to construct combined behavior using simple individual behavior to solve a variety of tasks. At last, Illustration is shown to prove effectiveness and an implementation of the method based on SociBuilder system is introduced.

Random forest-based methods for 3D temporal tracking over an image sequence have gained increasing prominence in recent years. They do not require object&rsquo, s texture and only use the raw depth images and previous pose as input, which makes them especially suitable for textureless objects. These methods learn a built-in occlusion handling from predetermined occlusion patterns, which are not always able to model the real case. Besides, the input of random forest is mixed with more and more outliers as the occlusion deepens. In this paper, we propose an occlusion-aware framework capable of real-time and robust 3D pose tracking from RGB-D images. To this end, the proposed framework is anchored in the random forest-based learning strategy, referred to as RFtracker. We aim to enhance its performance from two aspects: integrated local refinement of random forest on one side, and online rendering based occlusion handling on the other. In order to eliminate the inconsistency between learning and prediction of RFtracker, a local refinement step is embedded to guide random forest towards the optimal regression. Furthermore, we present an online rendering-based occlusion handling to improve the robustness against dynamic occlusion. Meanwhile, a lightweight convolutional neural network-based motion-compensated (CMC) module is designed to cope with fast motion and inevitable physical delay caused by imaging frequency and data transmission. Finally, experiments show that our proposed framework can cope better with heavily-occluded scenes than RFtracker and preserve the real-time performance.

Space manipulators are mainly used in the spatial loading task. According to problems of the spatial loading diversity, the testing loading installing position, and the utilization ratio of a test platform, the space manipulator is asked to evaluate the position and attitude of itself. This paper proposes the Point Omnidirectional Coefficient (POC) with unit attitude sphere/circle to describe attitude of the end-effector, which evaluates any points in the attainable space of the manipulators, in combination with the manipulation’s position message, and get relationships between its position and attitude of all points in the attainable space. It represents the mapping between sphere surface and plane for mission attitude constraints and the method for calculating volume of points space including attainable space, Omnidirectional space, and mission attitude space. Furthermore, the Manipulator Omnidirectional Coefficient based on mission or not is proposed for evaluating manipulator performance. Through analysis and simulation about 3D and 2D manipulators, the results show that the above theoretical approach is feasible and the relationships about link lengths, joints angles, attainable space, and Manipulator Omnidirectional Coefficient are drawn for guiding design.

Navigation is one of basic functions of auto-mobile robots. After dozens of years of development, now the navigation in static environment almost has been realized, taking the method in ROS (i.e. Robot Operation System) navigation stack as an example. However, when cruising in dynamic environment, there are more difficulties, as the objects in the environment can change their positions. To deal with such cases, we propose a scheme to navigate using layered costmap. By predicting the dynamic object's encounter position according to both its and the robot's kinematic information, and put it in another layer, and set different costs around the objects according the estimation of the motion of them, then path based on the layered costmaps can be planned. In this paper, the safety and efficiency of our scheme have been proven by the results of simulation.

This paper mainly investigated the rigid-flexible dynamics simulation method of multi-body system. The 3-RPS parallel robot dynamics model is created by ADAMS (multi-body dynamics software) and ANSYS (finite element analysis software). In accordance with the flexible-body theory, we analyzed mechanical characteristics of parallel robot with no-load or full-load working condition, and got the deformation of end measuring point, maximum stress position and dynamics stress curve. The analysis method is more intuitional and accurate, and can increase the accuracy of dynamic response analysis of links under the dynamic loads. The simulation results create conditions for structure design and optimization of 3-RPS parallel robot.

The paper is focused on the ambiguity problem in pose estimation of circle feature and a new method is proposed based on the concentric circle constraint. The pose of a single circle feature, in general, can be determined from its projection in the image plane with a pre-calibrated camera, but there are generally two possible sets of pose parameters. By introducing the concentric circle constraint, interference of the false solution can be excluded. On the basis of element at infinity in projective geometry and the Euclidean distance invariant, cases that concentric circles are coplanar and non-coplanar are discussed respectively. Extensive experiments on these two cases are carried out to validate the proposed method.

More and more robotic systems with lots of robotic individuals severs for human, such as intelligent terminal, intelligent storage, intelligence factory, etc. This is the trend of robotics technology, which will lead robotics system become huger with more individuals. Then, how to organize and manage this huge robotic system will be one important issue. This paper proposes hierarchical self-organizing approach to realize self-management, self-organization. Firstly, hierarchical self-organizing model is put forward and the process of formation is described in detail, which makes the organizing structure of system regularly. Secondly, this paper uses the improved hormone-inspired model (IHM) to establish relation between individuals, which considers topological structure of the organization, supports dynamic reconfiguration and self-organization, and requires no globally certain identifiers for individual robots. Finally, this paper presents the experimental results on swarm robotics system with a large scale of individuals to form a self-organization.

Collision sensing including collision position, collision direction, and force size could make robots smoothly interact with environment, so that the robots can strongly adapt to the outside world. Skin sensor imitates principles of human skin using special material and physical structure to obtain collision information, but this method has some disadvantages, such as complex design, low sampling rate, and poor generality. In this paper, a new method using force/torque sensor to calculate collision position, collision direction, and force size is proposed. Detailed algorithm is elaborated based on physical principle and unified modeling method for basic geometric surface. Gravity compensation and dynamic compensation are also introduced for working manipulators/robots in gravity and dynamic environment. In addition, considering algorithm solvability and uniqueness, four constraints are proposed, which are force constraint, geometric constraint, normal vector constraint, and current mutation constraint. In order to solve conflict solution of algorithm in redundant constraints, compatibility solution analysis is proposed. Finally, a simulation experiment shows that the proposed method can achieve collision information efficiently and accurately.

With the further theory research and engineering application, the study of the small electro-mechanical systems, with movement conversion, motion transmission, force/torque transmission and sensor function, becomes popular. The new highly integrated mechanical and electrical integration linear output sensor system is designed in this paper, to solve the problems of movement conversion/transmission, force transmission/measurement etc, and further theory study is also presented. The function of sensing parameters and transmission parameters is deduced and the key parameters of system design and the constrained conditions are shown. Then, the result of simulated experiment that made in allusion to the model was compared with the theory speculation, accordingly, validated the accuracy of the theory.

According to the human playing of ping-pong racket, this paper originally analyzes the multi-fingered grasping and manipulating of ping-pong racket and points out the three aspects have to be studied to make sure the grasp is stable and the manipulation is dexterous. Two of those three aspects are studied respectively: the least needed number and the distribution of degrees of the multi-fingered hand is determined at first; then the kinematic planning for robotic hand grasping ping-pong racket is studied, and a new method is proposed to selected the right solution in the planning. The simulation result in the end proofs that the kinematic planning and the method are right and effective.