Highly dynamic movements such as jumping are important to improve the agility and environmental adaptation of humanoid robots. This article proposes an online optimization method to realize a vertical jump with centroidal angular momentum (CAM) control and landing impact absorption for a humanoid robot. First, the robot's center of mass (CoM) trajectory is generated by nonlinear optimization. Then, a quasi-sliding mode controller is designed to ensure that the robot tracks the CoM trajectory accurately. To avoid unexpected spinning in the flight phase, a center-of-pressure-guided angular momentum controller is designed to stabilize the CAM. The modifications of CoM and CAM are realized by online optimization of dynamic components and inverse dynamics. Two quadratic programming optimizations are utilized to generate feasible contact force/torque and joint acceleration referring to uplevel CoM and CAM controllers. In addition, a viscoelastic model-based controller is designed to absorb the vibration caused by a large contact impact. A simulation and experiment of a 0.5-m high (foot lifting distance) vertical jump are achieved on a humanoid robot platform in this article (Fig. 1).

The current system design approaches for humanoid robots are mostly platform-based, making it difficult to consider the operational scenarios comprehensively. This leads to challenging modifications to the complex humanoid robot before meeting real-world task requirements. To boost the effectiveness and efficiency of task-oriented humanoid robot design, a novel robot design method based on Model-Based Systems Engineering (MBSE) is proposed. Firstly, a top-down model-based humanoid robot development method is presented, which integrates the development process and model to design and describe the requirements, functions, the logical architecture and the physical architecture of the humanoid robot. Then, the development of the search and rescue humanoid robot is taken as an example to illustrate the effectiveness of the method.

Long-span bridges are the lifeline throats of urban transportation network. Deflection (i.e., deformation) behavior of long-span bridges is complex. It can be found from long-term monitoring data that there is an obvious time-lag effect between the quasi-static behavior of deflection and environmental temperature, and abnormal signals, such as drift and jump-point, appear sporadically in the deflection data. In order to deal with the interference from the data time lag and abnormal signal, this article adopts the Bayesian multiple linear regression (BMLR) method to establish the mathematical model of bridge deflection based on temperature, other points’ deflection, or cable force data. A new paradigm of the recursive modeling strategy of BMLR for bridge deflection based on short-term data is proposed, which truly realizes the dynamic update ability of Bayes’ theorem in multiple regression modeling. Under the same conditions of modeling, the proposed paradigm performs higher accuracy of prediction and lower space of data storage occupied than the traditional multiple linear regression method and is less time taken than methods of deep learning. The whole process was validated to be robust to the data time lag and abnormal signal. When faced with the situation of sparse sensing points and not enough long time of monitoring, it is possible to fast predict deflection of new-added/adjusted sensing points using short-term observation data.

The high stiffness actuator (HSA), applied to each joint of an electrical driven humanoid robot, can directly affect the motion performance of the torque-controlled humanoid robots. For high control performance of HSA, a high-precision dynamic torque control (HDTC) is proposed. The HDTC consists of two phases: (1) A novel dynamic current control is used to linearize high stiffness actuator torque control system, which can estimate and compensate the nonlinear coupling parts; (2) An enhanced internal model control is designed to ensure high tracking accuracy in the system containing noisy torque signal and even numerical differentiation signals. Benefitting from dynamic current control and the enhanced internal model control, the proposed HDTC is accurate and adaptable. Finally, the superiority of the HDTC is verified with comparative experiments.

The biped robot adjusts stepping variables (i.e., stride length and stepping duration) to achieve robust walking under perturbations. A novel method of integrating center of mass (CoM) stabilization control into the optimization of stepping variables was proposed in this article. The method contains offline and online procedures. In the offline procedure, benchmark CoM stabilization controller coefficients and benchmark transition matrices were derived. In the online procedure, the states in subsequent steps were predicted in real time by transition matrix operations, and stepping variables were optimized by model predictive control (MPC). After MPC, the target stepping variables were input to CoM stabilization controllers to generate stable CoM motion. The combination of stepping adjustment and CoM stabilization control was accomplished with the method, and the advantages are presented in the introduction and demonstrations.

To face the challenge of adapting to complex terrains and environments, we develop a novel wheel-legged robot that can switch motion modes to adapt to different environments. The robot can perform efficient and stable upright balanced locomotion on flat roads and flexible crawling in low and narrow passages. For passing through low and narrow passages, we propose a crawling motion control strategy and methods for transitioning between locomotion modes of wheel-legged robots. In practical applications, the smooth transition between the two motion modes is challenging. By optimizing the gravity work of the body, the optimal trajectory of the center of mass (CoM) for the transition from standing to crawling is obtained. By constructing and solving an optimization problem regarding the posture and motion trajectories of the underactuated model, the robot achieves a smooth transition from crawling to standing. In experiments, the wheel-legged robot successfully transitioned between the crawling mode and the upright balanced moving mode and flexibly passed a low and narrow passage. Consequently, the effectiveness of the control strategies and algorithms proposed in this paper are verified by experiments.

In this paper, a parallel quadrupedal robot was designed that is capable of versatile dynamic locomotion and perception-less terrain adaptation. Firstly, a quadrupedal robot with a symmetric legs and a powerful actuator was implemented for highly dynamic movement. Then, a fast and reliable method based on generalized least square was proposed for estimating the terrain parameters by fusing the body, leg, and contact information. On the basis of virtual model control (VMC) with the quadratic program (QP) method, the optimal foot force for terrain adaptation was achieved. Finally, the results obtained by simulation and indoor and outdoor experiments demonstrate that the robot can achieve a robust and versatile dynamic locomotion on uneven terrain, and the rejection of disturbances is reliable, which proves the effectiveness and robustness of this proposed method.

Knee-stretched walking is considered to be a human-like and energy-efficient gait. The strategy of extending legs to obtain vertical center of mass trajectory is commonly used to avoid the problem of singularities in knee-stretched gait generation. However, knee-stretched gait generation utilizing this strategy with toe-off and heel-strike has kinematics conflicts at transition moments between single support and double support phases. In this article, a knee-stretched walking generation with toe-off and heel-strike for the position-controlled humanoid robot has been proposed. The position constraints of center of mass have been considered in the gait generation to avoid the kinematics conflicts based on model predictive control. The method has been verified in simulation and validated in experiment.

This paper proposes a vertical jump optimization strategy for a one-legged robot with consideration of its variable reduction ratio joints. Firstly, the characteristic of the joint is derived to obtain its influence on jump motion, which is similar to the reduction ratio. Secondly, referring to the joint’s characteristic, the initial posture of jumping is optimized to maximize the initial acceleration of jumping. Then, to generate the trajectory of the center of mass (CoM) and make the jump motion more efficient, nonlinear optimization of CoM is adopted with respect to human jumping data. Full-body dynamics is considered to track the trajectory with virtual force control. For flight phase, joint PD controller is adopted to decelerate and maintain the posture. A contrast simulation is implemented to demonstrate the characteristics of the variable reduction ratio joint. Vertical jump experiment on a one-legged robot platform is realized with a height of 30 cm.

A method for the force-free control of humanoid robot joints is presented that meets human–machine cooperation functions of a humanoid robot. To ensure accurate torque control, the current loop of the permanent magnet synchronous motor was initially optimised and its dynamic response improved with a compensating back electromotive force. The motor speed is then estimated by sampling the voltage and current, the values of which are used in the calculation of the joint angle and the gravitational-force compensation. Finally, a method to evaluate the dynamic force compensation is applied that then yields the motor output offset, the total gravitational load, the partial inertia force, and the partial friction moment. Experimental results show that this method reduces the drag torque to less than 1 Nm. This method can be widely applied to a variety of robotic joints.

The disturbance rejection performance of a biped robot when walking has long been a focus of roboticists in their attempts to improve robots. There are many traditional stabilizing control methods, such as modifying foot placements and the target zero moment point (ZMP), e.g., in model ZMP control. The disturbance rejection control method in the forward direction of the biped robot is an important technology, whether it comes from the inertia generated by walking or from external forces. The first step in solving the instability of the humanoid robot is to add the ability to dynamically adjust posture when the robot is standing still. The control method based on the model ZMP control is among the main methods of disturbance rejection for biped robots. We use the state-of-the-art deep-reinforcement-learning algorithm combined with model ZMP control in simulating the balance experiment of the cart–table model and the disturbance rejection experiment of the ASIMO humanoid robot standing still. Results show that our proposed method effectively reduces the probability of falling when the biped robot is subjected to an external force in the x-direction.

Accurate prediction/forecasting of the future response of civil infrastructure plays an essential role in health monitoring and safety assessment. However, the complex latent dynamics within the field sensing measurements makes the forecasting task challenging. To this end, this paper leverages the recent advances in deep learning and proposes a spatiotemporal learning framework to forecast structural responses with strong temporal dependencies and spatial correlations. The key concept is to establish a convolutional long-short term memory (ConvLSTM) network to learn spatiotemporal latent features from data and thus establish a surrogate model for structural response forecasting. The proposed approach is applied to predict the strain response for a concrete bridge with over three-year measurements available. A comparative study is also conducted against a traditional temporal-only network to highlight the forecasting performance of the proposed approach. Convincing results demonstrate that the ConvLSTM approach is a promising, reliable, and computationally efficient approach that is capable of accurately forecasting the dynamical response of civil infrastructure in a data-driven manner. (C) 2021 American Society of Civil Engineers.

In this paper, a jumping control method is proposed as based on Virtual Model Control (VMC). The virtual force and torque acting on the trunk are used to control the jumping height, speed, and posture. VMC is used to calculate the driving torque of joints to exert the desired virtual force and torque. Each jumping cycle is divided into three states: flying in the air, landing and balance control, and thrusting. Then, a finite state machine (FSM) is adopted to switch between each jumping states. The proposed continuous jumping method is verified through dynamic simulations whose results are discussed also to characterize the jumping performance.

Compliance is important for humanoid robots, especially a position-controlled one, to perform tasks in complicated environments where unexpected or sudden contacts will result in large impacts which may cause instability or destroy the hardware of robots. This paper presents a compliance control method based on viscoelastic model for humanoid robots to survive on these conditions. The viscoelastic model is used to obtain the relationship between the differential of contact force/torque and linear/angular position. Thus a state equation of this model can be established and a state feedback controller adjusting the position to adapt to the contact force/torque can be designed to realize the compliant movement. The proposed compliance control method based on viscoelastic model has been employed in ankle compliance for stable walking on indefinite uneven terrain and arm compliance for falling protection on BHR-6P, a position-controlled humanoid robot, which validates its effectiveness.

Most existing motion control methods for humanoids aim at avoiding falling. However, the humanoid is generally an unstable system that cannot completely avoid falling and it is difficult to cope with the sudden fall of a robot. This paper designs a planning method of fall protection for humanoids according to the human falling motion. This method changes the impact position between the robot and ground by adjusting the motion of the robot as it falls. To further reduce damage to the robot, an appropriate cushioning material is installed at the point of impact to buffer the robot. The effectiveness of the proposed method is verified for a BHR6P humanoid robot falling in simulations and experiments.

Based on the passive walking robots, a new quasi-passive walking robot based on impulse thrust is studied. Firstly, we establish a simple quasi-passive walking robot model. The dynamic equation and the transition equation of the simplest two-link model are analyzed by using angular momentum theorem. The existence and stability of the fixed points are analyzed by using Poincare map. Secondly, we use MATLAB to simulate the walking robot, calculate the fixed points of the quasi-passive walking robot, and analyze the stability of the system through the eigenvalue of Jacobian of the linearized map. The feasibility of impulse thrust and telescopic leg is verified.

At present, further reducing the mass of permanent magnet brushless dc motor has become a new demand for the development of humanoid robots. In order to achieve this goal, this paper presents a new design idea of motor. Compared with the motor currently used in robots, the new method will reduce the mass of the motor and improve the torque density as well, while ensuring the same torque output. This design method mainly uses HALBACH array for excitation. Through finite element analysis, the distribution rules of magnetic field, back electromotive force, torque and cogging torque of the motor are obtained. Experimental result shows that the actual torque of the motor reaches the goal of our design, and the rotor weight is reduced by 30%.

It is a fundamental issue that the dexterous multifingered hands grasp three-dimensional(3-D) objects, in the field of robotic grasping and dexterous manipulation. The dexterous hands need to grasp objects fast and stable, like the hands of humans. This paper presents a strategy for the stable and fast grasp. First, the grasp progress is divided into two stages including pre-grasp and grasp, which can work with a model not accurate. The algorithms of dimension reduction, Q-distance quality and simulated annealing are used together at the pre-grasp to simplify the calculation and the pre-grasp process. Finally, the effectiveness of the proposed methods is validated under simulations and in experiments with an dexterous multifingered hand.

Humanoid robots are being designed to perform tasks currently carried out by human workers in industry, manufacturing, service, and disaster assistance. To this end, the humanoid robot should be able to walk stably across many types of terrain. However, when traversing a complex unknown environment, it is difficult to realize accurate terrain perception immediately through large data collected by the sensor system, leading to a difference between planned foot landing positions and actual foot landing positions. As a result, an unexpected contact force/torque may affect the stability of the robot. This paper adopts active contact perception instead of terrain perception and proposes a contact force/torque control method based on the viscoelastic model to address this problem. In addition, we design a body stability controller based on tracking the trajectories of the virtual repellent point (VRP) and the divergent component of motion (DCM) to restrain the disturbance caused by the unexpected contact force/torque. Simulations and experiments on the BHR-6P humanoid robot platform demonstrate the proposed contact force/torque control method for walking on indefinite uneven terrain. Note to Practitioners This paper presents a contact force/torque controller based on a viscoelastic model, which unifies contact perception and adaptive reaction. Using the viscoelastic model, we get the relationship between the end position/posture and the first-order differential of the contact force/torque, from which a state equation can be established and used to design a state feedback controller. Combining with body stabilization, we adopt this controller to modify the trajectory of both landing and support feet simultaneously for humanoid robots to realize stable bipedal walking on indefinite uneven terrain, which is validated in both simulations and experiments on the BHR-6P humanoid robot. In addition, this method can be employed for other robots or equipment when the contact force/torque control is needed; e.g., industry manipulators and quadruped robots.

Linear Hall-effect sensors are integrated into permanent magnet motors and provide positional feedback with the advantages of compact size and low cost. Field-oriented control (FOC) is widely used in high-performance situations yet limited by its computational cost. This paper proposes a novel simplified FOC that uses the linear Hall outputs directly to obtain the coefficients required to perform coordinate transformations through linear combination; neither the computation of trigonometric functions nor the flux position estimations are required. The method is computationally efficient because it involves less latency and fewer hardware resources, when implemented on digital controllers. Furthermore, the method has been used for sensor delay compensation, which has been verified as important to eliminate direct current. The experimental results validate the feasibility and effectiveness of the proposed method with a 126 W motor under various speed and load conditions.