Impedance control makes the manipulator compliant by regulating the dynamic behavior between the environment and itself. As contact-based applications become more complex, variable impedance control has attracted more attention. However, variable impedance control can cause instability when directly switching impedance parameters, which is a conflict between maintaining the passivity and varying impedance parameters. To solve this conflict, a novel passivity-constrained variable impedance controller for hydraulic manipulators is proposed within the framework of model predictive control. With the impedance relationship and the manipulator dynamics integrated into a prediction model, the proposed strategy utilizes the passivity constraint and the impedance error cost to achieve an optimized tradeoff between passivity guarantee and impedance variation. Furthermore, the inner-loop motion of the hydraulic manipulator is controlled by the hierarchical decoupling controller with the manipulator dynamics and the nonlinearity of the valve-controlled cylinder system considered. The effectiveness of this method was verified by experiments on a heavy-duty hydraulic manipulator. Compared with the traditional method, the proposed method can improve the passivity while ensuring the timely modulation of impedance parameters.

The large-scale flexible hydraulic manipulator (FHM) is the key machinery for automation in construction. It generally works in a precise position and suffers from serious vibrations caused by external excitation. Thus, it requires both active damping control and position control, which are typically implemented by the single valve system (SVS) in the past. However, the SVS only has one signal input which cannot address the two controllers simultaneously. To solve the problem, a decoupling active damping controller utilizing the independent metering system (IMS) is presented. To reduce the influence between controls, the control degree of freedom is increased by breaking the mechanical coupling of the inlet and outlet. The relative gain array (RGA) method is then used to determine the best variable pair. Thus, the position and active damping controllers can be designed separately in two control loops to achieve control decoupling. To achieve active damping control, the dynamic pressure feedback (DPF) based on a high-pass filter is introduced to optimize system damping. The proposed method is verified on a concrete pump truck simulation model which is a typical large-scale flexible hydraulic manipulator. Simulation results show that the proposed method reduces interactions between different control loops and has a gentler vibration of the end-effector compared to the conventional SVS.

The large-scale flexible hydraulic manipulator (FHM) is the key machineryfor automation in construction. It generally works in a precise positionand suffers from serious vibrations caused by external excitation. Thus, itrequires both active damping control and position control, which are typicallyimplemented by the single valve system (SVS) in the past. However, theSVS only has one signal input which cannot address the two controllerssimultaneously. To solve the problem, a decoupling active damping controllerutilizing the independent metering system (IMS) is presented. To reduce theinfluence between controls, the control degree of freedom is increased bybreaking the mechanical coupling of the inlet and outlet. The relative gainarray (RGA) method is then used to determine the best variable pair. Thus,the position and active damping controllers can be designed separately intwo control loops to achieve control decoupling. To achieve active dampingcontrol, the dynamic pressure feedback (DPF) based on a high-pass filter isintroduced to optimize system damping. The proposed method is verifiedon a concrete pump truck simulation model which is a typical large-scaleflexible hydraulic manipulator. Simulation results show that the proposedmethod reduces interactions between different control loops and has a gentlervibration of the end-effector compared to the conventional SVS

Force monitoring or feedback on the end-effector is a key component for hydraulic manipulators to perform heavy-duty tasks such as handling and dismantling. However, the force sensors used in heavy-duty applications are expensive and fragile under complicated conditions. As an alternative, soft measurement methods are used to obtain the external force indirectly, but this is challenging for multi-DOF hydraulic manipulators due to the complex parallel-serial characteristics of the closed-chain structures with highly nonlinear frictions. To overcome the obstacles, a novel force soft measurement method for hydraulic manipulators is proposed using the cylinder decoupling model with LuGre friction compensation. In this method, the hydraulic manipulator is decoupled by the main chain consisting of manipulator links and several branch chains consisting of hydraulic cylinders, of which the inertia effect can be fully considered and modeled. The friction in each joint is also compensated by the LuGre friction model. Besides, the lumped inertia parameters and the LuGre friction parameters are acquired through the stepwise parameter identification method, which mitigates the complexity of the full model by first determining steady friction parameters and then solving the remaining parameters with different excitation trajectories. Compared with the traditional method without considering the closed-chain structures, the proposed force soft measurement method shows higher force estimation accuracy, which is conducive to the implementation of force-based strategies.

The dynamics of hydraulic robots are complicated due to the closed-chain joints formed by cylinder articulation. This article is focused on presenting a model-based control framework for rapid locomotion, integrating closed-chain dynamics without a substantial increase in computational costs. The virtual decomposition control (VDC) approach has been adapted and innovatively extended to a leg system for the first time, featuring a floating base and variable contact constraints. In this article, a position control framework is proposed, consisting of three VDC-based controllers designed specifically for the stance phase and the swing phase, respectively. During the stance phase, a constrained estimation model is developed to recursively compute the previously incalculable dynamic equations. Furthermore, the control laws are designed to ensure that the virtual power flows caused by contact constraints do not affect the stability. In the swing phase, a noninertial frame is established to transform the underactuated system into a fully actuated fixed-base system. Despite being position controlled, our framework enables the leg system to generate compliance by setting a separate low-gain VDC-based controller during the landing stage. Experiments reveal that the proposed framework exhibits better position trajectory tracking performance and jumping ability in highly dynamic motion compared with the state-of-the-art position controller. Additionally, the impedance characteristics of the leg system can be actively adjusted to adapt to uneven terrain.

Hydraulic manipulators are favored because of their high power density and strong explosive force. As more tasks interact with the environment, better force/motion control performance is demanded. However, hydraulic manipulators are usually driven by closed kinematic loops containing linear hydraulic actuators with passive resolute joints, leading to the difficulty in full dynamic modeling and parameter identification for accurate model-based force/motion control. To overcome this problem, a cylinder separated model is proposed with hydraulic cylinders separated virtually through the concept of virtual equivalent rotational joints. The decoupling model describes the complex dynamics of closed-chain structures based on the Lagrange equation, which is used to identify both inertial parameters of rigid links and cylinders in addition. Besides, a hierarchical decoupling controller is designed by splitting hydraulic cylinder control from the manipulator dynamics on the basis of the decoupling model, and then, the control problems of structure mechanisms and hydraulic system are handled independently. Based on the proposed dynamic model and control strategy, contact force/motion control is realized with the contact force estimated from cylinder pressure sensors. Experiments reveal that the position error of the end effector is reduced at least by 12.7% compared with the conventional controller that ignores coupling problems, whereas the root mean square error of contact force is within 100 N. The proposed strategy shows satisfied performance for hydraulic manipulators, which is also convenient to implement.

Galloping cheetahs, climbing mountain goats, and load hauling horses all show desirable locomotion capability, which motivates the development of quadruped robots. Among various quadruped robots, hydraulically driven quadruped robots show great potential in unstructured environments due to their discrete landing positions and large payloads. As the most critical movement unit of a quadruped robot, the limb leg unit (LLU) directly affects movement speed and reliability, and requires a compact and lightweight design. Inspired by the dexterous skeleton-muscle systems of cheetahs and humans, this paper proposes a highly integrated bionic actuator system for a better dynamic performance of an LLU. We propose that a cylinder barrel with multiple element interfaces and internal smooth channels is realized using metal additive manufacturing, and hybrid lattice structures are introduced into the lightweight design of the piston rod. In addition, additive manufacturing and topology optimization are incorporated to reduce the redundant material of the structural parts of the LLU. The mechanical properties of the actuator system are verified by numerical simulation and experiments, and the power density of the actuators is far greater than that of cheetah muscle. The mass of the optimized LLU is reduced by 24.5%, and the optimized LLU shows better response time performance when given a step signal, and presents a good trajectory tracking ability with the increase in motion frequency.

Contact detection is the key part in the unified event-based motion control framework of hydraulic quadruped robots. Without the use of an external force sensor, this process usually requires estimating the contact force, which is then compared with a certain set threshold to judge whether contact occurs. But there are some challenges when applied to hydraulic robots, due to the inherent complexity of the hydraulic system, such as high-order, non-linearity, strong interference, etc. The typical threshold method uses only single force information, which is easy to make incorrect inferences when the feedback hydraulic signal has a large unknown perturbation. What’s more, the dynamic model of a hydraulic robot is more complex and difficult to establish accurately, making some ways of estimating force based on complete dynamics inapplicable. To tackle such problems, this paper introduces the method of probabilistic model fusion for contact estimation, and applies it to our hydraulic single-leg platform. As far as we know, it may be the first time that this probability method is applied to a hydraulic legged robot. The method fuses multiple available information and makes optimal estimation, which can reduce the misjudgment caused by the interference of single information. Moreover, a simplified method for estimating contact force through impedance relationship is proposed. Since the method does not require complex dynamic modeling and torque feedback, it is computationally simple and effective and suitable for hydraulic robots. The above theory is tested in a general framework combined with an event-based state machine, and the experimental results verify the effectiveness. The work of this paper can provide a basis for the quadruped robot to cross the unstructured terrain without external perception in the follow-up work.

Nowadays, hydraulic quadruped robot shows high power density, good impact resistance and robustness in the research. The controller is the key to realize these features. This paper shows the design of an open-source single-leg controller for the hydraulic quadruped robot Spurlos using a distributed control scheme. The single-leg system of the hydraulic quadruped robot Spurlos contains three angle encoders, three servo valves and six pressure sensors, which has the same components as most single-leg systems. Through the chips designed in the controller, the signal can be received from the encoders and the sensors, meanwhile the signal can be delivered to the servo valves. The software part of the controller adopts the MBD (Model-Based Design) method, which can greatly improve the development efficiency. According to the experiments, the controller design is reasonable, stable operation, and can satisfy the requirements of the hydraulic quadruped robot for leg motion control. The controller designed in this paper provides a solution to the problem that there is no ready-made control board for hydraulic quadruped robot which have three degrees of freedom for each leg. It enables the control researches for hydraulic quadruped robots to be more easily implemented.

High power density is a development trend in legged robots. However, the huge weight and size of the hydraulic power unit is a major obstacle to improving the performance of robots, and heat dissipation is also an important issue for the power unit. As the core power component of the legged robot, the integration of the motor pump unit can reduce the volume and weight. This paper designs a coaxial motor pump unit with embedded cooling channels, which has the performance advantages of small envelope size, light weight, and high efficiency. Aiming at the heat dissipation problem of the motor pump, a spiral cooling flow channel is designed in the motor pump housing, and the temperature field inside the housing is simulated to verify that the oil cooling flow channel can realize the heat dissipation function.

Electro-hydraulic control valves are key hydraulic components for industrial applications and aerospace, which controls electro-hydraulic motion. With the development of automation, digital technology, and communication technology, electro-hydraulic control valves are becoming more digital, integrated, and intelligent in order to meet the requirements of Industry 4.0. This paper reviews the state of the art development for electro-hydraulic control valves and their related technologies. This review paper considers three aspects of state acquisition through sensors or indirect acquisition technologies, control strategies along with digital controllers and novel valves, and online maintenance through data interaction and fault diagnosis. The main features and development trends of electro-hydraulic control valves oriented to Industry 4.0 are discussed.

在光纤陀螺稳定控制平台应用领域中,光纤陀螺相位滞后及瞬态噪声严重制约着随动控制系统的控制品质.为了提高基于光纤陀螺随动系统的控制效果,研究了基于增益自补偿的光纤陀螺瞬态噪声抑制方法.首先,分析了由于死区补偿带来的光纤陀螺瞬态噪声,在此基础上研究了基于增益自补偿的光纤陀螺瞬态噪声抑制方法,并对该方法进行了理论分析.根据稳定控制平台对相位滞后和瞬态噪声的设定要求,通过一只数字闭环光纤陀螺进行了测试验证,测试结果满足系统对相位滞后和瞬态噪声的指标要求,瞬态噪声峰峰值为0.36(°)/s,并不随增益的变化而变化,验证了增益自补偿方法的有效性.

Rotary modulation and control algorithms are the key technologies of the rotating strapdown inertial navigation system. The complex rotation scheme earns limited accuracy, while causing larger size, lower reliability and higher additional cost. This paper analyzed the sources of inertial sensors' errors, compared similarities and differences, and evaluated the performance of all kinds of single-axis rotation modulation schemes. Initial alignment requires transposition in high accuracy and the rotary motor needs to stay a steady pace in actual practice under multiple interferences, so this paper designed a dual closed-loop controller for all concerned, with an ordinary PID in the position loop and ADRC controller in the speed loop, ensuring the modulation runs smoothly. The simulation results show that the system has some good characteristics such as fast response, no overshoot and good robustness. Compared with the original PID control, the new control algorithm helps rotary motor improve its anti-jamming performance.

Among many auxiliary particle filtering algorithms, the ensemble particle filter (EnPF) has been widely applied in various fields, which can be ascribed to its avoidance of the heavy computational burden resulted by the parallel computation of multifilters. However, the estimation precision of the EnPF, like other auxiliary particle filters, also decreases when the nonlinear extent of a system aggravates. In order to have the advantages of the EnPF promoted into strongly nonlinear systems, this paper first analyzes the reasons why the EnPF deteriorates in strongly nonlinear systems and then proposes an EnPF based on Kullback-Leibler distance (KLD), which is named KLEnPF. KLEnPF is mainly inspired by the Gaussian sum approximation. For the aim of improving the approximation precision, the KLEnPF calculates the optimal bandwidth parameters of Gaussian kernels to minimize the KLD between two distribution functions. Finally, we apply the KLEnPF to the initial alignment of the strapdown inertial navigation system (SINS) in large misalignment angles, which is a strongly nonlinear system. The experimental results of it demonstrate the effectiveness of the KLEnPF.

The fifth-degree cubature Kalman filter (CKF) has been proved to be a kind of algorithm that has higher precision than the third-degree CKF and unscented Kalman filter (UKF). In order to further improve the performance of CKF, the seventh-degree CKF is proposed in this paper by expanding the spherical-radial rule, and a new kind of deterministic sampling method is derived based on the seventh-degree cubature rule. Through the comparison in target tracking simulation, the seventh-degree CKF methods are shown to be able to enhance filtering precision compared to the fifth-degree CKF, the third-degree CKF and the UKF filter.

Temperature drift is one of the main factors that affect the accuracy of fiber optic gyroscope (FOG), and its modeling and compensation are effective methods to eliminate and reduce the drift. The key factors that affect the temperature drift of FOG were analyzed. Meanwhile, the test experiment of FOG temperature drift was carried out. Then, the support vector machine which had better generalization ability than the neural network was used to regress and model the temperature drift of FOG, and the radial basis kernel function was adopted as the kernel function of support vector machine which had better data set adaptability. In order to improve the modeling accuracy of support vector machine, the artificial fish swarm algorithm was used to optimize the penalty factor C of support vector machine and the factor of kernel function. Finally, the proposed compensation method was verified by the actual temperature drift data of FOG, which showed that the remaining error of FOG compensated by the proposed method was reduced by 4-5 orders of magnitude than that compensated by the linear regression method.

In order to deal with the problems of non-linearity and external disturbance in the rotary position system, a dual closed loop controller is designed. A conventional PID is used in the outer loop, and an improved tracking differentiator(ITD)-fuzzy PID controller is in the inner loop. Combining the merits of linear TD and nonlinear TD, ITD not only has the advantages of no chattering phenomenon and excellent dynamic effect, but also has quickness and high precision, which can realize any signal tracking and differential. Combined with fuzzy PID algorithm, the ITD-Fuzzy PID controller has strong filtering performance and good robustness, which can meet the requirement for rotary position system.

To deal with the problems of non-linearity and uncertainty disturbance in the rotary position system of car navigation,the double closed-loop control system was designed.The parameter self-adaptive fuzzy-PI controller was used for inner loop control.By simulating human thought of fuzzy control,the dynamic performance of the system was improved.PI regulator was introduced to guarantee the steady precision and make up for the shortcomings of fuzzy control so that the system robustness and stability precision were improved.The parameters of PI controller could be regulated by control value when controller was switched.The positional control accuracy was guaranteed through PI controller which was designed in outer loop.Simulation results show that the designed control system has run smoothly with strong anti-jamming ability and high positioning accuracy by the double closed-loop control while there exists external disturbance.

Fiber optic gyroscope (FOG) is sensitive to temperature, and there is a certain temperature drift error in the working process of FOG especially in the startup phase. In this paper, to reduce the bias drift in the startup phase of FOG and shorten the startup time, a scheme based on radial basis function (RBF) neural networks was designed to compensate the drift in the startup phase of FOG. The model took the temperature of FOG and the temperature change rate as the inputs and used the bias drift of FOG as the output. In the room temperature, the RBF neural network was used to compensate the startup drift of FOG, and the experiment shows that the method can effectively reduce the temperature drift and shorten the startup time of FOG. This method is used in a certain type of FOG north finder and can greatly reduce the preparation time, and so improves the north-seeking accuracy.

As the core components of Fiber Optic Gyroscope (FOG) are sensitive to temperature, there is a certain temperature drift error in the working process of FOG. In particular, during the period from supplying power to achieving the nominal precision, the temperature drift of FOG is much higher. In this paper, for reducing the drift in the startup process of FOG and shortening the time of FOG startup, a scheme based on Radial Basis Function (RBF) neural networks is designed to compensate the drift in the startup process of FOG. The RBF neural network use the two inputs and single output scheme that use the temperature of FOG and the temperature change rate as the inputs and use the drift of FOG as the output. In the room temperature, the RBF neural network is used to compensate for the startup process of FOG, and the results show that the method can effectively reduce the drift and startup time of the FOG. This method is used in a certain type of FOG North Finder and can greatly reduce the North Finder preparation time and improve the north-seeking accuracy.