Fig 2 - available via license: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
Content may be subject to copyright.
Joint configuration of the excavator. The bucket tip position is observed in the frame {B} attached to the base of the excavator.
Source publication
This work proposes a novel precision motion control framework of robotized industrial hydraulic excavators via data-driven model inversion. Rather than employing a single neural network to approximate the whole excavator dynamics, including input delays and dead-zones, we construct a physics-inspired data-driven model with a modular structure. The...
Contexts in source publication
Context 1
... strategy. The excavator is customized using sensors to measure states and environmental impacts, as shown in Fig. 1. Inertial measurement unit (IMU) sensors are attached to the boom, arm, and bucket links to estimate the joint configuration. Swing angle is also measurable, but we only consider the motion within the sagittal plane as visualized in Fig. 2 because the swing action is not involved in the excavation. Apart from the joint configuration, hydraulic pressure sensors are located in pumps and cylinders to consider the hydraulic behavior. The pumps and the cylinders are connected through the MCV as detailed in [3], which consists of spool valves that distribute the pump flow rate ...
Context 2
... bucket tip position p t := (p x,t , p z,t ) ∈ R 2 is calculated to evaluate the control performance, where p x,t , p z,t ∈ R are the horizontal and vertical tip positions as shown in Fig. 2. The path following error is denoted by e path p,t := min t0≤τ ≤t f ∥p t − p ref τ ∥ ∈ R, which indicates the error of the excavated ground geometry. The trajectory error, or the bucket tip position error, is written as e traj p,t := ∥p t − p ref t ∥ ∈ R. Here, we calculate the RMSE from one second after the initial time. This is to ...
Citations
... Among the previously investigated approaches, learningbased methods [16], [22], [3] have been proven suitable to learn and leverage complex machine dynamics while being robust to disturbances under different deployment scenarios. We decided to pursue the same direction to address the control of hydraulic machines with redundant and underactuated kinematics. ...
Automation of hydraulic material handling machinery is currently limited to semi-static pick-and-place cycles. Dynamic throwing motions which utilize the passive joints, can greatly improve time efficiency as well as increase the dumping workspace. In this work, we use Reinforcement Learning (RL) to design dynamic controllers for material handlers with underactuated arms as commonly used in logistics. The controllers are tested both in simulation and in real-world experiments on a 12-ton test platform. The method is able to exploit the passive joints of the gripper to perform dynamic throwing motions. With the proposed controllers, the machine is able to throw individual objects to targets outside the static reachability zone with good accuracy for its practical applications. The work demonstrates the possibility of using RL to perform highly dynamic tasks with heavy machinery, suggesting a potential for improving the efficiency and precision of autonomous material handling tasks.
... KURINOV et al. [15] accomplished the automatic loading of excavators through reinforcement learning within a simulation environment, and the study conducted by Samtani et al. [16] also employed reinforcement learning techniques to develop Dueling Double Deep-Q Networks and extensively investigated the fracturing actions of excavators. Through data-driven approaches, task planning research was conducted by ZHAO et al. [9], and inverse motion control of excavator models was conducted by Lee et al. [17]. EGLI et al. [18] trained models using reinforcement learning techniques to accurately identify soil hardness, while YU et al. [19] integrated physical and data-driven models for real-time soil resistance prediction. ...
Trajectory planning plays a crucial role in achieving unmanned excavator operations. The quality of trajectory planning results heavily relies on the level of rules extracted from objects such as scenes and optimization objectives, using traditional theoretical methods. To address this issue, this study focuses on professional operators and employs machine learning methods for job trajectory planning, thereby obtaining planned trajectories which exhibit excellent characteristics similar to those of professional operators. Under typical working conditions, data collection and analysis are conducted on the job trajectories of professional operators, with key points being extracted. Machine learning is then utilized to train models under different parameters in order to obtain the optimal model. To ensure sufficient samples for machine learning training, the bootstrap method is employed to adequately expand the sample size. Compared with the traditional spline curve method, the trajectories generated by machine learning models reduce the maximum speeds of excavator boom arm, dipper stick, bucket, and swing joint by 8.64 deg/s, 10.24 deg/s, 18.33 deg/s, and 1.6 deg/s, respectively; moreover, the error does not exceed 2.99 deg when compared with curves drawn by professional operators; and, finally, the trajectories generated by this model are continuously differentiable without position or velocity discontinuities, and their overall performance surpasses that of those generated by the traditional spline curve method. This paper proposes a trajectory generation method that combines excellent operators with machine learning and establishes a machine learning-based trajectory-planning model that eliminates the need for manually establishing complex rules. It is applicable to motion path planning in various working conditions of unmanned excavators.
... Dao explored linear operation [15] control from a novel perspective by categorizing tracking errors into contour error, tangential error, and azimuthal error, addressing each type of error individually using extended state observer techniques [16]. On the other hand, Lee et al. developed an excavator motion model based on data-driven approaches and accomplished precise trajectory tracking solely through proportional control strategies [17]. ...
... Recent studies highlight the use of data-based methods for tracking control tasks (Halbach, Kämäräinen, & Ghabcheloo, 2019;Lee et al., 2022;Taheri, Gustafsson, Rösth, Ghabcheloo, & Pajarinen, 2022). Taheri et al. (2022) employ a dual-Gaussian process model architecture for modeling the nonlinear actuator dynamics, accounting for uncertainties. ...
... Another approach is combining gain-tuning methods with linear controllers [5], [6], [7], [8] in which the control gains are optimized in metaheuristic algorithms. Some researchers [9], [10], [11], [12], [13] employ neural networks to approximate the actuator characteristics, but the actuator force saturation, which frequently happens and discontinuously alters the response characteristics, has not been taken into account. ...
... Another problematic factor in the control of hydraulic systems is the deadtime, which is typically 0.1 s to 0.6 s or more [2], [4], [14] in commercial hydraulic excavators. There have been some controllers based on models of the deadtime in hydraulic systems [2], [4], [11], [12], but they require the identification of the deadtime, and the control performance depends on the precision of the identification. ...
... Gain-tuning methods [5], [6], [7], [8] based on metaheuristic approaches, such as particle swarm optimization and genetic algorithms, have been proposed for linear controllers. To handle the strong nonlinearity of the actuator characteristics, some controllers [9], [10], [11], [12], [13] employ neural networks for the compensation of the nonlinearity. It has been reported that smooth trajectory tracking can be achieved with these controllers, but no controller analytically takes into account the effects of all valves, such as the regenerative pipelines, the force saturation and the square-root law [21] of the pressureflowrate relation. ...
This paper proposes a position controller for commercial hydraulic excavators. It is constructed by combining a proportional-derivative (PD) controller and a sliding-mode controller as a differential algebraic inclusion and also is integrated with a recently-proposed hydraulic actuator model. The use of the PD control is intended to make the controller insensitive to the deadtime in the hydraulic system, which is typically 0.1 s to 0.6 s in commercial excavators. The use of the sliding-mode controller combined with the actuator model is for handling the saturation of the actuator force, which may happen when the target position is not close enough to the current position and when the relief valves open. Moreover, this paper extends the controller to deal with the effect of the regenerative pipelines, which are embedded in commercial excavators to realize efficient operations but act as a source of disturbance on the controller. This paper also shows an analysis that can be used for tuning the controller parameters. The proposed controller was validated with simulations and experiments using a 13-ton class excavator, in which some set-point control tasks and trajectory-tracking tasks were performed.
Note to Practitioners
—This paper proposes a position controller for hydraulic excavators. The controller was validated with the boom and arm actuators of a 13-ton class commercial excavator, with trajectory-tracking and set-point control tasks. Most of the controller parameters can be set referring to available specifications of the hydraulic circuit, such as the set pressures of the relief valves and the cross-sectional areas of the chambers. There are three parameters (the proportional gain, the derivative gain, and the time constant of the convergence) that should be carefully tuned, but their physical interpretations are relatively straightforward, and they can also be tuned along our guideline using pole locations of a particular transfer function. It has been shown that the proposed controller properly works despite the existence of the deadtime in the hydraulic systems, which are typically 0.1 s to 0.6 s. An extended version of the proposed controller handles the effects of the regenerative pipeline, which exists in some hydraulic systems, e.g., the arm actuator, for efficient operation and is not accounted for in the original version of the controller. The accuracy of the proposed controller will be further improved by combining it with better means to estimate external forces by using additional sensors, such as pressure sensors installed to actuator chambers.
... 21 developed an improved genetic algorithm (IGA) to optimize the parameters of the excavator PID controller and obtained an IGA-tuned PID controller. Lee et al. 22 employed data-driven model inversion to execute the offline training model in a supervised way, and used model inversion control to make nonlinear compensation for DX380LC, which substantially improved the tracking accuracy of hydraulic actuators. ...
The motion control accuracy of the excavator manipulator is the primary guarantee for autonomous excavators to complete work tasks. This paper studies the motion control strategy of the manipulator of a certain autonomous excavator. By establishing a dynamic model of a 3-degree of freedom of excavator manipulator, the gravity term, inertia term, and centripetal force term in the model are equivalent to external disturbances for online compensation, which improves the motion control accuracy of the excavator manipulator. Based on the dynamic model, a control method of the bucket tooth tip trajectory of an autonomous excavator is proposed. The simulation and test results show that the maximum tracking errors of the joint angles of the boom, the bucket, and the bucket are reduced by 43.14, 26.56, and 51.05% respectively, and the average errors of the whole trajectory are reduced by 56.41, 61.33, and 64.26% respectively. The simulation and test results show that the proposed motion control strategy improves the operation accuracy of the excavator, and can effectively improve the operation accuracy and efficiency of the autonomous excavator.
... Parts of excavator represented as links and joints[16]. ...
The 4th industrial revolution has led to automation in various area, including construction. One of the most important parts of construction process is excavation. However, automating an excavation process still proves to be a challenging task. This research aims to automate an excavator in order to improve human safety and work efficiency using a class of machine learning, called imitation learning. The proposed method is able to learn the movement of the excavator using a dataset gathered from a simulated expert operator. The results show that Generative Adversarial Imitation Learning (GAIL) algorithm is able to successfully control excavator's joint angle in performing pick and place operation. The performance is validated from several control parameters such as rise time, settling time, and target angle error, with the largest error is 9.45 degrees. From this research, it can be concluded that excavator joint control system using imitation learning especially GAIL produces promising results.
... Wind ve ark. [5] yörünge üretimi için model tahminine dayalı kontrolcü geliştirmiştir. Konum ve hidrolik davranış için kinematik ve dinamik modelleri analiz etmişlerdir. ...
... Kontrolör, ileri besleme ve geri besleme parçaları içerecek şekilde tasarlanmıştır. Geri besleme kısmı, hız ve konum için bir P denetleyicisidir (5). Yörünge oluşturmada, kartezyen uzayda noktalar tanımlanır ve yörünge, mekanizma limitlerinde (5) ...
Abstract
Operator assist systems, automatic mechanism control and machine guidance systems have widely area of usage to increase safety for operators and environment and efficiency in construction work sites. In this study, two boom pieces excavator mechanism is analyzed in terms of forward kinematic analysis method with geometrical theorems. Also, this excavator mechanism is modeled and simulated in MATLAB script. Position/Inclination sensors are mounted on vehicle and mechanism to measure angle of bodies. Kinematic calculation is integrated into position sensors and bucket edge position is calculated in vehicle controller. Moreover, height value is controlled with constraint values determined by operator according to vehicle body. These operations are verified in excavator vehicle.
... Robotic excavators are a type of multi-joint device driven by hydraulic systems, which are widely used in earth-moving fields such as energy development, road construction, and infrastructure construction. However, the operator must have skilled techniques to control the multiple joints of the excavator to improve operational efficiency [1]. In addition, in some harsh construction environments, the operator's safety may be threatened. ...
Given the highly nonlinear and strongly constrained nature of the electro-hydraulic system, we proposed an observer-based approximate nonlinear model predictive controller (ANMPC) for the trajectory tracking control of robotic excavators. A nonlinear non-affine state space equation with identified parameters is employed to describe the dynamics of the electro-hydraulic system. Then, to mitigate the plant-model mismatch caused by the first-order linearization, an approximate affine nonlinear state space model is utilized to represent the explicit relationship between the output and input and an ANMPC is designed based on the approximate nonlinear model. Meanwhile, the Extended Kalman Filter was introduced for state observation to deal with the unmeasurable velocity information and heavy measurement noises. Comparative experiments are conducted on a 1.7-ton hydraulic robotic excavator, where ANMPC and linear model predictive control are used to track a typical excavation trajectory. The experimental results provide evidence of convincing trajectory tracking performance.
... Applying data-driven methods in nonlinear construction machines has collected considerable interest among researchers. Notably, researchers have employed neural networks to capture the underlying dynamics of excavators using a supervised learning approach [12], and have trained control policies for grading tasks through a data-driven actuator model in simulation using a reinforcement learning framework [13]. Despite the advancements established in these studies, human operators do not use the knowledge generated about optimizing excavator control. ...
The utilization of teleoperation is a crucial aspect of the construction industry, as it enables operators to control machines safely from a distance. However, remote operation of these machines at a joint level using individual joysticks necessitates extensive training for operators to achieve proficiency due to their multiple degrees of freedom. Additionally, verifying the machine resulting motion is only possible after execution, making optimal control challenging. In addressing this issue, this study proposes a reinforcement learning-based approach to optimize task performance. The control policy acquired through learning is used to provide instructions on efficiently controlling and coordinating multiple joints. To evaluate the effectiveness of the proposed framework, a user study is conducted with a Brokk 170 construction machine by assessing its performance in a typical construction task involving inserting a chisel into a borehole. The effectiveness of the proposed framework is evaluated by comparing the performance of participants in the presence and absence of virtual fixtures. This study results demonstrate the proposed framework potential in enhancing the teleoperation process in the construction industry.
