Figure 1 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
The solution of robot inverse kinematics has a direct impact on the control accuracy of the robot. Conventional inverse kinematics solution methods, such as numerical solution, algebraic solution, and geometric solution, have insufficient solution speed and solution accuracy, and the solution process is complicated. Due to the mapping ability of th...
Contexts in source publication
Context 1
... paper uses the FOA algorithm to optimize the threshold and weight of the BP neural network to improve the convergence speed and output accuracy of the neural network. The optimized neural network is used to solve the inverse kinematics of the robot shown in Figure 1, and compared with the ordinary BP neural network algorithm and the PSO optimized BP neural network algorithm. The robot structure shown in Figure 1 is a part of the service robot. ...
Context 2
... optimized neural network is used to solve the inverse kinematics of the robot shown in Figure 1, and compared with the ordinary BP neural network algorithm and the PSO optimized BP neural network algorithm. The robot structure shown in Figure 1 is a part of the service robot. It can be installed on a mobile chassis with laser radar [29] and equipped with robotic arms to form a complete service robot. ...
Similar publications
The signed distance field is a popular implicit shape representation in robotics, providing geometric information about objects and obstacles in a form that can easily be combined with control, optimization and learning techniques. Most often, SDFs are used to represent distances in task space, which corresponds to the familiar notion of distances...
Cuspidal robots can travel from one inverse kinematic solution to another without meeting a singularity. The name cuspidal was coined based on the existence of a cusp point in the workspace of 3R serial robots. The existence of a cusp point was proved to be a necessary and sufficient condition for orthogonal robots to be cuspidal, but it was not po...
The non-spherical 6R manipulators are widely used in many fields. However, the non-spherical structure often poses challenges in the inverse kinematics problem (IKP) for such robots. To address this challenge, transforming IKP into an optimization problem is a promising solution. Nevertheless, existing optimization methods often entail complex comp...
In the case of over-constrained mechanisms, the system can only be assembled or move under strict geometric conditions. A crucial step is to identify these conditions from the geometric model. Classically, the geometric model of a robot is computed from the Denavit–Hartenberg formalism. However, when imperfect mechanisms are studied, this formalism...
The paper presents the study of kinematic and dynamic characteristics of the robotic mobility platform (RMP) to determine inertia-force parameters depending on the nature of the implemented motion path. The method is implemented on the basis of parameterized digital simulation models with parallel kinematics, which allow determining acceleration an...
Citations
... However, these algorithms may be sensitive to parameter selection and prone to local optimum problems. Some researchers have used neural network to solve the inverse kinematics of manipulators [20][21][22][23][24][25] . Demby et al. ...
... 21 used neural networks and adaptive fuzzy inference recommendation systems to solve the inverse kinematics of manipulators with different degrees of freedom. Bai Y et al. 22 used the FOA optimized BP neural network algorithm to solve the robot kinematics can improve the control accuracy of the robot. Shiping et al. 25 proposed a method based on convolutional neural network models to solve the inverse kinematics of redundant manipulators and conducted trajectory tracking experiments to verify that convolutional neural networks can solve kinematic inverse solutions with high accuracy. ...
The advent of Industry 4.0 has significantly promoted the field of intelligent manufacturing, which is facilitated by the development of new technologies are emerging. Robot technology and robot intelligence methods have rapidly developed and been widely applied. Manipulators are widely used in industry, and their control is a crucial research topic. The inverse kinematics solution of manipulators is an important part of manipulator control, which calculates the joint angles required for the end effector to reach a desired position and posture. Traditional inverse kinematics solution algorithms often face the problem of insufficient generalization, and iterative methods have challenges such as large computation and long solution time. This paper proposes a reinforcement learning-based inverse kinematics solution algorithm, called the MAPPO-IK algorithm. The algorithm trains the manipulator agent using the MAPPO algorithm and calculates the difference between the end effector state of the manipulator and the target posture in real-time by designing a reward mechanism, while considering Gaussian distance and cosine distance. Through experimental comparative analysis, the feasibility, computational efficiency, and superiority of this reinforcement learning algorithm are verified. Compared with traditional inverse kinematics solution algorithms, this method has good generalization and supports real-time computation, and the obtained result is a unique solution. Reinforcement learning algorithms have better adaptability to complex environments and can handle different sudden situations in different environments. This algorithm also has the advantages of path planning, intelligent obstacle avoidance, and other advantages in dynamically processing complex environmental scenes.
... BP neural networks generally consist of three main layers [2]: ...
The Backpropagation Neural Network (BPNN) is a deep learning model inspired by the biological neural network. Introduced in the 1980s, the BPNN quickly became a focal point in neural network research due to its outstanding learning capability and adaptability. The network structure consists of input, hidden, and output layers, and it optimizes weights through the backpropagation algorithm, widely applied in image recognition, speech processing, natural language processing, and more. The mathematical model of neurons describes the relationship between input and output, and the training process involves adjusting weights and biases using optimization algorithms like gradient descent. In applications, BPNN excels in image recognition, speech processing, natural language processing, and financial forecasting. Researchers continuously experiment with optimization algorithms, including the Grey Wolf Algorithm, Genetic Algorithm, Particle Swarm Algorithm, Simulated Annealing Algorithm, as well as comprehensive strategies and improved gradient descent algorithms. In the future, with the ongoing development of deep learning, BPNN is poised to play a crucial role in tasks such as image recognition and speech processing.
... Takatani et al. [9] use NNs to map end-effector positions to joint configurations for serial redundant manipulators. Bai et al. [10] introduce an algorithm combining Fruit Fly Optimization Algorithm and Backpropagation NNs for solving robot inverse kinematics. Kramar et al. [11] improve inverse kinematics accuracy by designing an ANN with a correctional network for SAR-401 robot manipulators. ...
This research delves into the evaluation of Artificial Neural Networks (ANNs) in tackling the intricate domain of inverse kinematics problems within the context of robotic systems. Compared to direct kinematics, inverse kinematics presents significant challenges, especially when dealing with different arm lengths, which can impact the system's overall functionality. In this study, we conduct a comprehensive comparative analysis to assess the performance of ANNs across various robotic arm lengths. Following the forward kinematics of a simple 2-DoF robotic manipulator, three distinct datasets (fixed step size, random step size, and sinusoidal step size) are generated for training, testing, and validation phases. For every dataset, three optimizers, namely Levenberg Marquardt, Bayesian Regularization, and Scaled Conjugate Gradient, are considered with different hyperparameters to investigate the performance of the ANN models. Based on the validation of MSE values, ANN models with different hyperparameters are found to be a promising approach in estimating the inverse kinematic solutions for different arm lengths with data complexities. Such solutions can be further extended to higher DoF robotic solutions and save the computational burden of evaluating analytical solutions.
... They introduced a general framework for solving robot inverse kinematics, but it does not achieve satisfactory convergence accuracy. Khaleel (2018) and Bai et al. (2021) integrated Genetic Algorithm (GA) and Forest Optimization Algorithm (FOA) with the Backpropagation Neural Network (BPNN). When their approaches enhanced the solving precision, it also resulted in increased complexity of the network structure. ...
The non-spherical 6R manipulators are widely used in many fields. However, the non-spherical structure often poses challenges in the inverse kinematics problem (IKP) for such robots. To address this challenge, transforming IKP into an optimization problem is a promising solution. Nevertheless, existing optimization methods often entail complex computations and tend to overlook the geometric characteristics of the manipulators. In this study, we introduce a novel objective function based on a disconnect-reconnect method. Initially, based on the prior geometric knowledge of the non-spherical 6 degrees of freedom (DOF) manipulators, we employ a disconnect–reconnect strategy to decouple the kinematic equations. This process yields four nonlinear re-connection conditions equations. Subsequently, we utilize this equation to formulate a novel objective function. Then, we employ the adaptive covariance matrix evolution strategy (CMA-ES) alongside an analytical method to achieve precise solutions for the IKP. The proposed method was validated on the Comau NJ-220 manipulator. The simulation results demonstrate that the proposed effectively reduces computational complexity and enhances solution efficiency while maintaining accuracy in solving.
... Ref. [12] adopted the adaptive neuro-fuzzy inference system (ANFIS) to a 7-link manipulator to track trajectories in a narrow channel. In [13], the fruit fly optimization algorithm (FOA) is introduced as an optimization technique for the backpropagation (BP) neural network algorithm. It aims to address challenges related to low accuracy, slow convergence, and issues with local minima when solving inverse kinematics. ...
The Delta robot is a parallel robot that is over-actuated and has a highly nonlinear dynamic model, which poses a significant challenge to its control design. The inverse kinematics that maps the motor angles to the position of the end effector is highly nonlinear and extremely important for the control design of the Delta robot. It has been experimentally shown that geometry-based inverse kinematics is not accurate enough to capture the dynamics of the Delta robot due to manufacturing component errors, measurement errors, joint flexibility, backlash, friction, etc. To address this issue, we propose a neural network model to approximate the inverse kinematics of the Delta robot with stepper motors. The neural network model is trained with randomly sampled experimental data and implemented on the hardware in an open-loop control for trajectory tracking. Extensive experimental results show that the neural network model achieves excellent performance in terms of the trajectory tracking of the Delta robot under different operation conditions, and outperforms the geometry-based inverse kinematics model. A critical numerical observation indicates that neural networks trained with the specific trajectory data fall short of anticipated performance due to a lack of data. Conversely, neural networks trained on random experimental data capture the rich dynamics of the Delta robot and are quite robust to model uncertainties compared to geometry-based inverse kinematics.
... The neural network is such a data-driven modeling technique that it is flexible for modeling the inverse kinematics. Because of its flexibility and learning ability, the neural network can handle the problems of the inverse kinematics, starting from the simple robots [37,38] to the robots with complex structures [39][40][41]. The inverse kinematics solution resulted from the neural network is expressed in the neural network architecture that defines the mapping from the cartesian space to the joint space. ...
... In addition, the problem such as singularity and multiplicity does not exist in the neural network. That is why, in the design of robotic motion control, many researchers [37][38][39][40][41] preferred to use the neural network. It is important to note, the neural networkbased inverse kinematics structure is feedforward so it is classified as the open-loop control system. ...
... Meanwhile, the second approach is the neural network-based inverse kinematics. Because of its learning ability, most research used this approach [37][38][39][40][41]. Different from this approach, the approach we proposed in this research is the neural networkbased inverse kinematics combined with Jacobian. ...
Inverse kinematics plays a significant role in the robotic motion control. The flexible way to formulate the inverse kinematics can be achieved by using a neural network. The drawback of the neural network-based inverse kinematics is that it has no feedback mechanism to compensate for the remaining error. To improve its performance, further development should be conducted. In the design of the robotic motion control for the NAO robot arms, this paper proposed a new approach that combines the neural network-based inverse kinematics with the Jacobian. This combination yields a closed-loop control system. This control system utilizes the neural network as the feedforward controller and the Jacobian as the feedback controller. In particular, the neural network-based inverse kinematics has a function to estimate a set of required joint angles for the joint actuators, and the Jacobian function is to compensate for the remaining error of the neural network-based inverse kinematics in the estimation of joint angles. By using this proposed approach, we obtained more accurate joint angles for controlling the joint actuators. The comparison result showed that the averaged MSE for the particle swarm optimization (PSO) was 3.47 x 10-3 rad, 1.19 x 10-3 rad for the neural network, and 3.72 x 10-5 rad for the proposed approach. The performance comparison result indicated that our proposed approach has a lower averaged MSE than the other ones. Accordingly, the result of this research confirmed that our proposed approach can provide more accurate joint angles for controlling the joint actuators such that the robot’s end-effector can be driven along the desired path in the cartesian space.
... Применение методов машинного обучения для аппроксимации функций регрессии при решении задач инверсной кинематики осуществляется достаточно давно, но ранние работы [9] в этой области отмечали низкую точность и производительность данных подходов. Развитие программного и аппаратного обеспечения в сфере машинного обучения позволило обеспечить прогресс по данному направлению: сочетания различных нейронных сетей используются для реализации инверсной кинематики роботов [10], прогнозирования положения тела с учетом окружающей среды на основе виртуального скелета [11] или набора датчиков [12]. В [13] рассмотрено применение нейронных сетей для восстановления ключевых точек силуэта человека при удалении от 10 до 30 % данных. ...
... As the most typical deep learning method, the BP neural network [27][28][29] is widely used by many scholars to predict temperature, current, passenger flow, and so on. Therefore, this paper uses the BP neural network to construct a normal leakage current prediction model, taking the feature vector at the current moment as the input, and the prediction as the output. ...
The fluctuation of normal leakage current has a great influence on the fixed-threshold leakage protector. To address this issue, this paper proposes an adaptive leakage protection method based on the sparrow search algorithm (SSA)-backpropagation (BP) neural network. Based on the analysis of the normal leakage current generation mechanism, this method uses the SSA optimized BP neural network to construct a prediction model of normal leakage current. By dividing the normal leakage range into several intervals and setting the corresponding action threshold, the action threshold of the interval is automatically selected in advance, based on the predicted value of the model, so as to realize the adaptive protection of the leakage current faults. Experiments have proved that the leakage protector can identify the leakage fault more sensitively and increase the ratio of the protector put into operation by predicting the development of normal leakage current and adjusting the protection action threshold in advance.
... Domestic forestry operations today still rely on human labor to get the job done, especially drilling deep afforestation, which is done by hand. This method is not only time-consuming, but also labor-intensive and inefficient, greatly reducing the speed of afforestation [1][2][3]. Forest tree planting is a complex undertaking, with labor costs generally account for more than 60 % to 80 % of the actual cost. Labor force is also facing a phase of transformation from manual labor to skilled labor, with the cost of manual labor increasing year on year [4][5][6]. ...
Tree Planting Machine (TPM) is subject to a Tree-Planting Robot (TPR) with desired tracking trajectory planning. In this topic, taking the TPR proposed as the analysis object, the positive and inverse solutions of the kinematics are analyzed to explore the optimal trajectory planning. An improved position/posture algorithm, based on the analytical solution of the inverse kinematics of the TPR, is proposed. The trajectory planning strategy for TPR in Cartesian coordinate system and Joint coordinate system is discussed, which is used for parabolic transition linear programming optimization, and the simulation model of TPR trajectory planning is constructed by MATLAB module. Numerical simulation results indicate that the deviation of the TPR trajectory from the expected value is significantly reduced. The proposed improved position/posture algorithm is verified by kinematic analysis, and the TPR followability and trajectory planning accuracy are greatly improved. Toward this goal, a variable trajectory planning can be effectively, and stability adjusted by pre-designed TPM system in the field of ecological tree planting.
... Early works [19], rather pessimistically, assessed the possibility of solving this problem due to the insufficient performance and accuracy of neural networks at that time. However, in the future, significant progress can be observed in this direction; combinations of various neural networks are used to implement the inverse kinematics of robots [20], predicting body positions taking into account the environment based on a virtual skeleton [21] or a set of sensors [22]. The study in [23] considers the use of neural networks to restore the key points of a person's silhouette when deleting from 10 to 40% of the data. ...
In virtual reality (VR) systems, a problem is the accurate reproduction of the user’s body in a virtual environment using inverse kinematics because existing motion capture systems have a number of drawbacks, and minimizing the number of key tracking points (KTPs) leads to a large error. To solve this problem, it is proposed to use the concept of a digital shadow and machine learning technologies to optimize the number of KTPs. A technique for movement process data collecting from a virtual avatar is implemented, modeling of nonlinear dynamic processes of human movement based on a digital shadow is carried out, the problem of optimizing the number of KTP is formulated, and an overview of the applied machine learning algorithms and metrics for their evaluation is given. An experiment on a dataset formed from virtual avatar movements shows the following results: three KTPs do not provide sufficient reconstruction accuracy, the choice of five or seven KTPs is optimal; among the algorithms, the most efficient in descending order are AdaBoostRegressor, LinearRegression, and SGDRegressor. During the reconstruction using AdaBoostRegressor, the maximum deviation is not more than 0.25 m, and the average is not more than 0.10 m.
