9 October 2017 in IEEE Transactions on Industrial Informatics
Fingerprinting based indoor positioning system (IPS) is gaining more research interest under the umbrella of location-based services. However, existing works have certain limitations in addressing issues such as noisy measurements, high computational complexity, poor generalization ability. In this work, a random vector functional link network based approach is introduced to address these issues. In the proposed system, a subset of informative features from many randomized noisy features are selected to both reduce the computational complexity and boost the generalization ability. Moreover, the feature selector and predictor are jointly learned iteratively in a single framework based on Augmented Lagrangian method. The proposed system is appealing as it can be naturally fit into parallel or distributed computing environment. Extensive realworld indoor localization experiments are conducted on users with smartphone devices and results demonstrate the superiority of the proposed method over the existing approaches.
21 November 2024 in IEEE Transactions on Vehicular Technology
Both DQN and M-DQN is commonly used for mobile robot path planning, with the problem of relatively low efficiencies of the action selection policy and the reward function when coping with complex environments. To solve the problem, an improved $\epsilon$ -UCB action selection method based on neural network is proposed to fit the uncertainty of state-action pairs, which effectively improves the exploration efficiency. Meanwhile, aiming at the effectiveness of the reward function, a multi-goal fusion reward function is proposed to improve the guidance ability. Further, the proposed $\epsilon$ -UCB action selection policy and the multi-goal fusion reward function are introduced into the M-DQN framework to complete the path planning task of the mobile robot. Finally, some comparative experiments are carried out to verify the effectiveness of the proposed method. The experimental results show that the proposed method exhibits faster convergence speed and higher learning efficiency for the path planning of mobile robots in complex environments than other existing methods.
1 March 2018 in International Journal of Advanced Robotic Systems
The Paden–kahan subproblem is a simple and flexible method to solve the closed-form inverse resolution but limited by the geometrical structure of robots, which is very difficult to be kept because of processing and installation. Therefore, a closed-form solution on arbitrary configurations is an important issue in the field of robotic inverse kinematics. A novel second subproblem is firstly proposed in this study based on the product-of-exponentials model adapting to the two arbitrary axes without geometric constraints (parallel, vertical, disjoint, and so on). Furthermore, the algebraic methods involving the basic properties of the screw theory and Rodrigues’ rotation formula are employed for the solution, which makes the constraint equations of the second subproblem solvable for arbitrary configurations. This methodology can be applied to the inverse solutions of 5-degree-of-freedom robots that satisfies the Pieper criterion and can express the inverse solutions via two common formulas. Finally, the simulation and the real-world experiment demonstrated the accuracy of the method and the validity, respectively.
19 September 2023 in Expert Systems with Applications
9 October 2025 in IEEE Transactions on Instrumentation and Measurement
Applications such as electric vehicles require high-precision induction motor (IM) controllers. A large amount of measurement noise is introduced during position and speed detection, which seriously affects the tracking and anti-disturbance performance of the controller. To solve this problem, a novel enhanced ADRC framework is proposed in this paper. On the one hand, a novel third-order observer is developed for the enhanced ADRC by embedding an LPF into a phase-locked loop observer (PLLO) so that the motor’s actual speed, filtered speed, and total disturbance can be estimated simultaneously. Meanwhile, the observed actual speed is utilized as the feedback signal to enhance noise suppression while maintaining the original structure’s tracking performance and disturbance rejection ability. On the other hand, to address the inherent phase-lag problem in LPFs, a tracking differentiator with phase compensation (TDPC) is designed, which performs filtering while compensating for the phase lag induced by low-pass filtering through its derivative lead characteristic. Experimental validation on an IM test platform demonstrates significant performance enhancements: the total harmonic distortion represents the noise suppression ability is reduced by more than 68%, and the root mean square error (RMSE) representing the speed fluctuation is reduced by more than 63% at speeds below 30 rpm with rated-load. The phase lag is reduced by at least 80% at the frequency is 5Hz with rated-load. And the transient response improved more than 27%, with the disturbance rejection improved more than 18%.