Ballbot is a unique kind of robot with great mobility and possesses the ability to go in all directions. Obviously, controlling such a robotic device must be tricky. Indeed, ballbot systems pose unique challenges, particularly in the form of the difficulty of maintaining balance and stability in dynamic and uncertain environments. Traditional proportional integral derivative (PID) controllers struggle with these challenges, and other advanced methods, like sliding mode control, introduce issues like chattering. Therefore, there is a need to develop a controller that combines the simplicity and adaptability of PID with the learning capabilities of the now-popular neural networks, providing a robust solution to real-world robotic mobility problems.
Recently, in a novel study, a team of researchers, led by Dr. Van-Truong Nguyen of Hanoi University of Industry, Vietnam, has come up with a new robust and adaptive solution. Their innovative work was made available online on December 4, 2024 and published in Volume 61 of Engineering Science and Technology, an International Journal on January 1, 2025.
The team included Associate Professor Phan Xuan Tan from Shibaura Institute of Technology, Japan, Mr. Quoc-Cuong Nguyen and Mr. Dai-Nhan Duong from Hanoi University of Industry, Vietnam, Associate Professor Mien Van from Queen's University Belfast, United Kingdom, Professor Shun-Feng Su from National Taiwan University of Science and Technology, Taiwan, and Associate Professor Harish Garg from Thapar Institute of Engineering and Technology (Deemed University), India.
Their research introduces a novel adaptive nonlinear PID (NPID) controller integrated with a radial basis function neural network (RBFNN) for ballbots, offering lightweight computation, superior stability, chattering reduction, and robustness against external disturbances. The initial settings of the proposed controller are selected through balancing composite motion optimization, and the adaptive control law is improved continuously during operation to handle the real-time estimation of the external force.
In this study, the team underlines the stability of the system through the application of the Lyapunov theory. Through both simulations and real-world experiments, they demonstrate the efficacy of the NPID-RBFNN controller, which outperforms traditional PID and NPID controllers. Additionally, the proposed controller adapts to the surface variations through self-learning and self-adjusting capabilities.
Dr. Nguyen envisions various applications for their innovative technology, including assistive robotics, service robotics, and autonomous delivery. Expanding on each of these domains, he remarks: "Ballbots with this advanced controller can be used as assistive robots for tasks requiring high mobility and precision. For instance, they can assist individuals with mobility challenges in navigating complex environments. In addition, they can be used as service robots in dynamic settings such as restaurants, hospitals, or airports, offering smooth navigation." Further, he adds, "The robust self-balancing capabilities can be applied to delivery robots that need to operate efficiently despite unpredictable forces like wind or uneven terrain."
Notably, the study addresses significant challenges in controlling nonlinear and dynamic settings, focusing on reliability for broader adoption in industries requiring autonomous mobility solutions. By minimizing unnecessary movements and chattering, the proposed controller can optimize energy consumption, promoting sustainable robotics. This, in turn, enhances the reliability of ballbots, making them safer and viable for use in public and private spaces.
"Overall, industries such as logistics, healthcare, and retail could benefit from robots equipped with our technology, improving efficiency and service quality while reducing human workload," concludes Dr. Nguyen. Let us hope for future advancements in this research, enabling efficient use of robots in the real world.
