Common Challenges in Control of Industrial Manipulators: A Review

The study of the nonlinear control of manipulators is increasing day by day for scientific investigation and industrial purposes [1]. Robotics is a multidisciplinary system branch that requires multidomain expertise. The control of a robot manipulator is the most challenging task as even a single link manipulator is nonlinear in nature and has highly complex dynamic behavior [2]. There are many nonlinear control approaches like SMC, robust control, adaptive control, and intelligent control [3-5]. With the development of computational power, the use of AI-enabled robotics is increasing day by day.

Mostly Robots work on the architecture of SPA (Sense, Plan, Act). Sense means to sense their surroundings, such as the presence of barriers, Plan means the robot needs a strategy to do some task in that sensed environment, Act means once a strategy for completing the objective has been devised, the robot must carry out the strategy’s instructions.

The design of any robotic system needs the knowledge of kinematic and dynamic analysis, mathematical modelling, control system, image processing, simultaneous localization and mapping (SLAM), sensor interface and communication system, embedded system, hardware interface, real-time operating system, optimization of different parameters etc. The Autonomous industrial robotics system workflow is as shown in Figure 1.

Robotics systems required multidomain expertise, end-toend workflow, algorithm complexity, technical depth and system stability. Safety standards given by ANSI R15.06-1999, offers instructions for the construction and integration of industrial robots and robot systems with a focus on their secure use, the significance of risk assessment, and ensuring worker safety.

The control of robots can be done using various techniques such as model-based control, rule-based systems, feedback control, or hybrid approaches that combine multiple methods. In the literature, there is n -number of controllers and nonlinear control techniques available but implementing them in real-time is still far away. Robotic manipulator plays an important role in the industry in increasing quality and productivity if the overall process as a manipulator has greater flexibility than normal machines. In the design of a robot manipulator, different control strategies are designed according to different task requirements such that desired result can be obtained. The main objectives of control are stability, tracking, and performance Index. The stability of a robot can be affected by various factors, such as the robot’s configuration, its weight distribution, the forces acting on it, and the type of task it is performing. The stability analysis of robots is often based on Lyapunov’s stability theory, which provides a systematic method for determining the stability of a system. To get the desired performance of the system several control schemes including PID, FOPID, adaptive, optimal, backstepping, Intelligent control, Robust, and SMC have been implemented. SMC guarantees trajectory tracking with a rapid response for the robotic manipulators. SMC is well recognized as a robust control for dealing with uncertainty in nonlinear systems. Although SMC gives a robust response, it yields a chattering effect that can cause instability in the controlled system [6]. The high value of controller gains excites the high-frequency dynamics that create chattering. Boundary layer methods suggest the use of sigmoid function, saturation function, and use of fuzzy logic. These approaches reduce chattering since they function as low-pass filters. Unknown control direction can manifest intensified problems for the manipulators. It is referred to as the wrong connection between actuators and drivers, researchers are finding ways to overcome this.

In the case of AI–enabled robot machine learning control and Data-driven control approaches are most suitable when we have large data sets to predict future behavior Metaheuristic approaches like Whale optimization algorithms (WAO), Multiverse optimization (MVO), moth flame optimization (MFO), and grey wolf optimization (GWO) to control the trajectory of the manipulator.

The mentioned control methods given in the paper are useful for achieving the desired task. control is used to predict the boundaries of uncertainty and external disturbances. Furthermore, the suggested control removed the chattering phenomena while maintaining robustness and accuracy.

None.

No conflict of interest.

1. Rafael Kelly, Victor Santibáñez Davila, Antonio Loría (2006) “Control of Robot Manipulators in Joint Space, Sringer.
2. R Kelly, V Santibanez, A Loria (2005) Control of Robot Manipulators in Joint Space, Berlin, Germany: Springer-Verlag.
3. H. Nguyen, E. Kreuzer (2007) A robust adaptive sliding mode controller for remotely operated vehicles. Technische Mechanik 28(3–4): 185–193.
4. F Li, HL Xie (2010) Sliding mode variable structure control for visual servoing systems. International Journal of Automation and Computing 7(3): 317-323.
5. Raoufi M, Habibi H, Yazdani A, Wang H (2022) Robust Prescribed Trajectory Tracking Control of a Robot Manipulator Using Adaptive Finite-Time Sliding Mode and Extreme Learning Machine Method. Robotics 11(5)
6. Neila MBR, Tarak D (2011) Adaptive terminal sliding mode control for rigid robotic manipulators. Int. J. Autom Comput 8: 215-220.
7. Steven L Brunton, J Nathan Kutz (2019) “Data-driven science and engineering: Machine learning, dynamical systems, and control” Cambridge University Press.