Develop and Control of a Rigid-Flexible Manipulator

The objective of this research was to introduce flexibility within existing rigid-link robotic manipulators to benefit from the advantages of having elastic deformations within the structure of the arm. This integration enables the robot to alleviate the impact and potential damage to both the arm components and any encountered objects during interaction. However, this flexibility within the structure introduces challenges in maintaining precise control due to induced vibrations and resultant positioning inaccuracies. To counter these challenges, the study employed the Assumed Mode Method (AMM) in conjunction with Lagrangian formulation to develop an accurate and computationally efficient dynamic model of the flexible link. Having a precise dynamic model, the study focused on leveraging Model Predictive Control (MPC) strategies for effective trajectory tracking and vibration damping in a rigid-flexible link manipulator tasked with catching floating objects. The optimization of MPC parameters was conducted using Genetic Algorithms (GA) across various operational conditions of the manipulator. Subsequently, a Gain-Scheduled MPC approach was adopted to dynamically adjust the optimal MPC settings based on the robot's current state, ensuring enhanced performance. Additionally, For the estimation of the system state, as a crucial step for the control, the study proposed an asynchronous Kalman filter-based approach. Experimental findings demonstrated the superior performance of this proposed method compared to the traditional Kalman and asynchronous Kalman filters. The effectiveness of the proposed MPC strategy was validated through both simulation and practical experiments. This ability to precisely manage rigid-flexible manipulators and reduce vibrations triggered by dynamic interactions opens new avenues for their application in fields such as space debris mitigation, collaborative human-robot interactions, and agricultural tasks, showcasing the potential for broader adoption of rigid-flexible robotic arms in complex and dynamic environments.