daVinci Research Kit Patient Side Manipulator Dynamic Model Using Augmented Lagrangian Particle Swarm Optimization

In surgical robotics, accurate characterization of the dynamic model is crucial. It serves as a foundation for developing robust control algorithms that effectively handle the complex dynamics of the robot and its interactions with the environment. Additionally, an accurate dynamic model aids in estimating external forces and disturbances, enhancing the safety and stability of the control. Among surgical robots, the da Vinci Research Kit (dVRK) is one of the most used, and it has been a crucial instrument in removing a barrier to entry for new research groups in surgical robotics by facilitating the development of improved control algorithms. This paper presents a method for dynamic model identification of the dVRK *psm robot that employs a novel friction model definition. The model formulation has been modified by including the Stribeck effect at low velocities, and the friction has been estimated using the superposition method. The dynamic parameters are identified utilizing a restricted optimization method with physical consistency requirements in an Augmented Lagrangian Particle Swarm Algorithm (ALPSO) methodology. The identified model is thoroughly evaluated, and the results are compared with existing literature methods. Also, a model-based sensorless force estimation method was used to test the dynamic model.