Bin Zhao, Lianjun Chang, and Ruohuai Sun
Collision detection, particle swarm optimisation, desktop robot,transformer, LSTM
Given the robot’s rigorous safety and interactivity requirements, this paper introduces free teaching and collision detection methods for desktop robots. First, kinematics and trajectory planning are prerequisites for free teaching and collision detection. Second, the free-teaching parameter identification method combining the LuGre friction model and the particle swarm optimisation (PSO) algorithm is proposed to achieve high-precision dynamic modelling of robots. Third, the BiLSTM–Transformer hybrid deep learning model, used for robot collision detection, can simultaneously capture both temporal and spatial dependencies, thereby improving prediction accuracy and robustness. Finally, the proposed methods effectively resolve free teaching and collision detection, showing the remarkable accuracy and robustness.
Important Links:
Go Back
