A fixed-time GNN with enhanced performance via simplified activation for online equality-constrained quadratic programming and its robotic application

Key Points

• Performance improvements are linked to simplified activation functions in a fixed-time GNN.
• The enhanced GNN outperforms previous models, achieving faster convergence in quadratic programming tasks.
• Assessment using online programming techniques illustrates effective application in robotics and automation.
• This work highlights the need for efficient algorithms, particularly in real-time robotic settings.