Collaboration of multiple SCARA robots with guaranteed safety using recurrent neural networks

SCARA robot is one of the most popularly used robots in industry. The obstacle avoidance feature multiple collaboration essential and prominent, which can be to support accomplish not only more sophisticated tasks but also efficient than individual robot. This paper mainly focuses on studying problem simultaneous multi-robot coordination avoidance. A cooperative kinematic control manipulators, collision taken into account primary task as an inequality constraint trajectory planning considered secondary objective ensure priority safety, described a quadratic programming (QP) problem. Then, recurrent neural network (RNN) based dynamic controller designed solve formulated QP recursively. convergence proved through Lyapunov analysis. With three planar robots, effectiveness proposed validated numerical simulations. As observed results, when minimal distance between less setting safety distance, strategy reacts impel avoid collision, achieves for avoidance; otherwise, performs desired tracking task.