This paper studies cooperative control for dual-arm robots with state constraints and torque optimization. Firstly, the centralized comprehensive dynamic model of the dual-arm robots with the grasped object is established in the Cartesian coordinate system, and then the motion of the system is analyzed using the property that the sum of internal forces is zero. Secondly, to deal with the state constraints, a constraint transformation function is constructed involving the system state and the

more »

... straint boundary, with which the state constraints are satisfied as long as the new transformed system states are bounded. Then, a neuro-adaptive cooperative control scheme is designed with the non-unique pseudo-inverse grasping matrix. Finally, an energy function is built and the control is determined uniquely using the torque optimality principle. The significant advantages of this paper are described as follows:(i) Different from the traditional barrier-Lyapunov function for dealing with state constraints, where the satisfaction of feasibility conditions is required, here this condition is avoided by the introduction of constraint transformation function and system transformation technology. (ii) The torque optimization eliminates the requirement for calculating the inverse of the grasping matrix with insufficient rank and the control input is determined uniquely.