A kinematic joint fault tolerant control based on relative jacobian method for dual arm manipulation systems

The present paper presents a fault tolerant control for dual arm manipulation systems affected by joint faults. The fault is modelled like a reduction of the maximum joint velocity due to partial torque loss of a servomotor. The system is made locally fault tolerant by exploiting its redundancy degrees via two different methods. The first method permits to reduce the manipulators' manipulability loss in case of fault occurrence, by imposing an optimal fault tolerant configuration to both manipulators; this is directly obtained from the knowledge of the Jacobian null space. The second method permits to compensate the loss of the end effectors motion using the saturation null space approach. In order to handle the dual arm system as a unique equivalent manipulator, both methods are formulated according to the relative Jacobian method. The efficiency of the proposed joint fault tolerant control is shown in a case study, where joint faults occur at different time instants.