In the last two decades, robotics-assisted stroke reha-bilitation has been wide-spread, in particular for movement rehabilitation of upper limbs. Several studies have reported on the clinical effectiveness of this kind of therapy. The results of these studies show that robot assisted therapy can be more effective in recovering motor control abilities than conventional therapy. On the other hand, studies found no signifi-cant improvement on motor function abilities of pa-tients. These contradictory results stimulated our re-search to survey current status of robotics-assisted rehabilitation and to look for advancement opportuni-ties. We developed a reasoning model that help us conduct the study systematically and to consider the four most important aspects, namely (i) the post-stroke pathophysiological status of patients, (ii) the nature of the rehabilitation therapies, (iii) the versatil-ity of the robotic rehabilitation instruments, and (iv) the kind of stimulation provided for patients. Our major finding is that there are strong evidences that the efficacy of robotics-assisted rehabilitation can be increased by motivation and engagement. We con-cluded that by exploiting the opportunities offered by cyber-physical systems and gamification, a signifi-cant improvement of context sensitive engagement can be realized. Our follow-up research will study various implementation opportunities, the affordabilities of various cyber-physical solutions, and in-fluence on patients.