In this project we propose to develop a new reliable human-machine interface (HMI) to improve the level of autonomy of assistive robotic manipulation systems and provide patients with the ability to comfortably control them with minimal efforts by incorporating a real-time motion planner based on model predictive control (MPC) and hidden Markov models (HMM) methods so is to follow the human control commands while autonomously avoiding obstacles. In this way an impaired patient controlling the robotic arm can obtain precise reaches with little effort. In addition to avoiding potential damage of the robot and the objects in its environment, this approach would also shorten the time needed to train the patients, therefore improving the usability of such systems and, thus, offering a way to improve the patients’ quality of life. This HMI will constitute a new approach to prosthetic applications for patients with neurological conditions such as Cerebral Pulsy, Parkinson’s disease, stroke and multiple sclerosis. The secondary goal of the project is to advance theoretic and experimental research in robotics, control engineering and machine learning at Nazarbayev University through publication of no less than 3 high-impact factor international journal papers and training of young highly skilled Kazakhstani researchers including master and PhD students that will be able to develop novel innovative technologies in future.