Stewart platform based robot design and control for passive exercises in ankle and knee rehabilitation

Creative Commons License

Budakli M. T., Yilmaz C.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, vol.36, no.4, pp.1831-1846, 2021 (SCI-Expanded) identifier identifier identifier

  • Publication Type: Article / Article
  • Volume: 36 Issue: 4
  • Publication Date: 2021
  • Doi Number: 10.17341/gazimmfd.846641
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Art Source, Compendex, TR DİZİN (ULAKBİM)
  • Page Numbers: pp.1831-1846
  • Keywords: Lower limb rehabilitation robot, stewart platform, parallel, robot, Newton-Raphson method, position control, PERFORMANCE
  • Yıldız Technical University Affiliated: Yes


Nowadays, usage of robots in industry and daily life is increasing, additionally, different robot designs are developed in order to fulfill the applications optimally by means of robots. Parallel robots have a restricted workspace due to their closed chain kinematic structure. However, this kinematic structure provides the robot high positioning accuracy and rigidity. Precise positioning and rigidity are most important key aspects for rehabilitation robots, therefore parallel robots are very suitable for rehabilitation applications. However, the limited workspace of the parallel robot creates a limitation in joint range of motion. In this study, a new rehabilitation robot was designed with the addition of Stewart Platform structure, 7th linear actuator and mechanical part, thus extending the workspace and providing rehabilitation to ankle and knee joints. The robot is configured by combining mechanical design, electronic hardware, mathematical model based on the Newton-Raphson method for forward kinematics as well as vectorial approach for inverse kinematics, and PID position control. Experiments were performed on human leg; ankle and knee range of motion and positioning results are evaluated and compared with literature. According to experimental measurements, the time delay in trajectory tracking was less than 0.5 seconds and the maximum angle deviation was 1.2 degrees.