Lightweight and compact smart walking cane

Abstract

Devices such as canes and wheelchairs, used to assist locomotion, have remained mostly unchanged for centuries. Recent advances in robotics have the potential to develop smart versions of these devices that can offer better support and living conditions to their users. This is the underlying objective of this project. The existing devices, used during recovery and rehabilitation phases where gait stability is key, are often bulky and cannot be easily migrated from hospital to domestic environments, where maneuvering space tends to be restricted. This article discusses a compact, lightweight and minimally invasive, robotic cane to assist locomotion. The device can assist users with mild locomotion disabilities, e.g., in the final stages of rehabilitation, to maintain and recover their balance in standing and walking situations. This extends previous experiments with alternative control strategies, merged with indicators (based on the Gini index) able to recognize differences between users. Several experiments, with a range of users possessing different mobility impairments, confirm the viability of the robotic cane, with users comfortably using the cane after three minutes, on average, proving its ease of use and low intrusiveness, and with constant support offered during the whole movement. Furthermore, the real-time tuning of the controller gains, via the Gini inequality index, enables adjustment to the user's movement.

Keywords: Full-state feedback; Inertial unit; Lightweight and compact smart cane; Locomotion assistance; Mild locomotion disabilities; Robot cane; Unicycle kinematics.

Figure 1. Schematic of the model.

Figure 2. Control diagram.

Figure 3. Prototype of the smart cane.

Figure 4. Gain-scheduled LQR feedback controller.

Figure 5. Integration of the Gini index in the low level control.

Figure 6. Analysis of the ability of the cane to bear the user’s weight.

Figure 7. Support polygon area in different situations: no cane, cane at side, and cane at front.

Figure 8. Differences in the force applied and the angle between robotic and traditional cane with an elderly user.

Figure 9. Behaviour of robotic cane with a reduced mobility user getting up from a chair.

Figure 10. Representation of the behaviour of traditional (A) and robotic (B) canes with a user with reduced mobility getting up from a chair.

Figure 11. Behaviour of robotic cane with a user with reduced mobility on smooth pavement.

Figure 12. Behaviour of robotic cane with a user with reduced mobility on textured pavement.

Figure 13. Behaviour of robotic cane with a reduced mobility user on a pavement with moderate irregularities.

Figure 14. Behaviour of robotic cane with a user with reduced mobility on a pavement with debris.

Figure 15. Behaviour of the robotic cane with first (A) and second (B) control methods.

Figure 16. Test with user 1.

Figure 17. Test with user 2.

Figure 18. Test with user 3.

Figure 19. Test with user 4.

Figure 20. Test with user 5.

Figure 21. Test with user 6.

Figure 22. Test with user 7.

Figure 23. First test with a user without mobility problems.

Figure 24. Second test with a user without mobility problems.

Figure 25. Effect of the Gini index on the controller gains.

Figure 26. Test comparing first cane with and without Gini index.

Figure 27. Test comparing second cane with and without Gini index.