Improving active lower-limb exoskeleton performance with the Reaching Movement feature

The number of people with a mobility disorder caused by stroke, spinal cord injury (SCI) or other related diseases is increasing rapidly. Damage to central and peripheral nervous systems often leads to a gait impairment and reduced mobility or, in the worst cases, paralysis: even the activities of daily living (ADLs) become quite challenging. Nowadays, with the goal of giving a normal life back to SCI patients, a new technological achievement has been reached: lower-limb exoskeletons. Lower-limb exoskeletons are wearable robotic devices which provide impaired people with the possibility to perform normal ambulatory functions. The features of standing up/sitting down and walking are features that every exoskeleton commercially available has, since they were born in order to enable system-assisted walking for wheelchair users. Then, considering the ADLs, people would need, for example, to lower themselves to pick up objects or to lean forward or sideways without losing balance, as it happens for able-bodied subjects. The purpose of this master thesis is, then, that of extending exoskeletons' functionalities with the possibility to perform the act of reaching. This implies that the patient inside the exoskeleton will have the possibility to lower down to pick up objects which are positioned within a certain distance from exoskeleton's feet and within a certain range of height from the ground. It is important to point out that only hands-free exoskeletons, i.e. exoskeletons where the user does not need to use crutches to support himself/herself, are eligible to benefi t from this upgrade. Starting from the definition of the so-called Reaching Movement, this is first measured in an able-bodied subject, using the exoskeleton turned off as a tool for the data collection process. Then, from the analysis of the experiment, it is possible to extract a mathematical relationship between the variables which contribute to define the act of reaching, i.e. joints angles and thorax pitch angle. Afterwards, a prioritized task-space controller is designed, where the mathematical relationship found earlier is embedded inside a Stack of Tasks (SoT). A Hierarchical Quadratic Program (HQP) solver is used for solving the Inverse Kinematics problem at each level of the stack. Once the new feature is validated with Gazebo simulator, the control strategy is proven with the exoskeleton turned on and three testers. Finally, the achieved results are compared and discussed, and the feature is integrated into the functionalities of the robot. Future improvements are proposed too.