Analysis and Preliminary Design of a Passive Upper Limb Exoskeleton

This paper reports the analysis and preliminary design of a passive, wearable, upper limb ExoSkeleton (ES) to support workers in industrial environments in a vast range of repetitive tasks, offering an effective strategy to reduce the risk of injuries in production lines [1-2]. The system primary purpose is to compensate for gravity loads acting on the human upper limb. The proposed ES is based on 6 Degrees-of-Freedom (DoFs) kinematics with 5-DoFs for the shoulder joint (two displacements plus three rotations) and 1-DoF for the elbow (Fig. 1) [3]. Gravity compensation is implemented with passive elastic elements to minimize weight and reduce cost [4]. A detailed analytical tool is developed to support the designer in the preliminary design stage, investigating the ES kinetic-static behavior and deriving optimal design parameters for the springs over the human arm workspace (Figs. 2, 3). By defining specific functional requirements (i.e., the user’s features and simulated movements), computationally efficient optimization studies may be carried out to determine the optimal coefficients and positions of the springs, thus, maximizing the accuracy of the gravity balancing. Two 1-DoF balancers [5] (for R1 and R5), and one novel 3-DoFs balancer [6] (for the shoulder joint, i.e., R2, R3, R4) are employed. Comparing the balancing springs type to be used (Zero/non-Zero Free Length) [7], several results are investigated by evaluating i) two different 3-DoFs balancer optimal configurations (i.e., arrangement of the elastic elements); ii) the task to be reproduced (MOV 1, MOV 2, MOV 3); iii) the ES material properties (HP. Aluminum alloy links, Carbon Fiber links, null mass links). The obtained results are validated with the commercial multi-body tool RecurDyn for some relevant movements of the user’s arm (Fig. 3). Designing a lightweight, simplified assembled ES, the proposed analytical tool presents advantages in terms of versatility in the performed movement, customization according to the user’s features, and availability of the MATLAB open-source data for the research community.