The wearable upper limb exoskeletons designed using the traditional shoulder model have a serious issue related to joint axis misalignment with the human shoulder. The misalignment is primarily due to the limited degrees of freedom model assumed for the shoulder. The glenohumeral joint in the conventional upper limb model is considered a single spherical joint, which, when considering the human anatomy, neglects the humeral head translation from the glenoid fossa of the scapula. This shoulder model cannot be used for designing the wearable upper limb exoskeletons, which can damage the shoulder joint due to interactive forces generated from the human-exoskeleton joint misalignment. This paper presents a modified, anatomically similar, human upper limb model where the single spherical joint of the glenohumeral joint is modified into two spherical joints connected by a prismatic joint. The kinematic and dynamic model of the proposed shoulder model is explained in detail. The elbow reachable space for this model has been calculated and compared with that of the existing models, and it is found to be closer to the actual workspace. The proposed model is simulated with a conventional exoskeleton attachment to show the effect of misalignment during humeral head translation and the benefits of using the proposed model in the design of the shoulder exoskeleton. An exoskeleton test bench that uses the modified shoulder model is also presented.

The human need for rehabilitation, assistance, and augmentation has led to the development and use of wearable exoskeletons. Upper limb exoskeletons under research and development are tested on human volunteers to gauge performance and usability. Direct testing can often cause straining of the joints, especially the shoulder joint, which is the most important and flexible joint in the upper extremity of the human body. The misalignment of joint axes between the exoskeleton and the human body causes straining. To avoid this, we propose designing and developing a novel human shoulder phantom mimicking the shoulder complex motion and the humeral head translation that can help in the real-time testing of exoskeletons without the need for human volunteers. The device can be used to test the interaction forces and the maximum reachable position of the exoskeleton. It consists of three degrees of freedom (DOF) passive shoulder girdle mechanism and seven DOF glenohumeral joint mechanisms, of which six are passive revolute joints and one is an active prismatic joint mimicking the humeral head translation. All the passive joints are spring-loaded and are incorporated with joint angle sensors. A custom-made, three-axis force sensor measures the human–exoskeleton interaction forces. The design details, selection of joint springs, linear actuation mechanism, and the analysis of the phantom’s reachable workspace are presented. The device is validated by comparing the interaction forces produced during the conventional exoskeleton-assisted and human-assisted phantom arm elevation.

Upper limb wearable exoskeletons are devices used for assistive and augmentative applications in work environments. The shoulder mechanism of the upper limb exoskeleton is vital in fulfilling its objective by exhibiting its capability to move with the human shoulder in unity without any joint axis misalignment. Any misalignment can cause strain on the human shoulder complex due to unwanted interaction forces between the user and the exoskeleton. Most existing portable assistive or augmentation exoskeletons fail to exhibit the capability to self-align when there is a joint axis misalignment due to design-related issues. This paper presents the design, analysis, and experimental evaluation of a novel human-centric shoulder exoskeleton with a self-aligning shoulder mechanism for an average adult male. The novelty of the design is that it considers the human shoulder’s humeral head (HH) translation and incorporates a novel mechanism to reduce misalignment. The system’s design details, workspace analysis, and prototype testing and analysis are explained. The proposed exoskeleton is evaluated and validated using simulation and experimentation on a standardised test bench. The results are compared with the existing conventional exoskeleton and human-assisted motion, and it was found that the proposed exoskeleton is capable of reducing the interaction forces due to misalignment by 99%. Further, a simple assistive control architecture is proposed and implemented for the exoskeleton for efficient operation.