Simulation-Based Design of Exoskeletons Using Musculoskeletal Analysis


rehab02Exoskeletons are a new class of articulated mechanical systems whose performance is realized while in intimate contact with the human user. The overall performance depends on many factors including selection of architecture, device, parameters and the nature of the coupling to the human, offering numerous challenges to design-evaluation and refinement. In this work, we discuss merger of techniques from the musculoskeletal analysis and simulation-based design to study and analyze the performance of such exoskeletons. A representative example of a simplified exoskeleton interacting with and assisting the human arm is used to illustrate principal ideas. Overall, four different case-scenarios are developed and examined with quantitative performance measures to evaluate the effectiveness of the design and allow for design refinement. The results show that augmentation by way of the exoskeleton can lead to a significant reduction in muscle loading.

Exoskeletons and orthoses are defined as mechanical devices that are essentially anthropomorphic in nature, are 'worn' by an operator and fit closely to the body, and work in concert with the operator's movements. While both devices serve to augment the performance of wearer, ‘orthoses’ tend to be used as assistive devices that are used by a person with limb pathology whereas ‘exoskeletons’ are worn by able-bodied users. Performance enhancement studies in the past range from improved environment interaction and/or metabolic economies among others.

There are numerous industrial, military and medical applications of exoskeletal robotic systems. In 1960s and 1970s the main focus was the application of active exoskeletons for industry and medical purposes. In early 1990s, some of them were designed to augment the strength of the humans. Recently, their use in the area of rehabilitation and power assist became significant in the society for the individuals with a physical weakness (due to age, injury and/or handicap). The upper limb exoskeletons are primarily used for teleoperation and power amplification. Finally, due to their ability to apply independent dynamic forces on human limbs, these devices are providing a basis for neuromotor research. Every field of application has its specific requirements in terms of structural design and control algorithms.

Exoskeletons have been used in various operational modes including the assistive mode and the resistive mode. In the assistive mode the exoskeleton provides power to support the movement of the human limb, while in the resistive mode it opposes motion/forces. Though significant literature is available that discusses the design of upper-limb exoskeletons, but there still is a necessity to study the effects of exoskeletons on human musculoskeletal system to improve the design of these robotic devices. However, there are many challenges related to (i) kinematic compatibility, (ii) dynamic matching, (iii) lack of performance evaluation criterion, and (iv) lack of design and analysis tools.

Simulation-based-design, otherwise known as Virtual Prototyping (VP), is a methodology to iteratively refine design of a product using computer-based functional physical simulation(s). Rapid quantitative and computational investigation of numerous “what–if” design scenarios at relatively low cost is what makes VP a successful design refinement technique. This can include studying the effects of variability; determining the “best” geometries for performance; and to examine the linkage between form and function using “virtual experimentation” to name a few.

In this work, we present the use of musculoskeletal analysis for designing an upper-limb exoskeleton.  Four different case studies are performed to study the effect of using a simplified exoskeleton on the muscle loading for arm curl with dumbbell. The simulation results showed that with the use of exoskeleton significant reductions in both, individual muscle forces and elbow flexion moment are achievable. The results also showed that the exoskeleton applied-torque synchronization with the required torque is important for the performance of the device. Prior approaches to exoskeleton designs used a more qualitative designer assessment to describe performance and/or fit. This engenders the usual limitations inherent to any semi- quantitative/qualitative design methodology including lack of invariances etc. In contrast, musculoskeletal analysis provide rational basis for biomechanically quantifying the performance of a candidate exoskeleton design and thus in turn provides a means for quantitatively comparing alternate designs. Nevertheless, the resulting copious amounts of raw quantitative data need to be further processed to extract useful metrics. Careful assessment of the quality, sensitivity and most importantly usability, of both the raw information and extracted metrics, is the focus of our current research.


 Students Involved:

- Priyanshu Agarwal, MS Student, University at Buffalo
- Leng-Feng Lee, PhD Student, University at Buffalo

- Madusudanan S. Narayanan, PhD Candidate, University at Buffalo


 Research Issue :

Kinematic and Dynamic Matching

The task of correctly positioning the exoskeleton on the human arm model offers a challenge as alignment of critical axes is important for biomechanically correct operation. The issue of matching the axes and its effects on arm movement becomes more significant when axes move in space and time, due to the imperfect nature of the human arm joints. Past work on prosthetic design focuses on using interesting mechanisms such as 6 bars to allow alignment of such axes. The action of the misalignment above a certain degree can give rise to an unnatural restraint and affect limb mobility. Thus, the human-robot physical interaction imposes several constraints and requirements on the design of wearable exoskeletons. The problem may be analyzed in a detailed manner by conducting a parametric study and understanding the effects of the misalignment on human muscle force and activity. Typically such over-constraint is released by addition of mechanical compliance.


Lack of Quantitative Performance Metrics

Traditionally physiotherapists use criteria such as Rivermead Motor Assessment Score, time to completion etc. to quantify the performance of the device. However, such an assessment is made based on the performance of the wearer alone. Musculoskeletal analysis, on the other hand, allows monitoring of internal human variables – a wide variety of biologically relevant data (from lengths, forces, reactions of muscles/tendons/joints, to metabolic power consumption, and mechanical work) which captures both the nature of the device and the wearer. Alternatively, other higher level abstracted performance measures may be developed, allowing a designer to flexibly assess the performance of a design. In this work, we have considered individual muscle force and the elbow flexion moment as our performance criteria. Also, we select few important muscles in the human arm i.e. biceps, triceps, brachialis and supinator for comparing the case scenarios.


Lack of Design and Analysis Tools

Recently, a few tools such as SIMM (Software for Interactive Musculoskeletal Modeling), OpenSim, AnyBody Modeling System, SimTK, LifeModeler, Virtual Interactive Musculoskeletal System (VIMS) in the form of commercial packages have been made available for musculoskeletal analysis. These computational tools perform kinematic and dynamic analyses of vertebrate musculoskeletal systems, building on an articulated multi-body systems (AMBS) framework. Constrained musculoskeletal system-level computational models can be constructed modularly by placing physiologic and behavioral constraints on anatomical components (e.g., bone, muscle, and tendon). In our study, AnyBody Modeling System is employed for the analysis due to readily available library of human models.


 Movies :


1. Case A: Arm Curl with Dumbbell

- This is the benchmark case for all other cases.

- Larger Screen: view it on YouTube:


2. Case B: Arm Curl with Dumbbell and Constant Assistive Moment

- In this case, a constant assistive moment is applied at the human elbow joint for assistance to understand the reduction in muscle loading when no axes misalignment is present.

- Larger Screen: view it on YouTube:

3. Case C: Arm Curl with Dumbbell and Variable Assistive Moment

- In this case, a variable assistive moment is applied at the human elbow joint for assistance to understand whether it is possible to eliminate the muscle loading if the moment requirement is exactly matched.

- Larger Screen: view it on YouTube:


4. Case D: Arm Curl with Dumbbell and Exoskeleton Variable Assistive Moment

- The human arm coupled with the exoskeleton is simulated in this case with a variable assistive  moment being provided by the exoskeleton.

- Larger Screen: view it on YouTube:





























 Related Publications - Conference Proceedings:


P. Agarwal, M. S. Narayanan, L-F. Lee, F. Mendel and V. N. Krovi, "Simulation-based Design of Exoskeleton Using Musculoskeletal Analysis", proceeding of ASME 2010 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference, Montreal, Quebec, Canada, August 15-18, 2010. ( CIE Best Paper Award IDETC 2010)


Sponsor: This project was funded by Research Foundation of State University of New York, National Science Foundation CAREER Award (IIS-0347653) and CNS-0751132.

Last Updated: August 28, 2010