Design and validation of a general purpose robotic testing system for musculoskeletal applications

Lawrence D Noble Jr,Robb W Colbrunn,Dong-Gil Lee,Antonie J van den Bogert,Brian L Davis

DOI: https://doi.org/10.1115/1.4000851

Abstract:Orthopaedic research on in vitro forces applied to bones, tendons, and ligaments during joint loading has been difficult to perform because of limitations with existing robotic simulators in applying full-physiological loading to the joint under investigation in real time. The objectives of the current work are as follows: (1) describe the design of a musculoskeletal simulator developed to support in vitro testing of cadaveric joint systems, (2) provide component and system-level validation results, and (3) demonstrate the simulator's usefulness for specific applications of the foot-ankle complex and knee. The musculoskeletal simulator allows researchers to simulate a variety of loading conditions on cadaver joints via motorized actuators that simulate muscle forces while simultaneously contacting the joint with an external load applied by a specialized robot. Multiple foot and knee studies have been completed at the Cleveland Clinic to demonstrate the simulator's capabilities. Using a variety of general-use components, experiments can be designed to test other musculoskeletal joints as well (e.g., hip, shoulder, facet joints of the spine). The accuracy of the tendon actuators to generate a target force profile during simulated walking was found to be highly variable and dependent on stance position. Repeatability (the ability of the system to generate the same tendon forces when the same experimental conditions are repeated) results showed that repeat forces were within the measurement accuracy of the system. It was determined that synchronization system accuracy was 6.7+/-2.0 ms and was based on timing measurements from the robot and tendon actuators. The positioning error of the robot ranged from 10 microm to 359 microm, depending on measurement condition (e.g., loaded or unloaded, quasistatic or dynamic motion, centralized movements or extremes of travel, maximum value, or root-mean-square, and x-, y- or z-axis motion). Algorithms and methods for controlling specimen interactions with the robot (with and without muscle forces) to duplicate physiological loading of the joints through iterative pseudo-fuzzy logic and real-time hybrid control are described. Results from the tests of the musculoskeletal simulator have demonstrated that the speed and accuracy of the components, the synchronization timing, the force and position control methods, and the system software can adequately replicate the biomechanics of human motion required to conduct meaningful cadaveric joint investigations.