Abstract
Background
A new knee simulator has been developed at Ghent University. This simulator provides the unique opportunity of evaluating the knee kinematics during activities of daily living. The simulator therefore controls the position of the ankle in the sagittal plane while keeping the hip at a fixed position. This approach provides full kinematic freedom to the knee. To evaluate and validate the performance of the simulator, the development of and comparison with a numerical simulation model is discussed in this paper.
Methods
Both a two and three dimensional simulation model have been developed using the AnyBody Modelling System (AMS). In the two dimensional model, the knee joint is represented by a hinge. Similarly, the ankle and hip joint are represented by a hinge joint and a variable amplitude quadriceps and hamstrings force is applied. In line with this simulation model, a hinge model was created that could be mounted in the UGent knee simulator to evaluate the performance of the simulated model. The hinge model thereby performs a cyclic motion under varying simulated muscle loads while recording the ankle reaction forces. In addition to the two dimensional model, a three dimensional model has been developed. More specifically, a model is built of a sawbone leg holding a posterior stabilised single radius total knee implant. The physical sawbone model contains simplified medial and lateral collateral ligaments. In line with the boundary conditions of the UGent knee simulator, the simulated hip contains a single rotational degree of freedom and the ankle holds four degrees of freedom (three rotations, single translation). In the simulations, the knee is modelled using the force-dependent kinematics (FDK) method built in the AMS. This leaves the knee with six degrees of freedom that are controlled by the ligament tension in combination with the applied quadriceps load and shape of the implant. The physical sawbone model goes through five cycles in the UGent simulator using while recording the kinematics of the femur and tibia using a set of markers rigidly attached to the femur and tibia bone. The position of the implant with respect to the markers was evaluated by CT-scanning the sawbone model.
Results and Discussion
In a first step, the reaction forces at the ankle in the 2D model were evaluated. The difference between the simulated and measured reaction force is limited and can be explained from a slight variation of the attachment point of the simulated muscle loads. For the 3D model, the kinematic patterns have been evaluated for both the simulation and physical model using Grood & Suntay definitions. The kinematic parameters display realistic trends, however, no exact match has been obtained for all parameters so far. The latter might be attributed to a number of simplifications in the simulation model as well as elastic deformation of the physical sawbone model.
Conclusion
A three dimensional model of a knee implant in the UGent Knee Simulator has been developed. The simulated kinematic patterns appear realistic though no exact match with the measured patterns has been obtained. Future research will therefore focus on the development of a more realistic experimental and numerical model.
