When a knee flex deeply, the posterior side of thigh and calf contact. The contact force is unignorable to estimate the load acting on a knee because the force generates extensional moment on the knee, and the moment might be about 20–80% of the flexional moment generated by a floor reacting force. Besides, the thigh-calf contact force varies so much even if the posture or the test subject are the same that it is hard to use the average value to estimate the knee load. We have assumed that the force might change not only by the individual physical size but also by a slight change of the posture, especially the angle of the upper body. Therefore we tried to create the estimation equation for the thigh-calf contact force using both anthropometric sizes and posture angles as parameters.

The objective posture was kneeling, both plantarflexing and dorsiflexing the ankle joint. Test subjects were 10 healthy males. They were asked to sit on a floor with kneeling, and to tilt their upper body forward and backward. The estimation equations were created as the linear combinations of the parameters, determining the coefficient as to minimize the root mean square errors. We used the parameters as explanatory variables which could be divided into posture parameters and individual parameters. Posture parameters included the angle of upper body, thigh and lower thigh. Individual parameters included height, weight, axial and circumferential lengths of thigh and lower thigh. The magnitude of the force was normalized by a body weight, and the acting position was expressed by the moment arm length around a knee joint and normalized by a height.

As a result, the adjusted coefficient of determination improved and the root mean square error decreased when using both posture and individual parameters, though there were large errors when neglecting either parameters. The accuracy decreased little when using the same equation for plantarflexed and dorsiflexed kneeling in magnitude. The relation of measured and estimated values of the magnitude and acting position, using the common equation with all the parameters. It might be because the difference of the postures could be described by the inclination angle of a thigh. In both postures, the magnitude of a thigh-calf contact force was mainly affected by the posture and acting position by the individual parameters. When calculating the knee joint load, the errors would be about 8.59 Nm on the knee moment and 290 N on the knee load when using just an average, and they would decrease to 2.23 Nm and 74 N respectively using the estimation equation.

Many recent knee prostheses are designed aiming to the physiological knee kinematics on tibiofemoral joint, which means the femoral rollback and medial pivot motion. However, there have been few studies how to design a patellar component. Since patella and tibia are connected by a patellar tendon, tibiofemoral and patellofemoral motion or contact forces might affect each other. In this study, we aimed to discuss the optimal design of patellar component and simulated the knee flexion using four types of patellar shape during deep knee flexion.

Our simulation model calculates the position/orientation, contact points and contact forces by inputting knee flexion angle, muscle forces and external forces. It can be separated into patellofemoral and tibiofemoral joints. On each joint, calculations are performed using the condition of point contact and force/moment equilibrium. First, patellofemoral was calculated and output patellar tendon force, and tibiofemoral was calculated with patellar tendon force as external force. Then patellofemoral was calculated again, and the calculation was repeated until the position/orientation of tibia converged.

We tried four types of patellar shape, circular dome, cylinder, plate and anatomical. Femoral and tibial surfaces are created from Scorpio NRG PS (Stryker Co.). Condition of knee flexion was passive, with constant muscle forces and varying external force acting on tibia. Knee flexion angle was from 80 to 150 degrees.

As a result, the internal rotation of tibia varied much by using anatomical or plate patella than dome or cylinder shape. Although patellar contact force did not change much, tibial contact balances were better on dome and cylinder patella and the medial contact forces were larger than lateral on anatomical and plate patella. Thus, the results could be divided into two types, dome/cylinder and plate/anatomical. It might be caused by the variations of patellar rotation angle were large on anatomical and plate patella, though patellar tilt angles were similar in all the cases. We have already reported that the anatomical shape of patella would contact in good medial-lateral balance when tibia moved physiologically, therefore we have predicted the anatomical patella might facilitate the physiological tibiofemoral motion. However, the results were not as we predicted. Actually our previous and this study are not in the same condition; we used a posterior-stabilized type of prosthesis, and the post and cam mechanism could not make the femur roll back during deep knee flexion.

It might be better to choose dome or cylinder patella to obtain the stability of tibiofemoral joint, and to choose anatomical or plate to the mobility.

In some regions in Asia or Arab, there are lifestyles without chair or bed and sitting down on a floor directly, by flexing their knee deeply. However, there are little data about the joint angles, muscle forces or joint loads at such sitting postures or descending to and rising from the posture. In this study, we report the knee joint force and the muscle forces of lower limb at deep squatting and kneeling postures.

The model to estimate the forces were constructed as 2D on sagittal plane. Floor reacting force, gravity forces and thigh-calf contact force were considered as external forces. And as the muscle, rectus and vastus femoris, hamstrings, gluteus maximus, gastrocnemius and soleus were taken into the model. The rectus and vastus were connected to the tibia with patella and patella tendon. First the muscle forces were calculated by the moment equilibrium conditions around hip, knee and ankle joint, and then the knee joint force was calculated by the force equilibrium conditions at tibia and patella.

For measuring the acting point of the floor reacting force, thigh-calf contact force and joint angles during the objective posture, we performed the experiments. The postures to be subjected were heel-contact squatting (HCS), heel-rise squatting (HRS), kneeling and seiza (Japanese sedentary kneeling), as shown in the Fig.1. The test subjects were ten healthy male, and the average height was 1.71[m], weight was 66.1[kgf] and age was 21.5[years]. The thigh-calf contact force and its acting point were measured by settling the pressure distribution sensor sheet between thigh and calf.

Results were normalized by body weight, and shown in Fig.1. The thigh-calf contact force was the largest at the heel-rise squatting posture (1.16BW), and the smallest at heel-contact squatting (0.60BW). The patellofemoral and the tibiofemoral joint forces were shown in the figure. Both forces were the largest at the heel-contact squatting, and were the smallest at the seiza posture. And it might be estimated that the thigh-calf contact force acted anterior when the ankle joint dorsiflexed, and the force was larger when the hip joint extended. The thigh-calf contact force might be decided by not only the knee joint angle but also the hip and ankle joints.

As a limitation of this study, we should mention about the effect of the neglected soft tissues. It could be considerable that the compressive internal force of the soft tissues behind a knee joint substance the tibiofemoral force, and then the real tibiofemoral force might be smaller than the calculated values in this study. Then, the tensile force of quadriceps also might be smaller, and then the patellofemoral joint force is also small.

Thigh-calf contact force is the force acting on posterior side of the thigh and calf during deep knee flexion. It has been reported the force is important to analyze the kinetics of a lower limb and a knee joint. Some previous researches reported the measured thigh-calf contact force, however, the values varied among the reports. Furthermore, the reports indicated that there were large variations even in a single report. One of the reports tried to find the relationship between the magnitude of thigh-calf contact force and anthropometric measurement as height, weight or perimeter of the lower limb, however, there could not found clear correlations. We considered that the cause of the variations might be the difference of the posture. At heel-rise squatting posture, we can bend or stand upright the upper body. Therefore we tried to create the equation to estimate the thigh-calf contact force by multiple regression analysis, using the anthropometric and posture parameters as explanatory variables.

We performed the experiment to measure thigh-calf contact force, joint angles and anthropometric information. Test subjects were 10 healthy male. First we measured their height, weight, perimeter of the thigh and muscle mass of the legs and whole body. Muscle mass was measured by body composition meter (BC-118E, Tanita Co., Japan). Then, test subjects were asked to squat with their heels lifted and with putting the pressure distribution sensor between thigh and calf. And they bent their upper body forward and backward. The pressure sensor to be used was ConfroMat System (Nitta Co., Japan). After that, we measured the joint angles of the hip, knee and ankle, and the angle between the floor and upper body using the videos taken during the experiment. Then, we created the equation to estimate the thigh-calf contact force by linear combination of the anthropometric values and joint angles. The coefficients were settled as to minimize the average error between measured and estimated values.

Results are shown in Fig.1. Forces were normalized by the body weight of the test subjects. Because the horizontal axes show the measured and vertical axis show the estimated values, the estimation is accurate when the plots are near the 45-degree line. Average error was 0.11BW by using only physical values, 0.15BW by angles and 0.06BW using both values. And the maximum error was 0.69BW, 0.43BW and 0.32BW respectively. Thus we could estimate the thigh-calf contact force by multiple regressions, using both physical parameters and angles to indicate the posture. Using the equation, we would be able to analyze the kinetics of a lower limb by physical and motion measurement. Our future work might be increasing the number of subjects to consider the appropriateness, because the test subjects of this study were very limited.

One of the main concerns about the currently available simulators is that the TKA is driven in a “passive way” for assessment. For the simulators for the wear assessment, the tibio-femoral relative motion is automatically made by using the knee kinematics and loading profile of a normal gait. As for the simulators for the kinematics and kinetics assessment of TKA, also the predicted loading profiles introduced from the theoretical model are applied as the input data to drive the simulator. It should be noted that the human joints are driven by the muscles' forces and external loads, and their kinematics and kinetics are the “outcome”. This being so, the knee simulator should be driven by the muscles' forces and upon these conditions the TKA performance is to be assessed. Some other concerns about the current simulations are as follows. The effects of hip joint motion are not taken into account. The upper body weight is applied along a vertical rod in such a way as a crank-slider. Furthermore, few simulators are capable of knee flexion greater than about 110°.

Considering the above, we have developed a novel knee simulator which makes it possible to reproduce the active and natural knee motion to assess kinematics and kinetics of TKA. In the experiment, the custom-designed PS type TKA was attached and the simulator was operated so as to reproduce the sit-to-stand features, thereby introducing the tibio-femoral loading profiles during the motion.

Figure 1 illustrates the external appearances of the simulator and a close view of the knee joint compartment. Since our simulator is composed of a multiple inverted pendulum, the knee part bears the upper body weight in a physiological way. The holder bracket is set to prevent the simulator from collapsing for security. The dimension and weight of each link were set as close as those of each segment of a normal male subject. Our simulator is driven by the wire pull mechanism which substitutes the human musculo-skeletal system of lower limb. Figure 2 shows close views of tibial tray with load cells. In Fig.2a, cell FR, FC and FL are to measure the tangential components of tibio-femoral contact force, i.e., the Anterio-Posterior force (AP force). The rest five cells are to measure the normal components of tibio-femoral contact force (normal force). As shown in Fig.2c, the tibial insert of TKA is mounted on the lid of the tibial tray box.

In the experiment, a PS type TKA whose maximum flexion angle of 150° was attached to the simulator for evaluation. The simulator was operated so as to reproduce the sit-to-stand features and the data concerning about the AP force, F_t, and the normal force, F_n were recorded.

Figure 3 shows the variations of knee flexion angles and knee contact forces respectively as a function of normalized time. Our knee simulator may have a potential for substituting the in vivo measurement.

There have been a large number of studies reporting the knee joint force during level walking, however, the data of during deep knee flexion are scarce, and especially the data about patellofemoral joint force are lacking. Deep knee flexion is a important motion in Japan and some regions of Asia and Arab, because there are the lifestyle of sitting down and lying on the floor directly. Such data is necessary for designing and evaluating the new type of knee prosthesis which can flex deeply. Therefore we estimated the patellofemoral and tibiofemoral forces in deep knee flexion by using the masculoskeltal model of the lower limb.

The model for the calculation was constructed by open chain of three bar link mechanism, and each link stood for thigh, lower leg and foot. And six muscles, gluteus maximus, hamstrings, rectus, vastus, gastrocnemius and soleus were modeled as the lines connecting the both end of insertion, which apply tensile force at the insertion on the links. And the model also included the gravity forces, thigh-calf contact forces on the Inputting the data of floor reacting forces and joint angles, the model calculated the muscle forces by the moment equilibrium conditions around each joint, and some assumptions about the ratio of the biarticular muscles. And then, the joint forces were estimated from the muscle forces, using the force equilibrium conditions on patella and tibia. The position/orientation of each segments, femur, patella and tibia, were decided by referring the literature.

The motion to be analyzed was standing up from kneeling posture. The joint angles during the motion are shown in Fig.1. This motion included the motion from kneeling to squatting, rising the knee from the floor by flexing hip joint, and the motion from squatting to standing. The test subject was a healthy male, age 23[years], height 1.7[m], weight 65[kgw].

Results were shown in Fig.2. The patellofemoral force was little at standing posture, the end of the motion, however, was as large as tibiofemoral force during the knee joint angle was over 130 degrees. The reason of this was that the patellofemoral joint force was heavily dependent on the quadriceps forces, and the quadriceps tensile force was large at deep knee flexion, at kneeling or squatting posture. The maximum tibiofemoral force was 3.5[BW] at the beginning of standing up from squatting posture. And the maximum patellofemoral force was 3.8[BW] at the motion from kneeling to squatting posture. The conclusion was that the patellofemoral joint force might not be ignored in deep knee flexion and the design of the knee prosthesis should be include the strength design of patellofemoral joint.

1. Introduction

Such a Total Knee Arthroplasty (TKA) that is capable of making high knee flexion has been long awaited for the Asian and Muslim people. Our research group has developed the TKA possible to attain complete deep knee flexion such as seiza sitting. Yet as seiza is peculiar to the Japanese, other strategies will be necessary for our TKA to be on the overseas market. Still it is impractical to prepare many kinds of modifications of our TKA to meet various demands from every country/region. To this end, we contrived a way to modularize the post-cum alignment of our TKA in order to facilitate the following three activities containing high knee flexion: praying for the Muslim, gardening or golfing for the Westerner, sedentary siting on a floor for the Asian. We performed simulation and experiment, such as a mathematical model analysis, FEM analysis and a cadaveric study, thereby determining the optimal combination of moduli for the above activities respectively.

Knowledge of joint kinematics in the lower limb is important for understanding joint injuries and diseases and evaluating treatment outcomes. However, limited information is available about the joint kinematics required for high flexion activities necessary for floor sitting life style. In this study, the hip and knee joint kinematics of ten healthy male and ten healthy female subjects were investigated using an electromagnetic motion tracking system. We measured the hip and knee joints' functions moving into 1) kneeling on knees with legs parallel without using arms, 2) kneeling on knees with legs parallel with using arms, 3) kneeling on knees with one foot forward without using arms, 4) cross-legged sitting, 5) kneeling with legs to the side, 6) sitting with legs stretched out, and 7) deep squatting, and moving out of the above seven conditions. Conditions 1) through 3) were Japanese seiza style. On conditions 4) through 7), arms were not used. We further measured the functions of putting on and taking off a sock under such conditions as 8) with standing position and 9) sitting position (Fig 1). Here special attention was paid for flexion and extension motion. The data were used to produce the pattern of joint angulation against the percentage of the cycle for each individual conducting each activity. The kinematic curves were split into 3 phases: moving into the rest position, the rest position and out of the rest position. It should be noted that the moving into and the rest phases were split at the moment when the peak value was determined during the moving into phase. Thus the initiation of the rest phase on the curve was not coinciding with the moment the subject reached at the rest position. This was necessary in order not for the mean kinematic curve to become too dull in shape. Same was true when the end of rest phase was determined. The maximum hip and knee joint angles during the cycle were determined. Further a relationship between the hip and knee joint excursions were investigated. The results indicated condition 8) requires the maximum flexion angles to the hip among all conditions, 157.5 ± 20.4° and condition 3) to the knee joint, 157.1 ± 10.0° respectively (Fig 2). The results also indicated in many activities, the maximum joint angles were recorded not during the rest phase but during the moving into or out of phase. In any conditions even including donning on and off a sock, a strong relationship was found between the hip and knee joints motion (Fig 3), indicating the bi-articular muscles' co-contraction during the sit to stand activities. The data presented in this study will increase the knowledge of high-flexion needs especially in non-Western cultures and provide an initial characterization of the prosthesis kinematics in high flexion.

There are several concerns about the current simulators for TKA. First, the knee is flexed in a “passive way” under the condition of applying constant muscular tension forces. Second, the effects of hip joint motion are not taken into account. Thirdly, the external load for example, upper body weight is not applied in a natural way. Finally, few simulators are capable of knee flexion greater than about 100°.

To this end, we have developed a novel knee simulator system that reproduces the active and natural knee motion to evaluate kinematics and joint forces of TKA. Our simulator system has the following advantages and innovative features. First, it is driven directly by muscles' tension forces, and the knee is capable of active flexion. Secondly, a hip joint is incorporated into it and the lower limb motion is achieved in a synergistic way between the hip and knee joints. Thirdly, it is capable of complete deep knee flexion up to 180°.

Figure 1 shows the structure of the system. Both the hip and knee joints are moved by the tension forces of four wires that simulate the functions of the mono-articular muscles ((1), (3)) and the bi-articular muscles ((2), (4)) by means of a multiple pulley system (Fig 2). The femoral and tibial components of TKA are secured in the distal end of the upper link (thigh) and the proximal end of the lower link (shank) respectively. The ankle assembly has three sets of rotary bearings whose axes intersect at a fixed point, the center of the ankle, allowing spherical movement of the tibia about the ankle center. Springs were stretched around the ankle center to substitute the muscles around the ankle. Weights I and II are counterweights so as to duplicate the weights of the human upper body, thigh and shank respectively. The wires are pulled to produce the hip and knee motions. The linear bearings running along vertical rods also prevent the system from collapsing.

In the experiment, a custom-designed posterior stabilized type TKA was attached to the simulator system for evaluation. The system was operated so as to reproduce the sit-to-stand features in a quasi-static manner in order to study the kinematics of TKA. Beyond 130°, the knee proceeded to flex passively because of upper body weight. Conspicuous internal/external rotation or valgus/varus motion of the tibia relative to the femur was not observed as the knee flexed. When our simulator system was driven in a quasi-static manner, it was able to measure the kinematics of TKA however, when the system was driven in a dynamic manner, it oscillated because the springs around the ankle were not stiff enough to hold the inverted pendulum-like system upright and the ratios of the tension force exerted by the four wires simulating muscles could not be determined appropriately.

The objective of this study is to investigate the effect of the tensile force ratio between the two extensor muscles for the hip joint on the forces acting on the knee joint. We have created a mathematical model of lower limb and have performed some simulations to introduce the forces acting on the knee joint for various daily activities. With only one exception, our results for knee joint forces were in good or close agreement involving all range of knee flexion either with the in vivo data or other literature data. The exception was that, at high knee flexion angle (knee bend), the tangential components of knee joint force became pretty larger than those from the in vivo data, while the normal components did not differ much with each other though as shown in Fig. 1.

We considered that the above mentioned discrepancy was attributed to the fact that in order to solve an indeterminate problem, we had assumed the hamstrings and the gluteus maximus work together with the same force with each other, thereby introducing the hamstrings force too great. Then we expected that the above discrepancy could be eliminated if we change the tensile force ratio between the hamstrings and the gluteus maximus basing upon a certain biomechanical criterion, for example the biological cross-sectional areas.

Thus we modified our model so that we could introduce the knee joint forces as a function of the tensile force ratio. Simulation was performed for the various tensile ratio values and it was found that the knee joint force was sensitively affected by the tensile ratio and the above mentioned discrepancy between the simulation results and the in vivo data could be eliminated if the ratio value was appropriately chosen. Figure 2 shows the situation; Variations of F_n and F_t as a function of knee angle q for the various tensile force ratio r between the hamstrings and the gluteus maximus. Where, r=1.56 was determined from the biological cross-sectional areas of the hamstrings and the gluteus maximus and r=4.5 was determined so that the simulation results best fit to the in vivo data.

It has been criticized that there exist large variations of knee joint forces obtained from model analyses. And the reasons for this have been attributed to for example such facts that the model is 2D and the parameter values are incorrect. Yet, another important issue may be to find out the way how to determine the value of the synergetic muscles' force ratio with reflecting a biological rationality.

We have developed a novel knee simulator that reproduces the active knee motion to evaluate kinematics and joint reaction forces of TKA.

There have been developed many kinds of knee simulators; Most of them are to predict TKA component wear and the others are to evaluate the kinematics and/or kinetics of TKA. The most simulators have been operated using the data of the loading and kinematics profile of the knee obtained from normal gait. Here a problem is that such variables as joint force and kinematics are the outcome caused by the application of muscles' and external forces. If so, a simulator should be operated by the muscles' and external forces so as to duplicate the in vivo condition. Other disadvantages for the current knee simulators are; a knee joint motion is made passively, the effects of the hip joint motion are not taken into account, and the maximum flexion angle is usually limited at about 100°.

Considering the above, we have developed a knee simulator with the following advantages and innovative features. First, the simulator is driven by the muscles' forces and an active knee motion is made with bearing the upper body weight. As a result, the knee shows a 3D kinematics and generates the tibio-femoral contact forces. Under this condition, the TKA performance is to be assessed. Secondly, a hip joint mechanism is also incorporated into the simulator. The lower limb motion is achieved by the synergistic function between the hip and knee joints. Under this condition, a natural knee motion is to be reproduced. Thirdly, the simulator can make complete deep knee flexion up to 180°. Thus not only the conventional TKA but also a new TKA for high flexion can be attached to it for the evaluation.

Figure 1 shows the structure of the simulator, in which both the hip and knee joints are moved in a synergistic fashion by the pull forces of four wires. The four wires are pulled by the four servomotors respectively and reproduce the functions of the mono-articular muscles ((1), (3)) and the bi-articular muscles ((2), (4)) through the multiple pulley system. It should be noted that weight A and B are not heavy enough for the inverted double pendulum to stand up straight. They are applied as counter weights so that each segment duplicate the each segmental weight of the human lower limb. Figure 2 shows a sequential representation of stand to sit features: (a) at standing, (b) at high flexion, and (c) at deep flexion. At a state of 130° knee flexion between (b) and (c), hamstrings wire (4) becomes shortest and then exhibits an eccentric contraction, thereby attaining deep flexion.

Our knee simulator can be a useful tool for the evaluation of TKA performance and may potentially substitute the in vivo experiments.

The objective of this study is to determine the knee joint forces when rising from a kneeling position. We have developed a new type of knee prosthesis which is capable of attaining Japanese style sitting. To run the simulations and experiments needed to assess the performance of this prosthesis, it is necessary to know what forces act on the knee during deep flexion. Because these data are lacking, we created a 2D mathematical model of the lower leg to help determine knee joint forces during deep flexion. Healthy subjects of ten males (age of 25±4years, height of 170.3±9.1cm, and weight of 67.0±22.2kg) and five females (25±3years, 161±7.1cm, 47.7±6.2kg) participated in the experiment. Ground reaction force and joints angles were measured using a force plate and a motion recording system respectively. The collected data were entered into our mathematical model, and the muscle forces and the knee joint forces were calculated. To verify our model, we first used it to run simulation of middle and high flexions of the knee joint. In vivo data for these actions are available in the literature, and the results from our simulation were in good agreement with these data. We then collected the data and run simulation when rising from a kneeling position under the conditions shown in Fig. 1. They were a) double leg rising (both legs are aligned) without using the arms, b) ditto but using the arms, c) single leg rising (legs are in the front and the rear respectively) without using the arms, and d) ditto but using the arms. We obtained the following results. The statistics of the maximum values on the single knee joint for each condition were; a) F_max=5.1±0.4 [BW: (force on the knee joint)/(body weight)] at knee flexion angle of Q=140±8°, b) F_max=3.2±0.9[BW] at Q=90±10°, c) F_max-d=5.4±0.5[BW] at Q_d=62±20° for the dominant leg and F_max-s=3.0±0.5[BW] at Q_s=138±6° for the supporting leg respectively, and d) F_max-d=3.9±1.5[BW] at Q_d=70±17° for the dominant, and F_max-s=2.1±0.5 [BW] at Q_s=130±11° for the supporting. We may conclude that the single leg rising should be recommended since the maximum knee joint force did not become large as long as the knee was at deep flexion. The values introduced in this study could be used to assess the strength of the knee prosthesis at deep flexion. To obtain more realistic values of the joint forces, it is necessary to determine the ratio of the forces exerted by the mono-articular and the bi-articular joint muscles.

We have developed a new type of knee prosthesis which is capable to make 180° knee flexion, and have designated it as Complete flexion knee (CFK). Since the kinematics and kinetics of knee prosthesis vary depending not only on its articulating surface shapes but also on the stiffness of soft tissues, its performance should be assessed under various kinds of lower limb activities.

The objective of this study is to perform simulation analysis of various lower limb activities to evaluate the performance of CFK using the 2D and the 3D mathematical models. Kinematic analyses using X-ray picture or stress analyses using FEM are extensive however, kinematic analyses can not introduce stresses and FEM can not introduce kinematics. Mathematical model analyses can introduce vital information about kinematics and kinetics at the same time.

First, we carried out an in-vitro experiment using cadaver knee under the condition of passive knee flexion-extension. After that, we performed a simulation using the same parameter variables as the in-vitro experiment in order to assess the validity of our 2D and 3D models by comparing the results about the joint contact forces and kinematics with those from the experiment.

In the in-vitro experiment, the femoral bone of a cadaver knee was fixed on a jig. In order to secure the tibiofemoral contact, each muscle was pulled with constant force respectively. Then the tibia was carried through from 40° to 140° of knee flexion. The contact forces between the femur and the tibia were measured by a load sensor. During the process, fluoroscopic images were taken, and then 3D positions/orientations of the tibia relative to the femur were introduced from the images using the pattern matching method.

Our 2D and 3D models of total knee arthroplastic joint included the tibio-femoral and patello-femoral compartments, incorporating major muscles, patella tendon and primary ligaments. The patella tendon and primary ligaments were represented with non-linear springs, whose mechanical properties were determined from the literature. In our 2D model, “thigh and calf” contact was taken into account at deep knee flexion.

Using our 3D model, the simulation was performed up to 100° of knee flexion. After that we had to alternate the model from the 3D to the 2D because the patella stacked into the femoral intercondylar, the thigh-calf contact occurred and the 3D model did not introduce the converged solution.

Over all, both the experimental and simulation results were in good agreement with each other. The results from the simulation showed that the contact points were located unusually anteriorly. The post-cam contact occurred at 44° of knee flexion, indicating that the tibia was strongly pulled to the posterior. As for the contact resultant force, large differences between simulation and experiment were found. This may be because the soft tissues of the cadaver were not intact, while we determined their properties from the literature in the simulation.