The comparison of transfemoral amputees using mechanical and microprocessor- controlled prosthetic knee under different walking speeds: A randomized cross-over trial

Abstract

BACKGROUND:

The microprocessor-controlled prosthetic knees have been introduced to transfemoral amputees due to advances in biomedical engineering. A body of scientific literature has shown that the microprocessor-controlled prosthetic knees improve the gait and functional abilities of persons with transfemoral amputation.

OBJECTIVE:

The aim of this study was to propose a new microprocessor-controlled prosthetic knee (MPK) and compare it with non-microprocessor-controlled prosthetic knees (NMPKs) under different walking speeds.

METHODS:

The microprocessor-controlled prosthetic knee (i-KNEE) with hydraulic damper was developed. The comfortable self-selected walking speeds of 12 subjects with i-KNEE and NMPK were obtained. The maximum swing flexion knee angle and gait symmetry were compared in i-KNEE and NMPK condition.

RESULTS:

The comfortable self-selected walking speeds of some subjects were higher with i-KNEE while some were not. There was no significant difference in comfortable self-selected walking speed between the i-KNEE and the NMPK condition ( $P=$ 0.138). The peak prosthetic knee flexion during swing in the i-KNEE condition was between sixty and seventy degree under any walking speed. In the NMPK condition, the maximum swing flexion knee angle changed significantly. And it increased with walking speed. There is no significant difference in knee kinematic symmetry when the subjects wear the i-KNEE or NMPK.

CONCLUSIONS:

The results of this study indicated that the new microprocessor-controlled prosthetic knee was suitable for transfemoral amputees. The maximum swing flexion knee angle under different walking speeds showed different properties in the NMPK and i-KNEE condition. The i-KNEE was more adaptive to speed changes. There was little difference of comfortable self-selected walking speed between i-KNEE and NMPK condition.

Keywords

Transfemoral amputees microprocessor-controlled prosthetic knees gait symmetry maximum swing flexion knee angle

1. Introduction

A prosthetic knee joint is part of a lower limb prosthesis. It is used by people who have lost a leg at or above the knee [1]. The loss of the lower limb is commonly the results of problems with the blood vessels in the leg or trauma. A prosthetic knee allowing for an active lifestyle is of high priority for amputees [2].

Prosthetic knees mainly include two categories according to control mechanisms. One is mechanical control of the knee joint and the other is microprocessor control to manage the swing and/or stance phases [3]. Mechanical mechanisms include knee joints with single axis, constant-friction and weight activated control; pneumatic or hydraulic knee systems with fluid swing phase control and variable methods of stance stability; and polycentric knee components with pneumatic damper. However, all of these prosthetic knees cannot adapt to different surface conditions and changes in walking speed [4]. Microprocessor-controlled prosthetic knees (MPKs) are a category of prosthetic knee components, becoming more widely prescribed in the last 15–20 years. In contrast to non-microprocessor-controlled prosthetic knees (NMPKs), the microprocessor-controlled prosthetic knees have the ability to react to changes in walking speeds and allow prosthetic knee flexion during early stance and ideal kinematics during swing [5].

The commercially available prosthetic knees interpret angle, gyroscope and load sensors with an on board microprocessor to adapt dynamically to gait change. Typical microprocessor controlled knees include Genium (Otto Bock Healthcare, Duderstadt, Germany) [6], C-Leg (Otto Bock Healthcare, Duderstadt, Germany) [7], the Rheo Knee (Ossur. Reykjavik, Iceland) [8], the SmartIP (Chas A. Blatchford and Sons. Hampshire, United Kingdom) [9] and the POWER KNEE (Ossur. Reykjavik, Iceland) [10]. The aforementioned knee prostheses only provide damping torque during the gait except the POWER KNEE prosthesis.

A few studies have been conducted to quantitatively compare NMPKs with MPKs using measures of gait metabolism, kinetics, and kinematics. Orendurff et al. compared the Mauch SNS knee and the C-Leg microprocessor-controlled knee in eight TF amputees. They noted that participants used the same amount of energy while the self-selected walking speeds were higher when using the C-Leg microprocessor-controlled prosthesis. The increase of energy efficiency with the C-Leg was verified [11]. Seymour et al. found that use of the C-leg resulted in a statistically significant decrease in the number of steps and time to complete the obstacle course [12]. In a published study by Kaufman et al., changes in gait and balance when using the MPKs were researched. The transition from a hyperextended knee to a flexed knee resulted in a change from an internal knee flexor moment to a knee extensor moment during loading response [13]. To prevent the NMPK from buckling, subjects wearing a NMPK had to actively contract the hip extensors to pull back and force the prosthetic knee into extension. In contrast, the microprocessor-controlled knee allowed subjects to place more demand on the knee, as measured by an increase in the internal knee extension movement.

The aim of this study was to provide a new MPK (named i-KNEE) and compare walking with i-KNEE to NMPKs under different walking speeds. Firstly, we proposed a MPK (i-KNEE) with hydraulic damper. Secondly, we hypothesized the comfortable self-selected walking speed would be increased while walking with the i-KNEE. In addition, we hypothesized the maximum swing flexion knee angle would be held in little difference under all walking velocities when walking with the i-KNEE, while peak prosthetic knee flexion during swing would increase with walking speed in the NMPK condition. Finally, we compared the gait symmetry of transfemoral amputees while using a passive mechanical knee joint (NMPK) or the i-KNEE. We hypothesized that the patient would have improved gait symmetry when wearing the i-KNEE compared to the NMPK.

2. Methods

2.1 Development of the new microprocessor-controlled prosthetic knee

The knee joint acts as a damper most of the time in the normal walking. It just provides power during special situations such as climbing the stairs [14]. Therefore, the developed prosthetic knee is aimed to provide damping during different phases. To realize the continuity and independence of the damping adjustment during knee flexion and extension, the new MPK (i-KNEE) employs a hydraulic cylinder with two servo controlled valves that adjust both flexion and extension resistances dynamically during each stride, the prototype is shown in Fig. 1.

Figure 1.

The prototype of i-KNEE.

Each valve is controlled by the microprocessor and the valve can adjust the desired impedance for different gait phase at different speeds. The valve positions allow the fluid passages to be completely blocked, flexion blocked with adjustable extension impedance, flexion impeded with free extension, free flexion and extension, free flexion with impeded extension, or impeded flexion with blocked extension. Thus, the valve can adjust the desired damping for heel strike, stance flexion, swing flexion, and swing extension for different activities. There are also many sensors (knee angle sensor, axil load sensor, ankle pressure sensor and tri-axial accelerometer) integrated into the prototype. The real-time data of the knee joint angle, acceleration, and load force is collected to determine the amputee gait. The detailed description is shown in [15].

2.2 Subjects

We recruited persons with above-knee amputation from Shanghai for this randomized cross-over trial. Twelve people with unilateral transfemoral amputation participated in the evaluations. All of them had the characteristics that: 1) at least two year post amputation; 2) mobility grade of Medicare Functional Classification Level (MFCL) from K3 or higher; 3) never previously fitted with a microprocessor-controlled knee. They would not have other musculoskeletal problems that influence walking ability. The distance from knee center to floor would not be below 42 cm. Subject demographics are described in Table 1.

Table 1
Subject demographics

Subject	Age (years)	Height (cm)	Weight (kg)	Gender	NMPK
1	46	173	65	Male	3R80
2	23	174	68	Male	3R80
3	49	170	63	Male	CaTech
4	38	160	53	Female	3R60
5	32	178	77	Male	Mauch SNS
6	45	172	65	Male	3R80
7	47	163	57	Female	Mauch SNS
8	52	170	67	Male	3R80
9	50	172	65	Male	Mauch SNS
10	44	175	73	Male	CaTech
11	53	180	80	Male	3R80
12	27	160	50	Female	3R60

This study was approved by the authors’ Institutional Review Board. All subjects signed an informed consent prior to participation.

2.3 Motion analysis technology

The measurements were performed in the laboratory with the real-time three-dimensional gait and motion analysis system (JIANGSU NEUCOGNIC MEDICAL CO., LTD.). The 3D gait and motion analysis system is a three-dimensional clinical gait analysis device, which integrates measurement, evaluation and database management. It includes 17 accelerometers to capture the major joint movements of the body. For the real-time 3D gait and motion analysis system, it can output the maximum and minimum angles of the specific joint. Unlike VICON or other systems based on the camera, it does not need markers and can be used for any flat road condition. Ground reaction forces are measured using zebris FDM system (zebris Medical Gmbh, German) which is accurate and user-friendly for pressure analysis. Zebris FDM system is a treadmill system with 5300 built-in pre-calibrated capacitive sensors for the analysis of force and pressure distribution during standing, walking and running, which can be completed in less than 30 seconds.

2.4 Prosthetic adjustments

We assigned the subjects to start trials randomly with their own NMPK or with the i-KNEE. The Triton (Ottobock) prosthetic foot was provided for all prosthetic knee conditions. The socket kept to be same across the trials. As was done in a typical clinical setting, we used the expert opinions of a prosthetist and physiatrist to determine the optimal prescription for the NMPK and i-KNEE to maintain the study’s clinical relevance. Alignment of the prosthesis affects comfort, function, and cosmetics. The poor socket fit will be caused under improper alignment. This will result in undesirable pressure distribution at the residual limb/socket interface which will cause pain, discomfort, and potentially tissue damage [16]. The prosthetic knee joints were investigated with an identical prosthetic alignment. Alignment was quantified using the Otto Bock Laser Assisted Static Alignment Reference (LASAR) system. The forces and moments acting on the prosthesis were comparable because of the prosthetic alignment was not changed. The first set of trials was performed with their NMPKs. Because all the subjects were wearing NMPKs previously, they needed some time to adapt to the use of the i-KNEE. The subjects were sent home with the new prosthesis and returned for the test every day. After six weeks of acclimatization, the second set of measurements was performed with i-KNEE. There was no gait training program while walking with the i-KNEE or their own NMPK. This ensured that comparison was as little affected by gait training factors as possible.

2.5 Testing

Testing had two main goals. One was to determine comfortable self-selected walking speed while wearing i-KNEE and NMPK for the same person. The higher the preferred walking speed was, the better the walking ability with the prosthetic knee. The other was to compare the maximum swing flexion knee angle for subjects with i-KNEE and NMPKs under specific walking speeds. The slow speed, middle speed and fast speed were chosen to be the typical speeds. All subjects completed the walking tasks on the treadmill in the same order: 1) slow speed (0.7 m/s), 2) middle speed (1 m/s), 3) fast speed (1.4 m/s), 4) comfortable self-selected walking speed. For the three typical speeds, every subject was asked to walk five minutes for each speed on the treadmill. The comfortable self-selected walking speed was determined during a familiarization trial. In the trial, treadmill speed was gradually increased until participants indicated that the speed was comfortable. For each familiarization trial speed, the subject was given two minutes to feel. When the participant felt comfortable, the treadmill speed was increased with 0.2 m/s and the participant was asked whether it was more comfortable or uncomfortable. If the treadmill speed was uncomfortable, the earlier treadmill speed was used. If the higher treadmill speed was more comfortable, the treadmill speed was further increased 0.2 m/s until it became uncomfortable.

Table 2
The statistical comparison of comfortable self-selected walking speed between NMPK and i-KNEE

	Results		Statistical comparison
	NMPK	i-KNEE	P-value
Comfortable self-selected walking speed (m/s)	0.63 [0.57–0.82]	0.63 [0.60–0.84]	0.138

Figure 2.

Comfortable self-selected walking speed with i-KNEE and NMPK.

3. Results

3.1 Comfortable self-selected walking speed

Results of comfortable self-selected walking speed are shown in Fig. 2. The results are not as expected. Not all subjects with i-KNEE prefer higher walking speed than NMPK. The statistical comparison of comfortable self-selected walking speed between NMPK and i-KNEE is shown in Table 2. In the data analysis process, T-test is used to find out whether the differences of the mean scores are significant. There is no significant difference in self-selected walking speed between the i-KNEE and the NMPK condition ( $P=$ 0.138). The comfortable self-selected walking speeds are lower than non-amputees.

Figure 3.

0.7 m/s with i-KNEE and NMPKs.

Figure 4.

1 m/s with i-KNEE and NMPKs.

3.2 Maximum swing flexion knee angle

The results of the comparison of the NMPK and i-KNEE condition within different walking speeds are shown in Figs 3–5. The data represents the maximal observed knee angle.

Table 3
The statistical comparison of maximum swing flexion knee angle under different walking speeds between NMPK and i-KNEE

Speed (m/s)	Maximum swing flexion knee angle (degrees)		P-value
	NMPK	i-KNEE
0.7	54 [50–62]	62 [57–66]	0.022
1	60 [55–67]	63 [58–67]	0.041
1.4	75 [71–80]	63 [60–68]	0.015

Figure 5.

1.4 m/s with i-KNEE and NMPKs.

The statistical comparison of maximum swing flexion knee angle under different walking speeds between NMPK and i-KNEE is shown in Table 3. When looking at the influence of walking speed, we find that peak prosthetic knee flexion during swing in the i-KNEE condition is between sixty and seventy degree under any walking speeds. The range is in accordance with the physiological gait. In the NMPK condition, the maximum swing flexion changes significantly (0.7 m/s: $P=$ 0.022; 1 m/s: $P=$ 0.041; 1.4 m/s: $P=$ 0.015). And it increases with walking speed.

There are differences in peak prosthetic knee flexion during swing while walking with the i-KNEE and the NMPK within the same walking speed (0.7 m/s: $P=$ 0.022; 1 m/s: $P=$ 0.041; 1.4 m/s: $P=$ 0.015). When the walking speed is 1.4 m/s, the difference is maximum. When the walking speed is 1 m/s, the difference is minimum.

3.3 Kinematic symmetry

The symmetry index compared the kinematics of the healthy leg and prosthetic leg for prosthesis used. The symmetry index was calculated of the whole gait cycle for each subject. The symmetry index utilized the method proposed by Kaufman et al. [17]. The complete symmetry between the two waveforms was indicated by a final value of $+$ 1, while a value of $-$ 1 indicated complete asymmetry. A value of 0 indicated that the waveform shapes were unrelated. For each trial, the symmetry index of selected variables between the prosthetic leg and healthy leg was calculated for the kinematics at the knee joint in the sagittal plane.

Figure 6.

Symmetry index for kinematics in the sagittal plane for knee joint and two different prostheses.

Figure 7.

Sagittal plane kinematics of the healthy leg (blue line) versus the prosthetic leg (pink line) of a representative subject for the knee joint.

The kinematic symmetry of i-KNEE and NMPK during stance and swing phases is shown in Fig. 6. At the knee joint, there is a significant difference between stance and swing phase gait symmetry. In most cases, the difference in the knee position during loading response leads to the stance phase kinematic asymmetry. In the swing phase, most of the subjects have symmetrical kinematics. There is no significant difference in knee kinematic symmetry when the subjects wear the i-KNEE or NMPK. Data from a representative subject will more fully explain the gait difference between the i-KNEE and a NMPK (Fig. 7). The subject wears a NMPK in the first row and the i-KNEE in the second row. Graph is plotted as percentage of gait cycle, where 0% is heel strike and 100% is subsequent heel strike. There is no knee flexion during loading response when wearing the NMPK. However, there is a knee flexion during loading response when wearing the i-KNEE. This leads to the higher symmetry index of the i-KNEE during stance phase.

4. Discussion

The aim of this study was to propose a new microprocessor-controlled prosthetic knee and compare the use of NMPKs to the use of the microprocessor-controlled prosthetic knee (i-KNEE) under different walking speeds. The comfortable self-selected walking speed is similar between the i-KNEE and NMPK. The comparison of the NMPK and i-KNEE showed significant differences in maximum swing flexion knee angle under three evaluated walking speeds.

The new microprocessor-controlled prosthetic knee was still a passive device. It could not provide active power when climbing stairs. Many researches had been focused on the design and estimation of active prosthetic knee. Villalpando et al. presented the design and evaluation of a novel biomimetic active knee prosthesis capable of emulating intact knee biomechanics during level-ground walking [18]. Sup et al. presented an electrically powered knee and ankle prosthesis. The prosthesis design incorporated two motor-driven ball screw units to drive the knee and ankle joints. A spring in parallel with the ankle motor unit was employed to decrease the power consumption and increase the torque output for a given motor size [19]. However, the passive devices cannot restore a natural gait, and the active device is large and inefficient. Therefore, the semi-active prosthetic knee is the important trend. Lambrecht et al. developed a semi-active prosthetic knee that combined the proven safety and power efficiency of a passively damped hydraulic device with the improved gait and advanced mobility of a fully active device [20]. Awad et al. developed an electrical above knee prosthesis, which works as a passive knee prosthesis in part of gait cycle phases and as an active knee prosthesis during other portions [21].

We hypothesized that the self-selected walking speed would be higher while walking with i-KNEE when compared to the NMPK. Our results disproved the hypothesis: we found a fusion difference in self-selected walking speed between the i-KNEE and NMPK condition. Not all subjects would chose a higher comfortable self-selected walking speed. The comfortable self-selected walking speed between i-KNEE and NMPK condition was in a close level ( $P=$ 0.138).

The result of comfortable self-selected walking speed was different with some previous researches. Segal et al. compared the gait biomechanics of transfemoral amputees wearing the microprocessor-controlled prosthetic knee C-Leg with those wearing a common non-computerized prosthesis Mauch SNS and found that subjects chose to walk at a faster speed when wearing the C-Leg versus the Mauch SNS [22]. Kahle et al. compared subjects’ performance with a non-microprocessor knee mechanism (NMKM) versus a C-Leg on nine clinically repeatable evaluative measures and found that self-selected walking speed on even terrain improved 15% [23]. Orendurff et al. found subjects chose higher self-selected walking speed with the C-Leg compared with the Mauch SNS but did not incur higher oxygen costs, which suggested greater efficiency only at their self-selected walking speed [11]. Eberly et al. compared level walking function while wearing a microprocessor-controlled knee (C-Leg Compact) prosthesis to a traditionally prescribed non-microprocessor-controlled knee prosthesis for Medicare Functional Classification Level K-2 walkers and found that the C-Leg Compact produced significant increase in velocity, cadence, stride length, single-limb support, and heel-rise timing compared to walking with the non-microprocessor-controlled knee prosthesis [24].

However, some previous studies also supported the result of comfortable self-selected walking speed. Johansson compared two variable-damping knees, the hydraulic-based Otto Bock C-leg and the magnetorheological-based Össur Rheo, with the mechanically passive, hydraulic-based Mauch SNS and found that the comfortable self-selected walking speeds across the three prostheses were not statistically different [25]. Datta et al. compared the gait of amputees wearing conventionally damped pneumatic swing-phase control knees and microchip-controlled Intelligent Prostheses and found that there were no significant differences in subjective gait evaluation or temporal and spatial gait parameters [26]. Prinsena et al. compared the Rheo Knee II (a microprocessor-controlled prosthetic knee) with NMPKs across varying walking speeds and did not find a difference in preferred walking speed between the Rheo Knee II and NMPK condition [27].

The reason of this ambiguity in comfortable self-selected walking speed is unclear. It may be caused by difference in type of prosthetic knee, time of acclimatization and whether the training process to be used to the microprocessor-controlled prosthetic knee is done.

We also hypothesized the maximum swing flexion knee angle would be held in little difference under all walking velocities when walking with the i-KNEE, while maximum swing flexion knee angle would increase with walking speed in the NMPK condition. We observed that peak prosthetic knee flexion during swing significantly increased with walking speed in NMPK condition. The peak prosthetic knee flexion during swing was always between sixty and seventy degree in spite of the change of walking speed. Maximum swing flexion knee angle typically does not exceed 70 degrees in normal walking [5]. It demonstrated that the proposed microprocessor-controlled prosthetic knee (i-KNEE) was able to realize physiological gait under different speeds. Therefore, we were able to confirm our hypothesis in our study population. The range was in line with some previous studies. Herr and Wilkenfeld found that maximum swing flexion knee angle while walking with user-adaptive prosthetic knee remained around 70 degree irrespective of walking speed. Contrastingly, peak knee flexion angle while walking with the NMPK increased with walking speed [28]. Li proposed a prototype structure of the intelligent prosthetic knee and found that the best target angle of maximum swing flexion for Asian people was between sixty and seventy degree [29]. However, Prinsena et al. found that there was no differences on peak prosthetic knee flexion during swing between Rheo Knee II and NMPK condition. The maximum swing flexion knee angle increased significantly with walking speed for both microprocessor-controlled prosthetic knee and NMPK condition [27].

We finally hypothesized the patient would have improved gait symmetry when wearing the i-KNEE compared to the NMPK. Our results indicated that there was no significant difference in kinematic symmetry when using the i-KNEE and NMPK. Similar findings have been reported by Kaufman et al. who indicated amputees could not improve gait kinematic symmetry when using a C-Leg [17]. The difference of ground reaction force, duration of different gait phases while walking with i-KNEE and NMPKs are also important for the quality of gait with the new prosthesis. The comparison of supporting time and force of prosthetic side and healthy side will be studied in the future research.

5. Conclusions

The work proposed a new microprocessor-controlled prosthetic knee (i-KNEE) and compared the use of NMPKs to the use of the microprocessor-controlled prosthetic knee (i-KNEE) under different walking speeds. The results of this study indicated that there was little difference of comfortable self-selected walking speed between i-KNEE and NMPK. The maximum swing flexion knee angle under different walking speeds showed different properties in the NMPK and i-KNEE condition. The maximum swing flexion knee angle increased with walking speeds when walking with NMPK. The maximum swing flexion knee angle was held in little difference under different walking velocities when wearing i-KNEE. There is a significant difference between stance and swing phase gait symmetry. However, there is no significant difference in kinematic symmetry when using the i-KNEE and NMPK.

Footnotes

Acknowledgments

The project described was partially supported by grant number 61473193 from National Natural Science Foundation of China. Partial funding came from grants 15DZ2251700 from the Shanghai Engineering Research Center of Assistive Devices.

Conflict of interest

The authors have no conflict of interest to report.

References

Geeroms

Flynn

Jimenez Fabian

R E

, et al. Design and energetic evaluation of a prosthetic knee joint actuator with a lockable parallel spring[J]. Bioinspiration & Biomimetics, 2017, 12(2).

Choi

I H

. Postinfectious Deformities of the Lower Limb[M]// Pediatric Lower Limb Deformities. Springer International Publishing, 2016.

Thiele

Westebbe

Bellmann

, et al. Designs and performance of microprocessor-controlled knee joints[J]. Biomedizinische Technik Biomedical Engineering, 2014, 59(1): 65-77.

Furse

Cleghorn

Andrysek

. Development of a low-technology prosthetic swing-phase mechanism[J]. Journal of Medical & Biological Engineering, 2011, 31(2): 145-150.

Cao

Zhao

, et al. Target of physiological gait: Realization of speed adaptive control for a prosthetic knee during swing flexion[J]. Technology & Health Care Official Journal of the European Society for Engineering & Medicine, 2017(1): 1-12.

Bellmann

Schmalz

Ludwigs

, et al. Stair ascent with an innovative microprocessor-controlled exoprosthetic knee joint[J]. Biomedical Engineering-biomedizinische Technik, 2012, 57(6): 435-444.

Highsmith

M J

Kahle

J T

Bongiorni

D R

, et al. Safety, energy efficiency, and cost efficacy of the C-Leg for transfemoral amputees: A review of the literature[J]. Prosthetics & Orthotics International, 2010, 34(4): 362.

Edwards

T A

. Adjusting and Logging Rheo Knee Parameter Values Using a Smartphone Application[J]. Journal of Magnetic Resonance, 2015, 123(2): 226-229.

Chin

D T

Machida

Sawamura

, et al. Comparison of different microprocessor controlled knee joints on the energy consumption during walking in trans-femoral amputees: Intelligent Knee Prosthesis (IP) versus C-Leg[J]. Prosthetics & Orthotics International, 2006, 30(1): 73.

10.

Pasquina, Paul F, Carvalho, et al. Case Series of Wounded Warriors Receiving Initial Fit PowerKneeTM Prosthesis[J]. Jpo Journal of Prosthetics & Orthotics, 2017, 29(2): 1.

11.

Orendurff

M S

Segal

A D

Klute

G K

, et al. Gait efficiency using the C-Leg[J]. Journal of Rehabilitation Research & Development, 2006, 43(2): 239.

12.

Seymour

Engbretson

Kott

, et al. Comparison between the C-leg microprocessor-controlled prosthetic knee and non-microprocessor control prosthetic knees: a preliminary study of energy expenditure, obstacle course performance, and quality of life survey[J]. Prosthetics & Orthotics International, 2007, 31(1): 51.

13.

Kaufman

K R

Levine

J A

Brey

R H

, et al. Gait and balance of transfemoral amputees using passive mechanical and microprocessor-controlled prosthetic knees[J]. Gait & Posture, 2007, 26(4): 489-493.

14.

Zarrugh

M Y

. Kinematic prediction of intersegment loads and power at the joints of the leg in walking[J]. Journal of Biomechanics, 1981, 14(10): 713-725.

15.

Cao

Zhao

16.

Wei

Nothnagle

Savant

, et al. Pneumatic actuator inserts for interface pressure mapping and fit improvement in lower extremity prosthetics[C]// IEEE International Conference on Biomedical Robotics and Biomechatronics. IEEE, 2016: pp. 574-579.

17.

Kaufman

K R

Frittoli

Frigo

C A

. Gait asymmetry of transfemoral amputees using mechanical and microprocessor-controlled prosthetic knees[J]. Clinical Biomechanics, 2012, 27(5): 460-465.

18.

Villalpando

E C M

. Design and evaluation of a biomimetic agonist-antagonist active knee prosthesis[J]. Massachusetts Institute of Technology, 2012.

19.

Sup

Varol

H A

Goldfarb

. Upslope Walking With a Powered Knee and Ankle Prosthesis: Initial Results With an Amputee Subject[J]. IEEE Transactions on Neural Systems & Rehabilitation Engineering A Publication of the IEEE Engineering in Medicine & Biology Society, 2011, 19(1): 71.

20.

Lambrecht

B G A

Kazerooni

. Design of a semi-active knee prosthesis[C]// IEEE International Conference on Robotics and Automation. IEEE, 2009: pp. 639-645.

21.

Awad

Tee

K S

Dehghani

, et al. DESIGN OF AN EFFICIENT BACK-DRIVABLE SEMI-ACTIVE ABOVE KNEE PROSTHESIS[C]// Field Robotics-, International Conference on Climbing and Walking Robots and the Support Technologies for Mobile Machines. 2015: pp. 35-42.

22.

Segal

A D

Orendurff

M S

Klute

G K

, et al. Kinematic and kinetic comparisons of transfemoral amputee gait using C-Leg and Mauch SNS prosthetic knees[J]. Journal of Rehabilitation Research & Development, 2006, 43(7): 857.

23.

Kahle

J T

Highsmith

M J

Hubbard

S L

. Comparison of nonmicroprocessor knee mechanism versus C-Leg on Prosthesis Evaluation Questionnaire, stumbles, falls, walking tests, stair descent, and knee preference[J]. Journal of Rehabilitation Research & Development, 2008, 45(1): 1.

24.

Eberly

V J

Mulroy

S J

Gronley

J K

, et al. Impact of a stance phase microprocessor-controlled knee prosthesis on level walking in lower functioning individuals with a transfemoral amputation[J]. Prosthetics & Orthotics International, 2014, 38(6): 447-55.

25.

Johansson

J L

Sherrill

D M

Riley

P O

, et al. A clinical comparison of variable-damping and mechanically passive prosthetic knee devices[J]. American Journal of Physical Medicine & Rehabilitation, 2005, 84(8): 563-575.

26.

Datta

Heller

Howitt

. A comparative evaluation of oxygen consumption and gait pattern in amputees using Intelligent Prostheses and conventionally damped knee swing-phase control[J]. Clinical Rehabilitation, 2005, 19(4): 398-403.

27.

Prinsen

E C

Nederhand

M J

Sveinsdóttir

H S

, et al. The influence of a user-adaptive prosthetic knee across varying walking speeds: A randomized cross-over trial[J]. Gait & Posture, 2017, 51: 254-260.

28.

Herr

Wilkenfeld

. User-adaptive control of a magnetorheological prosthetic knee[J]. Industrial Robot, 2003, 30(1): 42-55.

29.

. Working mechanism research and prototype design of intelligent prosthetic knee joint[D]. Southeast University, 2012.