This paper aims to enhance the accuracy of human gait prediction using machine learning algorithms. Three classifiers are used in this paper: XGBoost, Random Forest, and SVM. A predefined dataset is used for feature extraction and classification. Gait prediction is determined during several locomotion activities: sitting (S), level walking (LW), ramp ascend (RA), ramp descend (RD), stair ascend (SA), stair descend (SD), and standing (ST). The results are gained for steady-state (SS) and overall (full) gait cycle. Two sets of sensors are used. The first set uses inertial measurement units only. The second set uses inertial measurement units, electromyography, and electro-goniometers. The comparison is based on prediction accuracy and prediction time. In addition, a comparison between the prediction times of XGBoost with CPU and GPU is introduced due to the easiness of using XGBoost with GPU. The results of this paper can help to choose a classifier for gait prediction that can obtain acceptable accuracy with fewer types of sensors.
DesmondDMMacLachlanM. Affective distress and amputation-related pain among older men with long-term, traumatic limb amputations. J Pain Symp Manage2006; 31(4): 362–368.
2.
LuiZWAwadMIAbouhosseinA, et al. Virtual prototyping of a semi-active transfemoral prosthetic leg. Proc IMechE, Part H: J Engineering in Medicine2015; 229(5): 350–361.
3.
HuoWMohammedSMorenoJC, et al. Lower limb wearable robots for assistance and rehabilitation: a state of the art. IEEE Syst J2014; 10(3): 1–14.
4.
WindrichMGrimmerMChristO, et al. Active lower limb prosthetics: a systematic review of design issues and solutions. Biomed Eng Online2016; 15(3): 5–19.
5.
UnalRCarloniRHekmanEEG, et al. Conceptual design of an energy efficient transfemoral prosthesis. In: 2010 IEEE/RSJ International conference intelligent robots and systems, Taipei, Taiwan, 18–22 October 2010, pp.343–348. IEEE.
6.
AwadMIAbouhosseinADehghani-SanijAA, et al. Towards a smart semi-active prosthetic leg: preliminary assessment and testing. IFAC-PapersOnLine2016; 49(21): 170–176.
7.
El-SayedAMHamzaidNAAbu OsmanN A. Technology efficacy in active prosthetic knees for transfemoral amputees: a quantitative evaluation. Sci World J2014; 2014: 297431.
8.
EkkachaiKNilkhamhangI. Swing phase control of semi-active prosthetic knee using neural network predictive control with particle swarm optimization. IEEE Trans Neural Syst Rehabil Eng2016; 24(11): 1169–1178.
9.
GaileyRSRoachKEApplegateEB, et al. The amputee mobility predictor: an instrument to assess determinants of the lower-limb amputee’s ability to ambulate. Arch Phys Med Rehabil2002; 83(5): 613–627.
10.
OrendurffMSRaschkeSUWinderL, et al. Functional level assessment of individuals with transtibial limb loss: evaluation in the clinical setting versus objective community ambulatory activity. J Rehabil Assist Technol Eng2016; 3: 205566831663631.
11.
VuHTTDongDCaoHL, et al. A review of gait phase detection algorithms for lower limb prostheses. Sensors (Basel)2020; 20(14): 3972.
12.
KrauszNEHuBHHargroveLJ. Subject and environment-based sensor variability for wearable lower-limb assistive devices. Sensors (Switzerland)2019; 19(22): 1–18.
13.
GuptaAJadhavAJadhavS, et al. Human gait analysis based on decision tree, random forest and KNN algorithms. Adv Intell Syst Comput2020; 1155: 283.
14.
NgYLJiangXZhangY, et al. Automated activity recognition with gait positions using machine learning algorithms. Eng Technol Amp Appl Sci Res2019; 9(4): 4554–4560.
15.
BehrJFriedlyJMoltonI, et al. Pain and pain-related interference in adults with lower-limb amputa- tion: comparison of knee-disarticulation, transtibial, and transfemoral surgical sites. J Rehabil Res Develop2009; 46(7): 963–972.
16.
Martinez-HernandezUDehghani-SanijAA. Adaptive Bayesian inference system for recognition of walking activities and prediction of gait events using wearable sensors. Neural Netw2018; 102: 107–119.
17.
WoodwardRBSpaniasJAHargroveLJ. User intent prediction with a scaled conjugate gradient trained artificial neural network for lower limb amputees using a powered prosthesis. In: Proceedings of the annual international conference of the IEEE engineering in medicine and biology society, EMBS, Oralando, FL, 16–20 August 2016, pp.6405–6408. IEEE.
18.
Martinez-HernandezURubio-SolisADehghani-SanijAA. Recognition of walking activity and prediction of gait periods with a CNN and first-order MC strategy. In: Proceedings of the IEEE RAS and EMBS international conference on biomedical robotics and biomechatronics, Enschede, Netherlands, 26–29 August 2018, pp.897–902.
19.
HuangHZhangFHargroveLJ, et al. Continuous locomotion-mode identification for prosthetic legs based on neuromuscular – mechanical fusion. IEEE Trans Biomed Eng2011; 58(10 PART 1): 2867–2875.
20.
WuYKrishnanS. Statistical analysis of gait rhythm in patients with parkinson’s disease. IEEE Trans Neural Syst Rehabil Eng2010; 18(2): 150–158.
21.
DuLShaoW. Recognizing gait using Haar wavelet and support vector machine. Adv Intell Soft Comput2012; 158: 221–227.
22.
SiWYangGChenXG, et al. Gait identification using fractal analysis and support vector machine. Soft Comput2019; 23(19): 9287–9297.
23.
NgHTongHLTanWH, et al. Gait classification by support vector machine. Commun Comput Inform Sci2011; 179: 623–636.
24.
ShiLFQiuCXXinDJ, et al. Gait recognition via random forests based on wearable inertial measurement unit. J Ambient Intellig Human Comput2020; 11: 5329–5340.
25.
AçiciKErdaşÇBAşuroğluT, et al. A random forest method to detect Parkinson’s disease via gait analysis. In:BoracchiGIliadisLJayneC, et al. (eds.), Engineering applications of neural networks. Cham: Springer International Publishing, 2017, pp.609–619.
26.
LuoGZhuYWangR, et al. Random forest-based classification and analysis of hemiplegia gait using low-cost depth cameras. Med Biol Eng Comput2020; 58(2): 373–382.
27.
ChenTCarlosG. XGBoost: a scalable tree boosting system. In: KDD ‘16 proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining. San Francisco, California, USA. Association for Computing Machinery. https://arxiv.org/abs/1603.02754 (2016, accessed 21 May 2019).
28.
WangCChanPPKLamBMF, et al. Real-time estimation of knee adduction moment for gait retraining in patients with knee osteoarthritis. IEEE Trans Neural Syst Rehabil Eng2020; 28(4): 888–894.
29.
WuYQiSHuF, et al. Recognizing activities of the elderly using wearable sensors: a comparison of ensemble algorithms based on boosting. Sensor Rev2019; 39(6): 743–751.
30.
HuBRouseEHargroveL. Benchmark datasets for bilateral lower-limb neuromechanical signals from wearable sensors during unassisted locomotion in able-bodied individuals. Front Robot AI2018; 5: 14.
31.
HuBRouseEHargroveL. Fusion of bilateral lower-limb neuromechanical signals improves prediction of locomotor activities. Front Robot AI2018; 5: 1–16.
32.
ZhangKWangJde SilvaCW, et al. Unsupervised cross-subject adaptation for predicting human locomotion intent. IEEE Trans Neural Syst Rehabil Eng2020; 28(3): 646–657.
33.
HalimAIbrahimAAwadMI, et al. Optimization of sensor number for lower limb prosthetics using genetic algorithm. In: 2020 International conference on innovative trends in communication and computer engineering (ITCE), Aswan, Egypt, 2020, pp.210–215. IEEE.
