Hidden Markov model-based similarity measure (HMM-SM) for gait quality assessment of lower-limb prosthetic users using inertial sensor signals

Table 5 Comparison of prosthetic levels

Algorithm	TT–TF	TT–VN	TT–LS	TF–VN	TF–LS	VN–LS	Differences per configuration
GPS	0.360	0.008	0.065	0.006	0.062	0.187	2
HMM-SM upper	0.067	0.747	0.441	0.093	0.068	0.692	0
HMM-SM lower	0.235	0.472	0.025	0.529	0.060	0.046	2
HMM-SM pelvis	0.496	0.942	0.012	0.712	0.022	0.027	3
MDP upper	0.780	0.316	0.465	0.292	0.409	0.923	0
MDP lower	0.611	0.098	0.327	0.040	0.263	0.837	1
MDP pelvis	0.875	0.233	0.200	0.349	0.214	0.422	0
DTW upper	0.736	0.365	0.347	0.339	0.305	0.484	0
DTW lower	0.447	0.016	0.192	0.003	0.153	0.743	2
DTW pelvis	0.908	0.246	0.114	0.383	0.143	0.397	0
INI	0.014	0.229	0.010	0.059	0.006	0.565	3
MGS	0.569	0.060	0.783	0.006	0.440	0.246	1
Differences per Subgroup	1	2	3	4	2	2

Results from post hoc Welch’s t-test comparing the different prosthetic levels. p-values for configurations which were statistically significant are bolded and underlined. X–Y indicates the two levels being compared. Differences per Configuration indicates the number of differences identified by each algorithm/sensor configuration (count of significant results in each row). Differences per Subgroup indicates number of algorithms that showed significant difference for each of the subgroup comparisons (count of significant results in each column)
TT Transtibial, TF Transfemoral, VN Van Nes, LS Limb Shortening

ISSN: 1743-0003