Table 6 Summary of characteristics for inertial sensor-based gait quality measures
Algorithm | Potential advantages | Potential disadvantages |
|---|---|---|
HMM-SM | - Could allow for additional analysis of spatiotemporal gait features - Particularly effective at discriminating compensatory strategies employed by limb-difference (e.g., at the pelvis) | - Inconsistent performance depending on sensors used |
MDP | - Can form deviation curve throughout the gait cycle, as done in [22] - Similar performance using different sensors - Only need to iterate over control data once, when training the SOM | - Very similar performance as DTW with added algorithmic complexity |
DTW | - Closely approximates GPS performance - Similar performance using different sensors - Simple to calculate | - Computationally expensive for larger datasets |
INI | - Closely approximates GPS performance | - Requires accurate estimation of multiple kinematic and spatiotemporal parameters - No inclusion of symmetry parameters |