Table 5.
State of the art of human activity classification systems.
| Sensors | Placement | Features | Classifiers | Participants | Activities | Accuracy (%) | Reference |
|---|---|---|---|---|---|---|---|
| Accelerometer | Upper arm, lower arm, hip, thigh, foot | Time-domain; frequency-domain. | KNN; Decision tree; NB. | 20 | 6 | 52–84 | Bao. et al. [74] |
| Gyroscope | Shank | Frequency-domain | Other | 20 | 1 | 97 | Coley et al. [61] |
| Accelerometer; gyroscope. | Wrist; lower leg; foot; neck; hip. | Time-domain; frequency-domain. | Decision tree; KNN; NB. | 2 | 20 | NA 1 | Heinz et al. [13] |
| Accelerometer | Wrist; hip; necklace | Time-domain | C4.5 2 | 6 | 6 | About 75 | Muurer et al. [83] |
| Accelerometer | Lower back | Time-domain; frequency-domain. | Decision tree | 20 | 20 | 93 | Bonomi et al. [72] |
| Accelerometer; gyroscope. | Chest, thigh | Time-domain | User-defined | 1 | 4 | 91 | Li et al. [63] |
| Accelerometer | Wrist; lower arm; knee; ankle. | Time-domain; frequency-domain. | NB, GMM, SVM, NV, C4.5 | 20 | 20 | 92.2–98.5 | Mannini et al. [58] |
| Accelerometer | Thigh | Time-domain; frequency-domain. | C4.5, MLP 3, LR 4 | 29 | 6 | 78.1–95.7 | Kwapisz et al. [75] |
| Accelerometer; gyroscope | Arm, thigh. | Time-domain; frequency-domain. | KNN; C4.5; NB, etc. | 16 | 13 | 63.2–90.2 | Wu et al. [60] |
| Accelerometer; gyroscope | Lower arm | Time-domain | RBM 5 | 12 | NA | 72.1 | Bhattacharya et al. [92] |
| Accelerometer; gyroscope | Wrist | Time-domain; frequency-domain. | HF-SVM 6 | 30 | 6 | 89 | Anguita et al. [93] |
| Accelerometer | Lower back | Time-domain | Decision tree | 24 | 4 | 96.61 | Khan et al. [94] |
| Accelerometer; gyroscope; barometric pressure sensors | Ankle; wrist; chest | Time-domain; frequency-domain | KNN | 6 | 16 | 93–95 | Moncada-Torres et al. [68] |
| Accelerometer | Wrist | Time-domain; frequency-domain | NB; SVM; Decision tree; KNN | 2 | 8 | 57–64 | Ravi et al. [86] |
| Accelerometer | Chest; upper arm; wrist; hip; thigh; ankle; ear. | Time-domain; frequency-domain | KNN; Bayesian | 11 | 15 | NA | Atallah et al. [85] |
| Accelerometer. | Thigh | Time-domain | Shapelet approach; SVM; NB; KNN; etc. | 4 | 8 | 72–77 | Liu et al. [57] |
| Accelerometer; gyroscope | Wrist | Time-domain | ANN; SVM; NB; C4.5 | 30 | 6 | 76.63–95.75 | Ronao et al. [76] |
| Accelerometer; gyroscope. | Thigh | Time-domain | RF; SVM; NB. | NA | 4 | 77–99 | Bedogni et al. [95] |
| Accelerometer; Gyroscope. | Chest, thigh, ankle | Time-domain; frequency-domain. | Decision tree; KNN; SVM; HMM, etc. | 6 | 12 | 73–98 | Attal et al. [56] |
| Accelerometer | Wrist | Time-domain; | NB; Decision tree; SVM; C4.5; KNN | 10 | 6 | 63.09–99.56 | Morillo et al. [96] |
| Accelerometer | Wrist | Time-domain; frequency-domain. | NN; KNN. | 13 | 6 | 79.2–90.4 | Wang et al. [87] |
| Accelerometer | Hip | Time-domain; frequency-domain. | KNN; SVM. | 5 | 9 | 70–75.68 | Wang et al. [97] |
1 NA: Not Available; 2 C4.5: Decision trees; 3 MLP: Multilayer perception; 4 LR: Logistic Regression; 5 RBM: Restricted Boltzmann Machine; 6 HF-SVM: Hardware-Friendly SVM.