Fig 1.
The structure of grasp force estimation.
Fig 2.
sEMG armband and its usage diagram.
Fig 3.
sEMG signal and its spectrum.
Fig 4.
Grasp force sensor and usage diagram.
(a) Grasp force sensor; (b) Usage diagram of the grasp force sensor.
Table 1.
Parameters of the grasp force sensor.
Fig 5.
Experimental scene and distribution of measurement positions.
(a) Experimental scene; (b) Measurement positions of sEMG signals.
Fig 6.
Distribution of the maximum grasp force for each action at each strength level for a subject.
Fig 7.
Six channels of sEMG and grasp force.
Table 2.
sEMG signal measurement positions.
Fig 8.
The segmented sEMG signal.
Fig 9.
The structure of GRNN.
Fig 10.
Grasp force estimation results when different FSs and MPSs are used.
(a) The NRMS of the estimation error; (b) The NMAE of the estimation error; (c) The CC between actual grasp force and estimated grasp force.
Fig 11.
Results estimated when different numbers of measurement positions and features are employed.
(a) The NRMS of the estimation error; (b) The NMAE of the estimation error; (c) The CC between estimated force and actual force.
Fig 12.
Variations of the estimated results with the variation of measurement position number.
(a) The variations of NRMS; (b) The variations of NMAE; (c) The variations of CC.
Fig 13.
Variations of the estimated results with the variation of feature number.
(a) The variations of NRMS; (b) The variations of NMAE; (c) The variations of CC.
Fig 14.
Estimated results of the grasp force when GRNNs with different input dimensions are used.
(a) The NRMS of the estimation error; (b) The NMAE of the estimation error; (c) The CC between actual grasp force and estimated grasp force.
Fig 15.
Estimated results of PS41 when different FSs are used.
(a) The NRMS of the estimation error; (b) The NMAE of the estimation error; (c) The CC between actual grasp force and estimated grasp force.
Table 3.
Tukey HSD testing results for NRMS (%) of the MPS of PS41.
Table 4.
Optimal FSs for different MPSs.
Fig 16.
Estimated results of VIWZ when different MPSs are used.
(a) The NRMS of the estimation error; (b) The NMAE of the estimation error; (c) The CC between actual grasp force and estimated grasp force.
Table 5.
Optimal MPSs for different FSs.