Waveform Analysis for Human Handling Force Data Using Wavelet Transform

Data for handling forces exerted by individuals during manipulation or assembly tasks are important for calculating the physical load on the body. However, because it is difficult to extract features from force data using conventional waveform analysis methods due to the noisy and nonstationary characteristics of the data, we focused on wavelet transform analysis, which is used to analyze nonstationary signals and pulse waves, and attempted to apply spectral analysis to the handling of force data. The purpose of this study was to develop a new waveform analysis method for handling force data using wavelet transform and demonstrate its effectiveness. First, to discriminate between human handling force components and other components in the data, a wavelet multiresolution analysis was applied to pushing force data obtained by using a push button switch and the human handling force component was extracted. Second, time and phase shifts in feature points such as edges and peaks were compared between the raw and extracted waveforms. The results show that temporal and phase shifts for feature points were sufficiently small, and this method proved superior to the approach of smoothing the handling force waveform. This method makes it possible to find a pattern for handling force exertion and understand its features.