Data collection scheme with minimum cost and location of emotional recognition edge devices

This paper develops a real-time and reliable data collection system for big scale emotional recognition systems. Based on the data sample set collected in the initialization stage and by considering the dynamic migration of emotional recognition data, we design an adaptive Kth average device clustering algorithm for migration perception. We define a sub-modulus weight function, which minimizes the sum of the weights of the subsets covered by a cover to achieve high-precision device positioning. Combining the energy of the data collection devices and the energy of the wireless emotional device, we balance the data collection efficiency and energy consumption, and define a minimum access number problem based on energy and storage space constraints. By designing an approximate algorithm to solve the approximate minimum Steiner point problem, the continuous collection of emotional recognition data and the connectivity of data acquisition devices are guaranteed under the energy constraint of wireless devices. We validate the proposed algorithms through simulation experiments using different emotional recognition systems and different data scale. Furthermore, we analyze the proposed algorithms in terms of topology for devices classification, location accuracy, and data collection efficiency by comparing with the Bayesian classifier-based expectation maximization algorithm, the background difference-based moving target detection arithmetic averaging algorithm, and the Hungarian algorithm for solving the assignment problem.