Research on Ocean Multitarget Tracking Control System Based on the Echo State Network

Abstract

Wang, T. and Wang, X., 2019. Research on ocean multitarget tracking control system based on the echo state network. In: Li, L.; Wan, X., and Huang, X. (eds.), Recent Developments in Practices and Research on Coastal Regions: Transportation, Environment and Economy. Journal of Coastal Research, Special Issue No. 98, pp. 235–238. Coconut Creek (Florida), ISSN 0749-0208.In recent years, with the rapid development of artificial intelligence technology, the echo state network (ESN) has shown some advantages in nonlinear time series processing and dynamic prediction system, which has attracted extensive attention of domestic and foreign researchers. As an improved recurrent neural network, the ESN uses a large-scale sparse connection structure as the hidden layer. By adjusting the weights of output connections only, the whole training stage is made simply and efficiently, which overcomes the shortcomings of the classical recurrent neural network learning rules, such as being difficult to implement and having a long running time. The ESN has gradually become an important method for the prediction of nonlinear time series. Aiming at the slow tracking speed of traditional ocean multitarget tracking control system, this paper designs an ocean multitarget tracking control system based on the echo state network. Through the target detection algorithm, we select the ocean multiobjective control area and calculate the target center position. In this way, the result of the target state estimation is obtained. By confirming the target trajectory, the multiobjective ocean control based on the ESN is achieved. The simulation results show that the speed of the system is 22.78% higher than the traditional speed when it is designed for target control, which meets the requirements of ideal ocean multiobjective control.