Particle Swarm Optimization Methods And Their Applications In Seabed Terrain-Aided Navigation

The seabed terrain-aided navigation (STAN) is one of the key technologies in submarine automation and intelligent navigation. It revises the error of reference navigation system using digital map, and the main purpose is to make navigation more exact. The search algorithm is the core of STAN, traditional STAN algorithm uses ergodic search method, which brings too much calculation, and affects the navigation efficiency a lot. Furthermore, as the particularity of STAN, STAN is more complex than land environment. So the optimization of complex search strategy is the key method to improve efficiency of STAN.Particle Swarm Optimization (PSO) is proposed in the end of the 20th century, it is simple in comcept, easy in implementation and few in parameters, so it has attracted much attention since proposed, and becomes a study hotspot in the world. Many successful applications verified the validity of PSO. This paper is mainly about PSO algorithm and its application in STAN systerm. The main content is as follows:1. The theory and characterization of PSO is studied. First the artificial life and artificial life computation are studied. Then the unified framework and design steps for solving optimization problem are analysed. After summarize the basic theory of PSO, the comvergence of PSO is investigated in details.2. The influence of swarm topology on PSO is studied in two aspects: typical swarm topology and communication style between two particles. A weighting model is established to represent the particle's force, and a more reasonable weighting function is designed for the application of STAN. A parameter setting methods based on pso for pso thinking is proposed. The parameter selection problem is transformed into nonlinear optimization problem, and the PSO algorithm is selected to solve the problem.3. As PSO has premature convergence problem when solving multi-modal problems, a fuzzy particle swarm optimization algorithm based on swarm diversity is proposed. After the swarm diversity is analysed, the rapid decline of the diversity is considered as the main reason for premature convergence. Mutation strategy is introduced to restrain the decline of diversity, and the fuzzy logic controller is designed for dynamical adjustment of mutation rateρ₀ and inertia weight w .4. The PSO algorithm is essentially continuous, so its application in discrete domain is an important study aspect. A discrete PSO algorithm with two sub-swarms is proposed in the paper. First, the double sub-swarms movement model is established, and based on the movement model, the changing rules of sub-swarms'size is designed. At the same time, a position updating strategy based on the hill-climbing theory and pseudo mutation strategy are introduced into the algorithm to increase the convergence speed and global optimization performance. The influence of parameters is analysed, and the whole flow of the algorithm is designed.5. According to the deep research of way and key technologies of STAN, the seabed terrain-aided navigation algorithm based on the Particle Swarm Optimization is proposed. Based on the anylysis of particularity of STAN, the determination methods of searching space and matching sequence is given, and the hausdorff distance which has strong anti-interference ability is chosen to be the similarity measure. The matching search policy based on PSO is designed to increase the efficiency of matching search.