An Adaptive Switching Cooperative Source Searching And Tracing Algorithms For Underwater Acoustic Source Localization
Swarm robotics is a study of how to organize a relatively large number of simple robots to achieve a robust, flexible and scalable solution for a given task. Searching a source with a complex spatial distribution pattern is one of the possible swarm robotics tasks. In a source searching task, two po...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/51388/1/An%20Adaptive%20Switching%20Cooperative%20Source%20Searching%20And%20Tracing%20Algorithms%20For%20Underwater%20Acoustic%20Source%20Localization.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Swarm robotics is a study of how to organize a relatively large number of simple robots to achieve a robust, flexible and scalable solution for a given task. Searching a source with a complex spatial distribution pattern is one of the possible swarm robotics tasks. In a source searching task, two possible scenarios can occur: source detected and source not detected. In this study, a complete solution to the two scenarios through an adaptive algorithm switching strategy is explored. Firstly, to detect the source, a Source Detection Algorithm (SDA) known as a Distributed Lévy Flight (DLF) is proposed. To improve exploration performance of the individual agent, a turning angle limit and boundary reflection is introduced in DLF. In order to optimize search space exploration and to maintain inter-robot communication connectivity at swarm level, a dispersion algorithm based on attraction and repulsion force is proposed. Secondly, to trace the source to its approximate location, a Source Tracing Algorithm (STA) known as an Asynchronous Dynamically Adjustable Particle Swarm Optimization (ADAPSO) is suggested. The ADAPSO parameters are adaptively and asynchronously adjusted based on feedback informations to improve convergence speed, to avoid robot trapped into local optima and to minimize target overshooting. In addition, the ADAPSO position update equation is modified to anticipate position adjustment to ensure communication connectivity. To adaptively switches between the two algorithms, an adaptive switching algorithm based on a Generalized Likelihood Ratio Test (GLRT) is proposed. To demonstrate the algorithm switching principle, underwater acoustic source localization using a swarm of Autonomous Surface Vehicles (ASVs) is considered. By considering the ASVs as swarm robotics testing platforms, each algorithm is evaluated and benchmarked against several existing algorithms through simulation studies. The obtained results show that the performance of the DLF for source detection outperformed other benchmark algorithms in term of search space exploration capability and the time taken to detect the source. The ADAPSO for source tracing achieved better tracing performance with better success rate and reduced the time taken to trace the source to its approximate location compared to the benchmark algorithms. Finally, the feasibility of the proposed algorithms for underwater acoustic source localization is confirmed through simulation and experimentation where the achieved average accuracy of source position estimation is 0.4 m and 4.2 m, respectively. |
|---|