| Autor: |
Xiao-Cheng Shi, He-gao Cai, Zong-Hu Chang, Zhaodong Tang, Xinqian Bian |
| Rok vydání: |
2005 |
| Předmět: |
|
| Zdroj: |
2005 International Conference on Machine Learning and Cybernetics. |
| DOI: |
10.1109/icmlc.2005.1527181 |
| Popis: |
The paper proposes a framework of obstacle avoidance for AUV based on the real-time information of forward looking sonar (FLS) in an unknown environment. The whole system includes such modules as obstacle detecting by FLS, tracking obstacles, motion estimation and local path planning to avoid obstacles. Least squares methods and its amends, Kalman filter and some adaptive methods are applied to estimate the motion parameters of different obstacles. We also take advantage of the real-time data stream to track obstacles and obtain their dynamic characteristics. A method of obstacle avoidance for AUV based on genetic algorithm (GA) has been proposed in the paper. The method utilizes floating-point genes, transforms multi-restrictions into the fitness function, such as obstacle avoidance, the minimum distances between way points and trace keeping. The introduction of elitist selection guarantees the constringency of the GA algorithm. Simulations show that the method is perfect for AUV to avoid not only static obstacles but also moving obstacles. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
