3D Space Obstacle Avoidance System for a Large-Scale Autonomous Underwater Vehicle
- Ben Li (Huazhong University of Science and Technology) | Guanxue Wang (Huazhong University of Science and Technology) | Guohua Xu (Huazhong University of Science and Technology) | Xin Zhang (Huazhong University of Science and Technology) | Han Xu (University of Southern California)
- Document ID
- International Society of Offshore and Polar Engineers
- The 28th International Ocean and Polar Engineering Conference, 10-15 June, Sapporo, Japan
- Publication Date
- Document Type
- Conference Paper
- 2018. International Society of Offshore and Polar Engineers
- fuzzy planner, obstacle avoidance system, large-scale AUV, navigation sensors
- 0 in the last 30 days
- 16 since 2007
- Show more detail
When an autonomous underwater vehicle (AUV) is operating in an unknown environment, a reliable obstacle avoidance system (OAS) is essential, especially for a large scale under-actuated AUV. This paper introduces an (OAS) for large-scale AUV in a 3D environment based on navigation sensors. For the OAS, reacting behavior given by fuzzy planner and predefined behavior are carried out to achieve the collision avoidance in both horizontal and vertical plane. Meanwhile, the speed of the AUV is simultaneously controlled by speed fuzzy planner to reach a desirable value. Finally, computer simulations in the 3D virtual environment are carried out to prove the effectiveness of the OAS.
The path planning of AUV can be divided into two tasks including global path planning and local path planning. The global path planning is to find an optimal path without obstacles from the start point to the goal point based on the known information and gives a series of waypoints that must be arrived. Obstacle avoidance can be considered as part of the local path planning task. When obstacles are detected, the vehicle will take collision avoidance behaviors. After avoiding the obstacles, the AUV will manage to get back to the originally planned path. So, a reliable OAS is essential for the safety of the vehicle and the success of the mission. Numerous algorithms have been created to design an OAS. Including geometric constraint, virtual force field, vector field histogram, potential field and fuzzy logic inference. Zhang, Wille and Knoll (1996) proposed that classical robot control architecture need to employ the Sensing-Modelling-Planning-Action (SMPA) strategy. However, in this way, there is a time delay between the perception of the obstacles and taking actions to avoid the obstacles, due to the complexity of the algorithms used for modeling and planning. Besides, this method requires high quality of computers and sonars to compute and sense, which is undesirable. Fuzzy logic was developed based on the relative importance of precision and allows a system to make inferences based on the uncertain or incomplete information gathered. It is also noted that (Zhao, Lu, and Anvar, 2010) Fuzzy logic relies on heuristic knowledge which is subject to the designer’s experience and interpretation of the system. A fuzzy system deals with the nonlinear match of input and output data. For large scale AUV, because of the high speed and large inertia, it is crucial for the AUV to respond in time to avoid the collisions when encountering obstacles. Therefore, the fuzzy logic algorithm is more suitable for OAS to cope with the obstacles in time based on sonar information. Xu and Feng (2009) proposed an AUV fuzzy obstacle avoidance method under event feedback supervision but did not discuss the details of the fuzzy planner. Abbasi, Danesh and Ghayour (2010) proposed a path fuzzy planner for AUV to avoid moving unknown obstacles in the horizontal plane. This paper designed an OAS for large-scale AUV in 3D space based on navigation sensors. Collision avoiding was achieved by making behaviors given by fuzzy planner along with the predefined behavior.
|File Size||896 KB||Number of Pages||6|