A behavior-based architecture for autonomous underwater exploration
Introduction
In unstructured, unknown, and dynamic environments, such as those encountered by outdoor mobile robots, an intelligent agent must address the issues of incomplete and inaccurate knowledge; it must be able to handle uncertainty in both its sensed and a priori information, in the current state of the agent itself, as well as in the effects of the agent’s actions. Autonomous underwater vehicles have the potential to play a substantial role in ecology management, geophysical surveying, offshore oil and gas exploration and well maintenance, undersea mineral exploration and mining, and in surveillance and defence. However, a subsea environment is particularly unstructured and dynamic, the degrees of freedom in the control and estimation problem are greater than on land, and there exists no reliable positioning data.
To function effectively in such conditions, an autonomous system must be responsive to its environment, proceeding in a data-driven manner, as well as goal-oriented, taking into account the higher-level goals of the system. When used appropriately, deliberative planning and reactive control complement each other and compensate for each other’s deficiencies. In order to achieve this desired symbiosis of deliberative and reactive elements, the distributed architecture for mobile navigation (DAMN) consists of a group of distributed task-achieving modules, or behaviors, communicating with a centralized command arbiter [8].
Within this framework, we have developed an architecture for behavior-based control of an autonomous underwater vehicle for the purpose of inspection of coral reefs. A survey project for reef management, carried out by the Australian Institute of Marine Sciences, is designed to provide long-term quantitative data about corals, algae and marine life over the extent of the Great Barrier Reef. This data are for studies of abundance and population change in selected organisms on a large geographic scale. Currently, visual transect information is recorded using underwater video cameras held by a diver following a rope, as shown in Fig. 1.
The reef surveillance task, as it is currently defined, consists primarily of following an assigned path while maintaining a fixed altitude above the reef and avoiding collisions [13]. Independent behaviors and arbiters, using decoupled controllers, have been developed as a modular means of accomplishing these various sub-tasks. For example, two behaviors have been developed for the path following aspect of the task; the first behavior uses video input to track a rope laid out along the coral, while the second behavior uses sonar to detect passive beacons.
The DAMN arbiters are then responsible for combining the behaviors’ votes to generate controller commands. Fuzzy logic arbiters are currently used for control of the vehicle; another set of behaviors and arbiters that perform utility fusion [9] are under development. In both cases, the distributed, asynchronous behaviors provide real-time responsiveness to the environment, while the centralized command arbitration provides a framework capable of producing coherent behavior. A task-level controller selects which behaviors are active at any given moment.
Section snippets
Oberon submersible vehicle
We have constructed a simple low-cost underwater robot named Oberon, shown in Fig. 2, as a test-bed for experimental work in autonomous undersea navigation. There are currently five thrusters on the vehicle. Three of these are oriented in the vertical direction while the remaining two are directed horizontally. This gives the vehicle the ability to move itself up and down, control its yaw, pitch and roll, and move forwards and backwards. This thruster configuration does not allow the vehicle to
Behavior-based control architecture
In order to avoid the bottlenecks and brittleness of centralized systems, the DAMN consists of a group of distributed specialized task-achieving modules, called behaviors, communicating with centralized command arbiters, as shown in Fig. 4. A behavior encapsulates the perception, planning and task execution capabilities necessary to achieve one specific aspect of robot control, and receives only that data specifically required for that task [2]. Each behavior operates independently and
Experimental results
This section presents some preliminary results from deployment of the vehicle in a natural terrain environment along Sydney’s coastline. The first behavior developed was Maintain Altitude, which keeps the vehicle at a fixed standoff distance from the ocean floor. In the experimental results showing altitude and depth in Fig. 12, the desired altitude was 1.5 m, which was maintained within a standard deviation of 0.2 m, as can be seen in the first plot if altitude vs. time. This is despite a
Conclusion
We have developed a behavior-based system for control of an autonomous underwater vehicle performing a survey of coral reefs. Implemented behaviors provide the ability to avoid collisions, maintain a proper standoff distance, and following the transect either using a rope with video or targets with sonar. Command fusion is performed using a fuzzy logic arbiter, and a utility fusion system is under development. A task-level controller selects which behaviors should be active according to
Acknowledgements
The authors would like to thank the other members of the AUV project team at the ACFR, as well as Dr. Hugues Talbot of the CMIS CSIRO, who assisted in developing the rope detection algorithm.
References (15)
Statistical Decision Theory and Bayesian Analysis
(1985)A robust layered control system for a mobile robot
Journal of Robotics and Automation
(1986)Task networks for controlling continuous processes
- et al.
A behavior-based system for off-road navigation
Journal of Robotics and Automation
(1994) Principles of Artificial Intelligence
(1980)Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference
(1988)- P. Pirjanian, The notion of optimality in behavior-based robotics, Journal of Robotics and Autonomous Systems, 2000 (to...
Cited by (46)
Control architecture of autonomous underwater vehicle for coverage mission in irregular region
2021, Ocean EngineeringCitation Excerpt :A group of parallel behaviors are acted independently without planning according to the environmental conditions (Brooks, 1986). Reactive architecture is widely applied in AUV and other agents because of its advantage of real-time response (Muller et al., 1994a; Müller et al., 1994b; Rosenblatt et al., 2002; Gu et al., 2003; Flanagan et al., 2003). However, it is difficult to design a set of behaviors in order to plan and to specify high-level goals.
Function-segment artificial moment method for sensor-based path planning of single robot in complex environments
2014, Information SciencesCitation Excerpt :In a realistic setting, path planning of a robot cannot be based on complete a priori knowledge of the environment [24]. Therefore, much attention has been paid to sensor-based path planning of mobile robots in unknown, complex and dynamic environments [8,12,9,23,24,30,31], and this paper focuses also on this topic. Among global sensor-based path planning methods, the visibility graph [10,14,19,31,32], the Voronoi diagram [10,19,31,33,35], and the tangent graph [15,20] are well-known.
Behavioral modeling with the new bio-inspired coordination generalized molecule model algorithm
2013, Information SciencesCitation Excerpt :2) In [9], Hervé Frezza-Buet and Frédéric Alexandre present a cortical model, which is inspired by biological data, to allow an animat to perceive and act on its environment to learn the regularities of the environment and to use them to satisfy certain goals; they then describe this model at different levels of abstraction, from both a biological and a computational point of view. ( 3) Julio Rosenblatt et al. develop a behavior-based system for controlling an autonomous underwater vehicle to perform a survey of coral reefs [17]; the implemented behaviors have the ability to avoid collisions, to maintain a proper standoff distance, and to follow the transect using video or sonar. ( 4) In [13], Seungho Lee et al. extend the decision field theory (DFT) for behavior modeling in a dynamic environment using Bayesian belief networks. (
Estimating spatio-temporal fields through reinforcement learning
2022, Frontiers in Robotics and AIA Local Planner Based on Environmental Geometric Features for Rescue Robots
2022, ICNSC 2022 - Proceedings of 2022 IEEE International Conference on Networking, Sensing and Control: Autonomous Intelligent SystemsHuman diver-inspired visual navigation: Towards coverage path planning of shipwrecks
2021, Marine Technology Society Journal