Connectionist learning control systems: submarine depth control

Farrell, Jay A.; Goldenthal, B.; Govindarajan, Krishna K.

doi:10.1109/cdc.1990.204050

Cited by 16 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In each case we selected some fixed thruster configurations to experiment with. We also ran the thruster estimation algorithm in segments of 100 runs for each thruster number (2,4,6,8). For each of the 100 runs we generated randomly the thruster configuration.…”

Section: Simulationsmentioning

confidence: 99%

See 1 more Smart Citation

Estimation of Thruster Configurations for Reconfigurable Modular Underwater Robots

Doniec

Detweiler

Rus

2014

Experimental Robotics

View full text Add to dashboard Cite

We present an algorithm for estimating thruster configurations of underwater vehicles with reconfigurable thrusters. The algorithm estimates each thruster's effect on the vehicle's attitude and position. The estimated parameters are used to maintain the robot's attitude and position. The algorithm operates by measuring impulse response of individual thrusters and thruster combinations. Statistical metrics are used to select data samples. Finally, we compute a Moore-Penrose pseudoinverse, which is used to project the desired attitude and position changes onto the thrusters. We verify our algorithm experimentally using our robot AMOUR. The robot consists of a main body with a variable number of thrusters that can be mounted at arbitrary locations. It utilizes an IMU and a pressure sensor to continuously compute its attitude and depth. We use the algorithm to estimate different thruster configurations and show that the estimated parameters successfully control the robot. The gathering of samples together with the estimation computation takes approximately 40 seconds. Further, we show that the performance of the estimated controller matches the performance of a manually tuned controller. We also demonstrate that the estimation algorithm can adapt the controller to unexpected changes in thruster positions. The estimated controller greatly improves the stability and maneuverability of the robot when compared to the manually tuned controller.

show abstract

Section: Simulationsmentioning

confidence: 99%

“…Van de Ven et al give an overview of neural network control of underwater vehicles and simulation studies [13]. Gua et al and Ishii et al learn yaw control [6,7] and Farrell et al and Kodogiannis et al learn depth control [4,8]. They all verify their controllers in simulation as well as through experiments.…”

Section: Introductionmentioning

confidence: 99%

Estimation of Thruster Configurations for Reconfigurable Modular Underwater Robots

Doniec

Detweiler

Rus

2014

Experimental Robotics

View full text Add to dashboard Cite

show abstract

“…The thrusters have an operating range of± 20 r.p.s., enabling a top speed of 4 knots ahead and 2 knots astern. 3. THE PROPORTIONAL-INTEGRAL-DERIVATIVE CONTROLLER.…”

mentioning

confidence: 99%

A Neural Auto-depth Controller for an Unmanned Underwater Vehicle

1997

View full text Add to dashboard Cite

Artificial neural networks offer an alternative strategy for the nonlinear control of unmanned underwater vehicles (uuvs). This paper investigates the use of a multi-layered perceptron (MLP) network in controlling an uuv over a sea-bed profile and compares the use of applying chemotaxis learning to that of the more commonly employed back propagation algorithm. The results show that, for differing sized MLPS, the chemotaxis algorithm produces a successful controller over the sea-bed profile in an improved training time. Also it will be shown that, in the presence of noise and change in vehicle mass, the neural controller out-performed a classical proportional-integral-derivative controller.i. I N T R O D U C T I O N . The control of uuvs has always offered a challenge to the control engineer due to the combined nonlinear nature of both the vehicle itself and the environment in which it operates. Whilst the traditional tethered remotely operated vehicle (ROV) frequently employs a human pilot in the loop, there has been a drive towards giving the vehicle some form of autonomy to remove the responsibility for die dynamic control from the operator and providing the ROV with a limited autonomous capability via an autopilot.uuvs are currently at the centre of considerable research interest in both the commercial and military sectors. They offer advantages over ROVS in terms of flexibility of use (range and depth) and can replace vulnerable human divers or expensive manned submersibles. Techniques including gain-scheduling, sliding mode and fuzzy logic have all been considered for use in uuv control and these are reviewed in Farbrother and Stacy. 1 The applications of neural networks to the dynamic control of uuvs has received a certain amount of attention. 2 ' 3 Generally these have considered midwater control for heading and depth-keeping, however, one further important area which has drawn minimal consideration is that of bottom profiling. For tasks such as pipe inspection, oceanographic surveying and mine deployment, some form of sea-bed profiling is required and, in such cases, the uuv has to maintain a safe height above the bottom. Thus, herein, are reported the results of a study into the application of neural networks to control an uuv in the sea-bed profiling mode. Also results are presented to show the capability of the neural controller to cope with noisy sensor signals and a change in vehicle mass during a simulated mission.A particular feature of the work herein is the employment of the chemotaxis 292

show abstract

Neural network augmented identification of underwater vehicle models

Ven

Johansen

Sørensen

et al. 2007

Control Engineering Practice

View full text Add to dashboard Cite

In this article the use of neural networks in models for underwater vehicles is discussed. Rather than using a neural network parallel to the known model to account for unmodeled phenomena in a model wide fashion, knowledge regarding the various parts of the model is used to apply neural networks for those parts of the model that are most uncertain. As an example, the damping of an underwater vehicle is identified using neural networks. The performance of the neural network based model is demonstrated for an AUV that changes its physical characteristics during a simulated intervention operation.

show abstract

Connectionist learning control systems: submarine depth control

Cited by 16 publications

References 11 publications

Estimation of Thruster Configurations for Reconfigurable Modular Underwater Robots

Estimation of Thruster Configurations for Reconfigurable Modular Underwater Robots

A Neural Auto-depth Controller for an Unmanned Underwater Vehicle

Neural network augmented identification of underwater vehicle models

Contact Info

Product

Resources

About