Simulating an optical guidance system for the recovery of an unmanned underwater vehicle

The underwater environment is hazardous, remote, and hostile, Having a look and interacting in this environment is a challenge for a human supervisor. Moreover, to design an unmanned underwater vehicle (UUV), or evaluate its performance in operation, access to the underwater world is required. A pow erful way to visualize the behavior of the vehicle is to create a virtual w orld with all functionalities of the real world, and to operate the vehicle in this virtual world. This implementation of a virtual laboratory is an e xcellent way to perform meaningful simulations and complex system testing. In order to study the problem of UUV recovery by a submarine, simulations f an be a great help. After the vehicle has finished its mission, it has to p roceed to a predetermined rendezvous area to return to the submarine. When the UUV and submarine have detected each other, the recovery begins. The ve hicle needs a very accurate guidance mode in order to steer itself to the r ecovery device. An additional guidance system coupled with a nominal naviga tion system may be a wag to ensure safe vehicle navigation through the flow around the slowly moving submarine, When considering the different technological possibilities concerning the a dditional guidance system, a functional design approach leads to the choice of an optical technology, The assumptions for the. optical guidance mode a re that the UUV is fitted with a camera and a high-powered light is located at the edge of the recovery device. The principle is that the UUV tracks t he highest intensity; light source. This system is easy to operate, but the distance between the UUV and the submarine must not exceed 200 m, due to l ight attenuation. In order to simulate and stimulate such a guidance system , it is interesting to create realistic views representing what the UUV may see according to this environment. A software program was designed, taking into account the physical phenomena occurring during the light propagation under the water, to simulate the kind of images that can be obtained from a camera, An underwater scene is generated, including any object and any li ght source, and including the physical properties of the sea water (reflect ion, refraction, absorption, and scattering). A ray-tracing algorithm simul ates the operation of a camera by calculating and rendering the path invers e to the Light path. Because both camera optics and hydrodynamics response are simulated using high-resolution physics models, this virtual camera pro vides physically based sensor inputs to the robot software in the laborator y. Control orders concerning the vehicle result from the real-time robot inter pretation of the generated image. To steer the vehicle to the light source, the navigation system has to take into account the image and the informati on carried by the image: shall the vehicle go up or down, starboard or port , of slowly or quickly to navigate in the direction of the light? To answer these questions, the image synthesizer module is integrated with an underw ater virtual world. The vehicle performs a mission described in a file with simple keywords. When the mission controller reads a key,word activating t he additional guidance mode, the image synthesizer computes the image of th e camera and returns ordered data for the depth, heading, and speed to the navigation system. At the next step, another image is processed and new ord ers are returned, until the vehicle reaches the area around the light sourc e. If the light source is put directly on the recovery device, stable guida nce through recovery becomes possible. A variety of simulations were performed, with varying light sources and pos itions, to verify. proper guidance system operation during different UUV/su bmarine configurations. The results obtained during the simulations were us ed to create an optical guidance control mode. All the steps for designing such a simulated guidance system are described in this communication.