Underwater vehicle with sonar array

Abstract

An underwater vehicle including an axi-symmetric framing system rotatable about a centerline to define a shell of revolution having a uniformly-convex outer boundary. A narrow-beam sonar array is mounted on the axi-symmetric framing system, and includes a multitude of simultaneously-fireable and/or asynchronously-fireable transducers distributed substantially evenly over a 4π-steradian viewing angle. The present invention provides the necessary configuration for a vehicle wherein an internal algorithm can compare a “new” geometry to an “old” geometry collected earlier to construct a best fit of the new world map with the old world map and locate the vehicle within the context of the new world map. This then provides a completely independent mechanism for correction of the gradual drift in x and y that is not dependent on any form of external navigation aid.

Description

CROSS-REFERENCES TO RELATED APPLICATIONS

This original nonprovisional application claims the benefit of U.S. provisional application No. 60/953,070, filed Jul. 31, 2007 and entitled “Autonomous Underwater Vehicle for 3D Mapping and Navigation in Labyrinthine Environments,” which is incorporated by reference herein.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

The U.S. Government has a paid-up license in this invention and the right in limited circumstances to require the patent owner to license others on reasonable teens as provided for by the terms of Grant No. NNG04GC09G awarded by the National Aeronautics and Space Administration.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to the exterior geometry and structural configuration for an underwater vehicle. More specifically, the invention relates to the shape of the vehicle in combination with the positioning of a plurality of sonar devices on the exterior geometry of the vehicle.

2. Description of the Related Art

Autonomous underwater vehicles (AUVs) are common scientific devices used for oceanographic research and bathymetric measurements. Because AUVs are, by definition, unmanned and autonomous, they are ideal for high-risk activities within the depths of the world's oceans. Oil and gas companies, for example, frequently use AUVs to make detailed maps of the seafloor prior to installing the infrastructures for oil rigs and pipelines. AUVs have also been used to map an area to determine whether enemy mines are present. Scientists also frequently use AUVs to study the ocean floor.

AUVs are frequently tasked to autonomously navigate in labyrinthine environments, such as rock caverns, fissures, ice pack cracks, underground flood tunnels, dam bypass tunnels, and underwater structures. In such environments, complicated terrain frequently surrounds the vehicle on all sides. As such, these conditions give rise to certain problems for traditional underwater vehicles.

First, labyrinthine environments create a true three-dimensional navigation problem for AUVs. Because of the presence of these surfaces, it is very unlikely, if not impossible, to obtain continuous or even periodic navigational updates from external navigation sources. It is common, for example, for ocean-going underwater vehicles to periodically surface for the purpose of obtaining a navigation fix from a global positioning system. It is further standard practice in the design of ocean-going underwater vehicles to have as their mission abort mechanism a system that makes the vehicle positively buoyant, so that in the event of a problem it will rise to the surface and initiate communication to indicate its status. In a complex labyrinthine environment, however, the net result would simply be the permanent loss of the very-expensive vehicle.

Traditional AUVs suffer from one additional, and frequently fatal, design flaw for working in labyrinthine environments: They are almost always torpedo shaped—that is, they almost always have long, cylindrical bodies, blunt nose cones, and a single, aft propeller. Moreover, traditional AUVS are generally designed for oceanographic research and bathymetric measurements. As such, they may employ downward-looking swath-type sonar systems (e.g., multibeam, sidescan, acoustic x-z

While usually intractable to solve in closed form, the SLAM
which is a combination of the distribution of vehicle trajectories and the distribution of maps (where ηis a Bayesian normalization factor). It is often very hard to express these distributions in closed form, and so a particle filter maintains a discrete approximation of the SLAM posterior using a large set of samples, or particles. Because the posterior is factorized into trajectory and map distributions, the particles are actually a specific trajectory and map pair. In fact, given a vehicle trajectory, the map can be interactively iteratively updated as discussed supra. This is known as Rao-Blackwell factorization (Doucet et al. 2000) (see Doucet, A., de Freitas, N., Murphy, K., and Russell, S. (2000). Rao-blackwellised particle filtering for dynamic bayesian networks. Proc. of the Sixteenth Conf. on Uncertainty in AI, pages 176-183).

The algorithm of the present invention comprises the following steps:

Initialize. The particles start with their poses initialized according to some initial distribution and their maps optionally containing some prior information about the world. This is called prior distribution.

Predict. The dead-reckoned position innovation is computed using the navigation sensor sensors. A new pose is predicted for each particle using a vehicle motion model. This new distribution of the particles is called the proposal distribution.

Weight. A weight is computed for each particle by comparing the real range measurements to ranges simulated by ray-tracing with the particle pose and map. A particle that has a pose and map that are consistent with the real range measurements will have a high weight, whereas particles which are inconsistent will have low weights.

Resample. The algorithm resamples the set of particles according to the weights such that particles with low weights are likely to be discarded and particles with high weights are likely to be duplicated (poses and maps). The resample set of particles is now the new estimate of the new SLAM posterior.

Update. The measurements are inserted into the particle maps as described supra to update the evidence of all the voxels that lie in the conic sonar beam model of each measurement relative to the particle position.

Estimate. A position estimate is generated from the particles, When SLAM is being used to provide a pose for the rest of the vehicle control software, it is desirable to turn the set of particles into a single point estimate.

The predicting, weighing, resample weighting, resampling, updating and estarioatbig steps are then repeated.

FIG. 15A-15B show experimental results of the effectiveness of the algorithm described herein. The vehicle represented in these figures was programed to execute a proximity operation starting from position S at coordinates [1, −1] and proceeding to a wall standoff distance of two meters in a circular test tank. The vehicle was then to follow the wall clockwise to a point ninety-degrees from the contact point, and then cycle back and forth along the same segment of the wall, while maintaining a standoff distance of two meters. The circle 230 represents the two meter target standoff circle. The trace 232 in FIG. 15A represents the dead-reckoned vehicle trajectory, while the trace 234 in FIG. 15B represents the SLAM-localized vehicle trajectory. The ending point E′ of the vehicle in FIG. 15B is significantly closer to the intended standoff distance than the ending point E of the vehicle in FIG. 15A.

The present invention is described above in terms of a preferred illustrative embodiment of a specifically described autonomous underwater vehicle, as well as alternative embodiments thereof. Those skilled in the art will recognize that alternative constructions of such an assembly can be used in carrying out the present invention. Other aspects, features, and advantages of the present invention may be obtained from a study of this disclosure and the drawings, along with the appended claims.

Claims

1. An underwater vehicle for operating in an underwater environment, said underwater vehicle comprising:

a source of power;

an axi-symmetric framing system rotatable about a centerline to define a shell of revolution having a uniformly-convex outer boundary;

a narrow-beam sonar array connected to said source of power and mounted on said axi-symmetric framing system, said narrow-beam sonar array having a first and third plurality of transducers distributed substantially evenly over along said outer boundary within a first plane, a second and fourth plurality of transducers distributed substantially evenly over along said outer boundary within a second plane, and a plurality of lateral imaging transducers orientated to project within a third plane that is orthogonal to said first and second planes;

a plurality of digital signal processors connected to said narrow-beam sonar array; and

a plurality of bidirectional circumferential thrusters connected to said source of power and mounted, to said axi-symmetric framing system, said plurality of bidirectional circumferential thrusters being oriented to selectively cause rotation of said axi-symmetric framing system and propulsion of said vehicle within said underwater environment, said plurality of circumferential thrusters positioned within said shell of revolution;

wherein each transducer of said plurality of transducers each of said transducers is configured to project a beam having a width of no more than about two degrees or less radially outwardly along an associated beam path, and the angle of separation between beam paths transducers is at least greater than about ten degrees.

2. The underwater vehicle of claim 1 wherein each of said transducers comprises projector and a corresponding hydrophone, each of said hydrophones is operative to receive a reflected signal originating from its corresponding projector to (a) determine, said transducers configured to communicate said signals to said plurality of digital signal processors for use by a main processor bank in (a) determining the position of said underwater vehicle within said underwater environment; (b) map mapping said underwater environment; and/or or (c) move moving said underwater vehicle within said underwater environment.

3. The underwater vehicle of claim 1 further comprising at least one vertical thruster mounted to said axi-symmetric framing system and positioned within said shell of revolution.

4. The underwater vehicle of claim 1 further comprising at least one flotation panel mounted to said axi-symmetric framing system.

5. The underwater vehicle of claim 1 wherein said plurality first and second pluralities of transducers further comprises: comprise a plurality of obstacle avoidance transducers; and wherein said third and fourth pluralities of transducers further comprise a plurality of fine imaging transducers.

6. The underwater vehicle of claim 5 wherein said plurality of digital signal processors further comprises:

a first digital signal processor stack connected to said plurality of obstacle avoidance transducers; and

a second digital signal processor stack connected to said plurality of fine imaging transducers.

7. The underwater vehicle of claim 1 further comprising a main processor bank contained within a pressure housing mounted to said axi-symmetric framing system.

8. The underwater vehicle of claim 7 further comprising a motor control system contained within a pressure housing and electrically connected to said main processor bank, said motor control system operative to actuate said circumferential thrusters.

9. The underwater vehicle of claim 7 further comprising a Doppler velocity log electrically connected to said main processor bank.

10. The underwater vehicle of claim 7 further comprising an inertial measurement unit contained within a pressure housing and electrically connected to said main processor bank.

11. The underwater vehicle of claim 7 further comprising:

a variable buoyancy system contained Within a pressure housing and electrically connected to said main processor bank;

a ballast chamber connected to said variable buoyancy system; and

at least one gas supply tank in fluid communication with said ballast chamber.

12. The underwater vehicle of claim 1 wherein said plurality of circumferential thrusters comprises:

a first pair of circumferential thrusters mounted to said axi-symmetric framing system at positions distal from said centerline and orientated to provide thrust in a first direction perpendicular to said centerline; and

a second pair of circumferential thrusters mounted to said axi-symmetric framing system at positions distal from the centerline and orientated to provide thrust in a second direction perpendicular to said centerline, wherein said second direction is coplanar with and orthogonal to said first direction.

13. The underwater vehicle of claim 7 further comprising a processor-readable medium electrically connected to said main processor bank, said processor-readable medium comprising a set of computer readable instructions for estimating a trajectory of the vehicle, the set of instructions comprising:

constructing a 3D compact map data structure representative of the surrounding environment within said processor-readable medium;

initialing a set of particles within said processor-readable medium, each particle of the set having an associated pose;

predicting a new pose for each particle of the set of particles using dead reckoning navigation;

taking real range measurements of the vehicle environment;

assigning a weight to each particle of the set of particles by comparing the real range measurements to simulated range measurements derived by ray-tracing with the particle pose and map, wherein the weight for a particle having a pose and map consistent with real range measurements is high and the weight for a particle having pose and map that are inconsistent with real range measurements is low;

resampling the set of particles according to the assigned weights, wherein particles with low weights are likely to be discarded and particles with high weights are likely to be duplicated;

updating said 3D compact map data structure based on the real range measurements; and

generating a position estimate of the vehicle within the vehicle environment from the set of particles.