When an organization sets out to map the sea floor, it will typically use a device known as a bathymetric lidar (Light Detection and Ranging) unit. These are large and can weigh almost 600 lb (272 kg), so they're mounted on crewed aircraft that fly over the area to be mapped. Led by Dr. Grady Tuell, a team at Georgia Tech has now developed a cost-effective new system that they claim could lead to much smaller, more efficient bathymetric lidars, capable of being carried by a UAV.
On a regular bathymetric lidar, laser light pulses are emitted down through the water to the sea bed, then reflected back up to the device. By analyzing the amount of time that elapses between a pulse being emitted and its reflection being received, it's possible to gauge the depth at that particular point. By combining a multitude of these readings, a three-dimensional point cloud map of the sea floor is subsequently created. It takes time to analyze the signals, however, as both the refractive index and the turbidity of the water must be accounted for.
In the Georgia Tech system, the accuracy of each measurement point is checked much more quickly and efficiently. It utilizes a computing technique known as total propagated uncertainty (TPU), which incorporates "statistics, calculus and linear algebra." Additionally, its hardware consists of field programmable gate arrays (electronic circuits that are configured by the user for the specified task), plus central-processing and graphics-processing units.
As a result, it's able to process approximately 37 million points per second – by contrast, some traditional bathymetric lidars only manage speeds of 1,000 per second.
The reflection of each pulse is analyzed before the next one is emitted, allowing for data to be transmitted to ground-based crews in real time. Because of that immediacy, it has been suggested that the system could be particularly useful for applications such as searching for enemy mines or submarines. That said, it could additionally be used on land, as its laser pulses are reportedly also able to "see through" foliage to reveal what's on the forest floor.
The present prototype has been gantry-mounted over a pool for testing and evaluation. A CAD design has been developed for a deployable version that's approximately half the size and weight of traditional models (and which consumes half as much power), that could be mounted on larger UAVs such as unmanned helicopters. Down the road, however, the team hopes to create a much lighter, more compact unit that could be carried by smaller, more economical UAVs.
Source: Georgia Tech
