GulfSCEI Capabilities

Compute Capabilities

GulfSCEI Servers:

GulfSCEI-USACE is a cutting-edge server setup, a testament to our commitment to harnessing the power of advanced technology. Our server is powered by the AMD EPYC™ 7713 processor, a beast of a CPU with 64 cores and 128 threads, capable of handling the most demanding workloads with ease. Complementing this is a massive 1 TB of RAM and four NVIDIA A100 GPUs, each with 80 GB of dedicated VRAM. These GPUs are not just about raw power; they're about versatility, supporting a wide range of data types and workloads, and offering up to 624 TFLOPS of performance with Tensor Cores. Our server is connected to an HPE MSA 2060 Large Form Factor storage system via a dedicated SAS link, operating at a high-speed of 12 Gb/s. This ensures rapid data retrieval and transfer between the server and the storage system, which boasts a substantial capacity of 300 TB.

  • AMD EPYC™ 7713 with 64 cores
  • 1 TB of RAM
  • 4x Nvidia A100 GPUs w/ 320 GB total VRAM
  • 12 Gb/s SAS link to 300 TB storage system
An IT specialist monitors the GulfSCEI servers.
Fig. 1.1: An IT specialist monitors the GulfSCEI servers.

GulfSCEI-LINUX is a veritable titan of computational power, boasting the formidable Intel Xeon E5-2630 v3 processor at its heart. With 8 cores, 16 threads, and a turbo boost frequency of up to 3.20 GHz, this server effortlessly handles even the most demanding workloads. Coupled with an astounding 240 GB of RAM, it ensures seamless operation and swift data access, eliminating bottlenecks and enhancing overall performance. But the true marvel lies in its storage capacity. With a staggering 32 TB of local storage, it stands ready to accommodate vast databases, high-definition multimedia, and extensive archives with ease.

  • Intel Xeon E5-2630 v3 @ 2.40GHz with 8 cores (16 threads)
  • 240 GB RAM
  • 32 TB local storage

GulfSCEI Workstations:

At GulfSCEI, we are proud to equip our researchers with the performant Dell Precision 5820 Tower workstations. These machines are powered by the formidable Intel Core i9-10900X processor, which, with its 10 cores, Hyper-Threading technology, and a turbo boost frequency of up to 4.7GHz, ensures exceptional computational performance. The workstations are further enhanced with 32 GB of DDR4 memory, providing ample space for complex datasets and ensuring smooth multitasking. For graphics-intensive tasks, these workstations boast the powerful Nvidia GeForce RTX 3080Ti graphics card, delivering unparalleled visual performance. Storage is no less impressive, with a lightning-fast 512 GB M.2 PCIe NVMe Class 40 Solid State Drive ensuring swift data access and efficient operations. Running on the reliable and secure Ubuntu Linux 20.04, these workstations offer an optimal environment for research and development. In essence, our researchers are equipped with the best-in-class technology, enabling them to push the boundaries of innovation and discovery.

  • Intel Core i9-10900X processor
  • 32 GB of DDR4 memory
  • Nvidia GeForce RTX 3080Ti graphics card
  • 512 GB M.2 PCIe NVMe Class 40 Solid State Drive
  • Ubuntu Linux 20.04

GulfSCEI Robotics

Jackal Robot:

The GulfSCEI center utilizes the Clearpath Robotics Jackal ground vehicle. Jackal is a rugged, lightweight, fast, and easy-to-use unmanned ground vehicle for ROS Noetic designed to conquer diverse terrains and tasks with unparalleled precision and agility. At its core lies a 270-watt-hour lithium battery, providing ample power for extended missions. Whether navigating rugged outdoor landscapes or maneuvering through indoor environments, Jackal’s robust energy source ensures sustained performance and endurance. Controlled by a Sony Bluetooth controller, the Jackal offers seamless remote operation, enabling intuitive control and real-time interaction with the environment.

A GulfSCEI researcher controls the Jackal robot.
Fig. 2.1: A GulfSCEI researcher controls the Jackal robot.

Jackal integrates the Velodyne VLP16 LIDAR, delivering unparalleled 360-degree environmental sensing capabilities. With its high-resolution laser scanning, the Jackal effortlessly maps its surroundings, enabling obstacle detection and navigation in dynamic environments with high accuracy and efficiency. It features an Axis PTZ Camera, providing high-definition imagery for enhanced situational awareness and surveillance capabilities. The Jackal is also equipped with SwiftNav GPS technology, ensuring precise localization and navigation capabilities across diverse environments.

Driving its computational power is the NVIDIA Jetson AGX Orin, a cutting-edge AI computing platform that powers the Jackal’s advanced autonomy and decision-making capabilities. With its high-performance computing architecture, the Jetson AGX Orin enables real-time data processing, machine learning, and autonomous decision-making, empowering the Jackal to adapt and respond intelligently to dynamic environments and evolving mission requirements.

Drone Capabilities

The GulfSCEI Center utilizes the senseFly eBee+ and eBee TAC fixed-wing sUAS systems, some of the best drones on the market for high-resolution aerial photography. We also have access to traditional platforms such as the DJI Mavic 2 Pro for quadcopter operations.

A researcher launches the eBee+.
Fig. 3.1: A researcher launches the eBee+.

All fixed-wing flights are digitally-planned, using the latest eMotion software suite. This means our flights can be conducted in any environment, so long as adequate takeoff and landing vectors are available. The flight-path is adjusted in-software to provide ideal routes and optimal imagery for each mission. A lower flight altitude can be used for better ground sampling distance for critical areas of observation.

Using Pix4D software, extremely accurate mosaics and 3D point clouds are generated for the entire survey area from GPS location, depth inference, and feature detection. Each pixel of these mosaics may be aggregated from upwards of 4 to 5 different raw images. These mosaics are fully orthonormalized to correct distortion and the resulting images are suitable for making high-precision measurements.

The GulfSCEI’s robust data center is equipped to process these images and perform a variety of post-processing, machine learning, and predictive tasks. All GulfSCEI pilots are sUAS-rated airmen certified under FAA 107.

Specifications for each of our drones are listed below:

senseFly (AgEagle) eBee TAC Black
  • Fixed-wing drone.
  • Flight time of 90 minutes.
  • DIU Blue compliant with AES-256 encrypted data link.
  • Cruise speed of 25-28 m.p.h.
  • senseFly GeoBase allows RTK and PPK location correction.
  • Maximum flight range of 55 km (≈ 34 mi).
  • 116 cm (45.7 in) wingspan, 2.2-3.5 lbs at takeoff.
  • Wind resistance up to 28.6 m.p.h.
Payload Capabilities
senseFly S.O.D.A. 3D senseFly S.O.D.A. 3D
  • 1" RGB spectrum camera.
  • Captures both oblique and nadir images for wider FOV.
  • 20 MP resolution (5,472 × 3,648).
senseFly Duet M senseFly Duet M RGB
  • 1" RGB spectrum camera.
  • 20 MP resolution (5,472 × 3,648).
Multispectral
  • 4-Band multispectral sensor.
  • 1.2 MP single-channel resolution (1,280 × 960).
  • ±40 nm Green band (550 nm)
  • ±40 nm Red band (660 nm)
  • ±10 nm Red-Edge band (735 nm)
  • ±40 nm Near Infrared band (790 nm)
senseFly Duet T senseFly Duet T RGB
  • 1" RGB spectrum camera.
  • 20 MP resolution (5,472 × 3,648).
Thermal
  • Thermal sensor.
  • ≈ 1/3 MP resolution (640 × 512).

senseFly (AgEagle) eBee+
  • Fixed-wing 1.1 kg drone.
  • Flight time of 59 minutes.
  • Can document 220 ha (540 ac) in a single 400 ft AGL flight.
Payload Capabilities
senseFly S.O.D.A. senseFly S.O.D.A.
  • RGB spectrum camera.
  • 20 MP resolution (5,472 × 3,648).
  • 2.9 cm/px @ 400 ft. AGL.
senseFly ThermoMAP senseFly ThermoMAP
  • Thermal sensor.
  • Accurate down to 0.1 °C.
  • 14 cm/px resolution @ 250 ft. AGL.
Parrot Sequoia Parrot Sequoia RGB
  • RGB spectrum camera.
  • 16 MP resolution (4,608 × 3,456)
Multispectral
  • 4-Band multispectral sensor.
  • 1.2 MP single-channel resolution (1,280 × 960).
  • 40 nm Green band (550 nm)
  • 40 nm Red band (660 nm)
  • 10 nm Red-Edge band (735 nm)
  • 40 nm Near Infrared band (790 nm)

DJI Mavic 2 Pro
  • Quadcopter drone.
  • Flight time of 31 minutes.
  • Maximum flight range of 18 km (≈ 11 miles).
  • Wind resistance of 29-38 k.p.h.
Payload Capabilities
RGB Camera
  • Capable of videos and still images.
  • 20 MP (5,472 × 3,648) still images.
  • Multiple video formats:
    • 4K (3840 × 2160, ≈ 8.3 MP).
    • 2.7K (2688 × 1512, ≈ 4 MP).
    • FHD (1920 × 1080, ≈ 2 MP).