close

Palantir POC - Case Study

Synopsis

Xailient specializes in extremely accurate, low-power computer vision on Edge devices. Palantir Technologies builds and deploys software that allows organizations to integrate their data, decisions, and operations effectively.

Palantir Technologies’ Meta-Constellation harnesses satellite technology at the Edge by empowering autonomous decision-making in low-bandwidth, low-power conditions. Palantir enables stakeholders to monitor regions of interest using satellites from multiple providers, maximizing coverage.

For example, Meta-Constellation enables continuous monitoring for submarines by handing off between satellites and providers as they orbit over the region.

Palantir’s system increases revenue for satellite operators by enabling multiple tenants to subscribe to imagery – maximizing utilization.

By deploying Xailient’s patented Detectum™ detection technology with Palantir’s Edge AI Meta-Constellation system, operators can operationalize AI insights to make smarter, faster decisions across domains, from space to seaports.

Palantir’s Meta-Constellation integrates with existing satellites, optimizing hundreds of orbital sensors and AI models. This allows users to ask time-sensitive questions across the entire planet. Essential questions such as:

  • Where are wildfire indicators located?
  • How are climate changes affecting crop productivity?
  • When and where are naval fleets conducting operations?

Xailient’s low-power Detectum™ technology was deployed into orbit by Palantir to detect submarines instantly. By deploying Edge AI to existing satellites using Xailient’s patented Detectum™ technology, submarines can be instantly detected – all in under one second.

Key Outcomes

1Xailient helps Palantir’s Meta-Constellation maximize coverage and revenue by enabling multiple tenants to subscribe to imagery.

2Xailient provides Palantir with real-time imagery analysis that can run onboard the satellite.

3Onboard AI allows for automated re-tasking between tenants - continuously monitoring the field of view and activating different workflows depending on what the satellite observes. Only Xailient’s computer vision AI technology meets the size and power requirements necessary to perform these tasks.

Problem Statement

Submarines present a threat to the US and its allies. To protect strategic interests, Allied forces need to track every submarine’s deployment worldwide. Anti-submarine warfare officers use Meta-Constellation to monitor ports across the Pacific.

 

The challenge is, to successfully deploy computer vision AI into space, the AI must be small, light-weight, and able to run on little power.

Activity

1

Meta-Constellation dynamically determines which orbiting sensors are available in response to allied monitoring requests. Integrated through Palantir, the constellations then collaboratively schedule coverage over each port.

2

Once the strategy is in place, satellites are automatically reconfigured into the correct orbit. And, as a software payload onboard, Palantir’s Edge AI platform connects complex satellite subsystems to models, integrating new AI with the hardware.

3

Xailient’s computer vision AI, running onboard the satellites, can process imagery, detect submarines, geolocate them, and determine any movement since the last detection pass – all in under a second.

4

When Xailient’s AI sees submarine movement, that insight is directly downlinked to allied forces as the satellite passes overhead. The anti-submarine warfare officers are notified in just minutes, empowering them to respond in a way never seen before.

Results

By deploying Edge AI onto existing satellites and using Xailient’s low-power Detectum™ technology, analysts can operationalize AI insights that require minimum size, weight, and power.

Trainable AI from Xailient, for marine vessel detection, spots submarines (or illegal fishing vessels) in under one second – solving a critical problem and improving intelligence.

For Palantir, Xailient demonstrated an ability to use open-source satellite imagery to detect marine vessels within extremely constrained systems and to do so fast enough to allow for dynamic retasking of the satellite.

The upload limitation for deployment and update was 10MB. Other models couldn’t be deployed because of their size, but Xailient was able to upload with less than 2MB, and this package size continues to improve.

Continuously operating AI consumes power and generates heat, which can be challenging to radiate in the near-vacuum of space. Many AI models must be turned on and off, allowing heat to dissipate. A 50% duty cycle is a common limitation. Xailient’s low-power AI generates less heat and can operate continuously with a 100% duty cycle.

Xailient was able to have its computer vision AI technology ready for POC in a short 48-hour period. The POC used unclassified (open source) training data and demonstrated that Xailient’s computer vision AI can be trained to many different “classes” of objects.

Next Steps

Xailient and Palantir technologies combined can answer essential questions as Meta-Constellation pushes Palantir’s Edge AI technology to a new frontier.

We are excited about Palantir’s Meta-Constellation as the infrastructure software brings hundreds of satellites to bear on some of the most complex problems.

Whether it’s the anti-submarine warfare officer discussed earlier using AI to find submarines, or first responders leveraging AI to spot wildfire signs, Xailient is there to empower users.

Discussion

Xailient’s Computer Vision AI technology benefits Palantir by offering:

  • Computer Vision that fits inside systems allows for imagery analysis at the source.
  • The ability to analyze data in orbit to reduce downlink bandwidth use and provide autonomous control when out of range of ground stations.
  • The capacity to run continuously (on low power) to inspect all incoming imagery.

If a region of the South China Sea, for example, has what looks like a submarine, Xailient then activates the software package that services the DoD.

In another region of the South China Sea, the satellite may observe fishing vessels. This activates a software package that looks for illegal fishing activity.

Xailient Detectum serves as a ‘visual multitasker,’ activating the correct software application based on what is observed in the field of view. This, combined with Palantir’s Meta-Constellation, maximizes the use of satellites while simultaneously providing more coverage of a region. Satellite providers improve their revenue potential, and stakeholders obtain better intelligence.

*Images are the copyright of Palantir Technologies

Press releases

March 14, 2022

Xailient specializes in extremely efficient low-power computer vision. Intel's OpenVINO specializes in maximizing the performance and speed of computer vision AI workloads. OpenVINO improved Xailient FPS 9.5x on Intel hardware to 448 FPS. Together, Xailient-Intel outperforms the comparable MobileNet_SSD by 80x. Even after Intel worked the OpenVINO magic on MobileNet_SSD, Xailient-OpenVINO is 14x faster.

November 29, 2021

Xailient’s Face Recognition enables high-speed edge AI processing with low-power consumption using Sony’s IMX500 – a chip so small it can fit on the tip of your finger.

Explore our blogs

We see things differently in the dynamic field of computer vision AI

Get started with Xailient

We empower companies to bring computer vision AI products to market
faster and with less investment

OnEdge Newsletter

A weekly newsletter with the best hand-picked resources about Edge AI and Computer Vision

OnEdge is a free weekly newsletter that keeps you ahead of the curve on low-powered Edge devices and computer vision AI.

 

You’ll get insights and resources into:

  • Edge computing use cases.
  • Market trends in Edge computing.
  • Computer Vision AI at the Edge.
  • Machine learning at the Edge.
Cookie Policy