Radio
Signals of opportunity
Doppler shifts from LEO constellations such as Iridium, Starlink and OneWeb become a position and timing source that doesn’t depend on GPS.
Any data. Any system. One super intelligence.
Autobots Technology delivers next-generation physical artificial intelligence, signal intelligence and space domain intelligence, today.
Product: Navigate anywhere on Earth or space without GPS using LEO satellite constellation signals.
Solution
Navigation and sensing that work where GPS doesn’t.
GPS is the backbone of modern defense and commerce, and the easiest thing to take away. Jamming and spoofing deny it. Weather, terrain and the plasma sheath around a hypersonic vehicle block it. We build the AI, signal processing and low-SWaP hardware that keep position, timing and awareness alive without continuously needing GPS.
Our thesis
Resilient navigation should be multi-source.
No single technology replaces GPS. Star trackers need a clear sky. Cameras need scenery. Radio needs to be in range. Each one fails somewhere, so we fuse independent physics, and when one goes dark the others hold the solution.
Field
Earth’s magnetic field
Crustal magnetic anomalies form a passive map that no radio jammer can deny. Chip-scale quantum magnetometers make reading it practical on aircraft and vehicles.
Light
Event-based sensing
Sensors that report only per-pixel change, at microsecond latency and can track low contrast hypersonic objects through extreme motion dynamics and solar background radiation.
Many disciplines, one system.
Illustrative · four cores, one fabric
Physical intelligence
NeuroFusion: four neural cores on one hardware
NeuroFusion is a reconfigurable, FPGA-based processor that runs convolutional, spiking, recurrent and attention networks as coordinated cores on a single commercial chip. Modeled on the neocortex’s shared microcircuit, it gives each input the computation it needs, and no more, inside a low-SWaP budget.
- Four cognitive coresCNN, SNN, RNN, attention
- One shared substrateCommercial FPGA fabric
- Resource-aware allocationEasy inputs exit early
- Uncertainty-aware fusionClass + confidence out
Simulation · spiking network
Signal intelligence
Neuromorphic RF recognition
Spiking neural networks compute only when a signal changes. On a space-qualifiable FPGA they can recognize the modulation of a contested RF signal in real time, at a fraction of the power of GPU or DSP pipelines, then hand off to a waveform generator to respond.
- Modulation recognitionLow SNR, under jamming
- Spiking neural networksEvent-driven, sparse
- SIGINT and ELINTUnknown-signal classification
- Waveform responseDDS arbitrary waveform
Where the technology goes to work.
Illustrative · onboard navigation
Air, land and sea
Autonomous Systems
Drones, ground vehicles and maritime platforms that must navigate and perceive on their own. Event-based sensing and onboard learning keep them accurate when GPS, links and cloud compute are gone.
- GPS-denied navigationDoppler, inertial, magnetic field
- Event-based perceptionLow latency, low power
- Onboard learningAdapts at the edge
Illustrative · contested spectrum
Contested spectrum
Electronic Warfare
Neuromorphic processors that recognize unknown emitters as they appear, even under jamming, then hand off to a waveform response, at a fraction of the power of GPU or DSP pipelines.
- Emitter classificationLow SNR, under jamming
- Cognitive responseAdaptive waveforms
- Low-SWaP payloadsReal-time, at the edge
Illustrative · compute in orbit
In orbit
Space Data Centers
Processing data where it is collected avoids downlinking raw streams, but it demands power-frugal, radiation-tolerant silicon. Reconfigurable neural processors are built for that budget.
- On-orbit inferenceProcess before downlink
- Radiation toleranceSpace-qualifiable FPGA
- Power-frugal AISparse, event-driven
Company
A fast growing company focused on intelligence systems for defense and industrial applications.
Dr. Sanjeevi KarriPresident
Ph.D. in Electrical Engineering from the University at Buffalo. More than 20 years in semiconductors at Philips, NXP, IBM, AMD and GlobalFoundries, followed by leading SBIR projects from DARPA, DTRA, NASA, Navy, NGA, DMEA and DOE. Today she leads projects in signal intelligence, AI, cognitive communications, intelligent navigation and advanced microelectronics.
Contact
Working on
Working on
physical artificial intelligence, signal intelligence, and space domain intelligence?
Tell us about the mission, the platform and the constraints. We’ll tell you what the engineering can carry.
Start a conversation