ABR’s TSP Chips Are Specially Built to Run Our LMUs for Even Greater Efficiency Gains


Process signals more accurately and efficiently. With ABR’s SignalChip, you can do more with less.

Use this chip from our Time Series Processor family to run large AI models faster, at low power, and bring down energy costs. Perform effectively with less data and smaller models, and do your signal processing on-device.

A black smartwatch displaying atrial fibrillation information on its screen.

Get an edge on your competitors

ABR's SignalChip is bringing next-generation digital signal processing and advanced AI functionality to the edge, as your solution to any streaming data problem.

A computer chip
  • Anomaly detection for < 30mW and < 50ms latency
  • Heartbeat R-peak detection < 0.42mW and < 60ms latency
  • Background noise suppression for < 24mW and < 10ms latency
  • High efficiency ABR neural accelerator fabric
  • I/O voltages between 1.6 V and 3.3 V
Put SignalChip on your device, and run advanced digital signal processing locally. Avoid recurring cloud costs, reduce latency, and protect your privacy.

Key features and applications

We offer a full software stack, including AI model design and deployment, with custom and pre-trained networks available.

Signal classification

Categorize inputs based on their characteristics

Anomaly detection

Identify unusual patterns in your signal

Noise reduction

Filter out disruptive inputs

Predictive maintenance

Anticipate need for upkeep and repairs

User success story

ABR has partnered with Myant to do signal processing in the form of R-peak detection for their smart garments, and unlock the potential of textile computing.

"I was incredibly impressed with ABR’s ability to develop efficient lightweight heartbeat detection models for Skiin and at the same time improve our heartbeat accuracy. They have a systematic and scientific approach while being quite flexible in this collaboration with us. We’re very excited to continue our partnership, licensing their LMU algorithm, and applying it to more bio-signal processing as we further develop our technology. Anyone who is building AI at the edge, or has any power, compute, or accuracy concerns, I recommend reaching out to their excellent engineering team."

Tony Chahine
CEO, Myant

A pair of black garments, integrated with smart technology, possibly for monitoring or enhancing user experience.

Talk to our sales team

Learn more about how low power edge computing can support your AI integration.

Contact us