Algorized and AKM Partner to Deliver Edge AI Human Sensing Platform Powered by mmWave Radar

Image Courtesy: Public Domain

Algorized, the Physical AI company building the edge-native nervous system for machines operating around people, and Asahi Kasei Microdevices (AKM), the semiconductor and sensor solutions arm of the Asahi Kasei Group,  announced a partnership to integrate Algorized's edge-AI foundation model with AKM's AK581xAIM millimeter-wave radar (mmWave) Antenna-in-Module (AiM) preserving people sensing to elderly care monitoring and industrial safety applications in Japan and global markets.

The Algorized foundation model processes raw data from AKM's AK5816AIM / AK5818AIM mmWave radar at the edge. It outputs  sub-millimeter chest movements to capture breathing and heart rate; tracks posture across states and distinguishes human presence from background activity. The radar perceives the world through dust, darkness, and clutter. The foundation model perceives the human: standing or fallen, resting or in distress, entering or leaving, approaching a robot or walking past.

The Shift: Awareness, Not Surveillance. Camera-based perception is fragile in the conditions that matter most. It fails in low light, gets blocked by occlusion, and captures identities even when it has no need to. It sees pixels, it cannot sense breath.

The Algorized Predictive Safety Engine runs entirely at the edge, using physics to digitize the environment. Three capabilities, demonstrated as part of this partnership:

1. Awareness on the Move. AMRs and forklifts operate in fast-moving environments where perception moves with the platform itself. The Algorized edge-AI foundation model combines platform motion handling and human tracking into a single model utilizing the AK581xAIM with sub-100ms latency at the edge.
2. Counting Without Cameras. Designed for crowded, high-traffic spaces, the AK581xAIM uses mmWave radar and the Algorized edge-AI foundation model to count and track people moving through doorways in real time. The same sensing intelligence powering mobile robotics and doorway analytics running entirely on the edge with privacy preserved through radio physics.
3. Knowing Without Watching. Continuous care environments require reliable awareness of posture, presence, and vital signs from a single sensing system. The Algorized edge-AI foundation model delivers all three using a single AK581xAIM mmWave radar, detecting whether a person is sitting, standing, walking, or lying down, while also monitoring heart rate during rest.

All three capabilities run on one foundation model, on the same module. Today's available market solutions deliver application-specific products: a people-counter device, a bed-occupancy sensor, a factory safety system - each with its own model, its own training pipeline, its own integration path. Algorized ships the infrastructure underneath. The same model that watches over a care home in Osaka watches over a robot cell in Detroit and a doorway in Düsseldorf. Edge-native. Sub-100ms. Privacy preserved by radio physics, not by policy.

"We are moving from sensors that watch to sensors that understand," said Natalya Lopareva, CEO and Co-Founder of Algorized. "Algorized senses human life - heartbeat, breathing, posture - from raw physics, directly from radio waves, with no cameras, no cloud, and no record of who a person is. With Asahi Kasei Microdevices, that capability now ships on production-level silicon engineered in Japan for the customers building the next generation of safer, more responsive systems worldwide."

"The AK581xAIM was engineered for environments where privacy and reliability are non-negotiable: care facilities, factory floors, doorways. The Algorized foundation model is the perception layer that completes the picture - together we deliver a sensing platform that reads human life, not just motion. That is the standard we are setting for intelligent sensor," said Gregg Rouse, President of AKM's North American business.