A two-year program studies how sensor fusion enhances human presence detection and safe human-machine interaction.
ETOY, Switzerland, Oct. 20, 2026 /PRNewswire/ — Robots working beside people need a reliable estimate of human movement. Their controllers also need to know how far they can trust that estimate. Algorized and ZHAW Zurich University of Applied Sciences Institute of Signal Processing and Wireless Communications ISC have launched a two-year research program to study both questions. The collaboration is funded by InnoSuisse.
The program, “Real-Time Human Perception via Edge-AI Wireless Sensor Fusion for Collaborative Robotics”, runs from September 2026 for two years. It builds on earlier research by Algorized and ISC on ultra-wideband (UWB) radar.
The program aims to combine raw wireless and optical signals in one fusion layer between a robot’s sensors and its controller. Researchers aim to weight each input by confidence and produce a calibrated estimate of a person’s position and predicted path.
Algorized processes raw data with one foundation model to detect human presence and classify whether it is a human or an object. The program tests whether combining raw inputs in sensor fusion, before each sensor makes its own decision, creates a holistic layer for the human-aware machines.
Each sensor type adds different information. UWB radar and Wi-Fi sensing can help detect a person behind a shelf or a machine. Millimeter-wave radar holds its lock through dust, steam, glare and darkness. RGB and time-of-flight cameras add spatial and semantic detail. The program studies how these inputs work together.
The research has four goals. The first three are to sense people in approach to the robot, tell living from non-living objects, and track and predict each person’s path. The fourth is to send the robot controller a graded instruction: continue, slow down, or stop. The inference target is under 100 milliseconds on standard industrial edge hardware. Algorized’s planned safety functions cover presence detection, localization and tracking.
A model of this kind needs its own training data. That data must come from real facilities with people moving through them. Researchers who understand the physics must interpret it. The goal is a core model that transfers across sensor types, vendors and sites without retraining for each deployment.
“Machines enter the physical world,” said Natalya Lopareva, CEO and Founder of Algorized. “For robots to coexist with humans, they need to understand the people around them. We are exploring every sensor and every layer of compute to make shared spaces safe. That takes research, and we are doing it together with the ISC at the ZHAW School of Engineering.”
Leave a comment