PHYS-EXO Physical Embodied Intelligence for Next-generation Smart Exoskeletons
- Coordinator
- Scuola Universitaria Professionale Della Svizzera Italiana, Switzerland
- Runs
- 1 September 2026 to 31 May 2027
- Status
- Running
- EU funding
- €300k
- Total cost
- —
- Partners
- 1 organisations in 1 countries
- Type
- European Innovation Council
- Grant agreement
- 101326967
What the project is doing
In the project's own words, from its summary on CORDIS.
PHYS-EXO proposes a new generation of Embodied Physical AI exoskeletons that move beyond today’s reactive, context-blind wearable robots toward situationally aware, proactive, and human-centric physical assistants. Current industrial and rescue exoskeletons largely rely on posture or motion signals, lacking a real understanding of surrounding hazards, tasks, and spatial context. This limits their effectiveness in safety-critical, unstructured environments where perception, reasoning, and seamless human–AI collaboration are essential.
PHYS-EXO introduces a modular Agentic AI-ready Physical AI pipeline—including Perception, Semantic Mapping (SLAM), Risk Reasoning, Task Planning, and Human-centric Control—embedded in a wearable platform operating in real time and on the edge. The system integrates egocentric multimodal perception, 3D environment reconstruction, task planning and reasoning, and adaptive user-centric control to provide context-aware physical assistance, risk-aware navigation support, and explainable decision guidance to the user.
The project targets two high-impact application areas: (1) ""Disaster response and civil security, through an Urban Search and Rescue"" scenario in a simulated collapsed building, and (2) ""Industrial manual handling"" in a dynamic factory environment with moving robots, restricted zones, and heavy lifting tasks. In both cases, PHYS-EXO aims to reduce physical strain, improve safety, enhance situational awareness, and increase user trust and operational effectiveness.
During Phase 1, PHYS-EXO will extend the existing (TRL 4) SUPSI-IDSIA vision-enabled active back-support exoskeleton with egocentric depth perception, preliminary semantic SLAM, a first-generation AI reasoning layer, and an integrated perception–control loop, reaching TRL 5 through validation in the two above-described relevant environments. Clear KPIs will assess perception accuracy, safe planning, ergonomic benefits, rescue effectiveness, and industrial productivity.""
Topics
Who is involved
| Organisation | City | Country | Role |
|---|---|---|---|
| Scuola Universitaria Professionale Della Svizzera Italiana | Manno | Switzerland | Coordinator |
Source: CORDIS, European Commission. Data of 5 Nov 2026, reused under the Commission's reuse policy. View on CORDIS · Grant DOI