Real Edge AI
Local object detection on the wearable device, without depending on cloud services for critical safety decisions.
R&D Project / Edge AI / Assistive Mobility
Adaptive AI Mobility Runtime
Wearable assistive navigation system combining LiDAR, computer vision, YOLO, audio guidance, directional LED feedback and live web monitoring for obstacle and fall-risk detection.
LiDARYOLOv8nEdge AIAssistive mobilityRuntime safety
Client / ProgrammeInternal R&D / Horizon-ready concept
Year2026
LocationRomania / EU
StatusWorking demo in real-world testing
Overview
Adaptive AI Mobility Runtime is an edge-AI platform for assistive mobility. The system runs on Raspberry Pi and integrates LiDAR, camera vision, YOLOv8n, proximity sensing and a live web dashboard for real-time monitoring. It is designed for visually impaired users, older adults and rehabilitation or recovery scenarios.
Challenge
People with disabilities or reduced mobility face daily risks from obstacles, stairs, edges, dynamic objects and collision or fall hazards. Simple assistive tools provide limited context, while many digital solutions are not validated in real-world conditions.
Solution and impact
The system detects obstacles and fall-risk situations in real time, evaluates the level of risk and provides audio and visual guidance. The dashboard stores events, incidents, images and runtime data for analysis and validation.
The platform can support independent mobility, accident prevention, rehabilitation support and integration into European projects focused on independent living, disability and community care.
How it works
Real-time assistive navigation, from sensing to guidance.
The wearable edge device reads data from LiDAR, camera vision, motion sensing and distance sensing, runs local perception, sends structured runtime events to the backend and triggers audio or LED guidance whenever a risk is detected.
Runtime safety orchestration
Policy engine, state machine, obstacle memory and real-time risk escalation for safer navigation support.
Live monitoring
Dashboard with live video, incidents, logs, model registry, device health and structured runtime JSON.