Incubating technology. Accelerating Time to Market.
The T&S Innovation Department is a dedicated R&D structure combining experimental platforms, applied AI, intelligent systems and industrial validation environments.
An innovation model built around incubation
The Innovation Department operates as an integrated R&D structure designed to explore, validate and industrialise emerging technologies.
Our approach combines experimental platforms, software environments, AI capabilities and multidisciplinary teams working on operational industrial challenges. Technologies are developed and tested in real-world conditions before integration into client environments. The objective is to accelerate the transition from experimentation to operational deployment while maintaining a strong connection between innovation, industrial constraints and real usage conditions.
An environment designed for industrial innovation
Dedicated R&D team
Engineers, data scientists, roboticists and PhD students.
Experimental platforms
Three platforms in a 1,500m² facility near Strasbourg.
TRL-driven approach
Technologies validated up to TRL 7 before client integration.
R&D funding
Structure compatible with public R&D funding mechanisms (R&D tax credit, public grants).
.avif)
Three experimental platforms,
one integrated program
Each platform combines dedicated equipment, software infrastructure and specialised expertise to validate technologies in operational environments.
Industry 5.0
Instrumented assembly line with AI-based inspection, process digitalisation and operator augmentation systems.
Automated inspection using computer vision models to detect and classify surface defects on manufactured components.
Digital integration of industrial processes through Manufacturing Execution System connectivity and real-time production data collection.
Sensor-based monitoring, anomaly detection and failure prediction using the SPARQ Predictive platform.
Robotization and automation of industrial production lines.
Real-time gesture recognition, biomechanical analysis, projection-based guidance and
LLM-powered operator chatbot.
Robotics
Inspection robot equipped with multi-sensor perception stacks for unstructured outdoor or industrial environments.
Mapping, localisation and path planning for outdoor and industrial environments using ROS2 and SLAM algorithms.
AI-based scene understanding for unstructured outdoor environments, with annotated datasets and semi-automatic labelling tools.
Person detection, obstacle avoidance, anomaly localisation and scene description running on-board Jetson hardware.
AI-based correction and optimisation of tracked trajectories using stereo vision and IMU data fusion.
Hands-free robot control via natural language commands, using on-board and server-based LLM architectures.
Software Defined Product
Hardware-in-the-loop testing environment combining real-time simulation, sensor integration and digital twin capabilities.
FOC vector control, RPM and torque loops, PWM generation and current monitoring via dSPACE.
Dynamic motor model running in parallel with the physical system, comparing simulation and real behaviour in real time.
Lightweight ML models detecting temperature anomalies, vibration drifts and current asymmetry in real time using SPARQ Predictive.
Controlled fault injection for ISO 26262 validation, degraded mode testing and functional safety assessment.
Dynamic parameter and control logic updates without stopping the motor, demonstrating software-defined flexibility.
Artificial Intelligence as a
transversal layer
AI is embedded across all three platforms as a set of reusable building blocks, each validated in real experimental conditions before being deployed in client projects.
Computer vision
Object detection, semantic segmentation, 3D inspection, person tracking and scene understanding across all platforms.
Synthetic dataset generation
Automated generation of realistic, labelled training data from 3D models and CAD files, reducing annotation cost and bias.
LLM and NLP integration
Natural language interfaces for operator assistance and robot control, using both cloud and embedded LLM architectures.
Predictive maintenance AI
Anomaly detection, remaining useful life estimation and failure prediction integrated in the SPARQ Predictive platform.
Three types of partnerships,
one integrated innovation model
The Innovation Department operates within a structured ecosystem combining academic research, institutional support and industrial co-development.
Research laboratories and universities
Public funding bodies and national programmes
Co-development and equipment integration for POC
From experimental platform to industrial deployment
Technologies developed on the Department's platforms have been transferred to industrial clients through co-development or direct integration projects. Content being finalised.
Research validated through academic publication
The Innovation Department's work is anchored in scientific rigour, disseminated through peer-reviewed journals, international conferences and PhD dissertations.
PhD programs &
CIFRE theses
- "Towards Industry 5.0: Design of an assistance system for optimising attentional resources and the sustainable inclusion of operators in the assembly process" — Julian Salazar, ICube, 2024 (ANRT)
- "Adaptive autonomous navigation in unstructured environments: Artificial intelligence and optimised decision-making for the traversability of autonomous ground vehicles" — Titouan Leost, DRIVE, 2025 (ANRT-AID)
Several active PhD programs co-supervised with research laboratories, covering autonomous mobility, off-road perception, AI and embedded systems.
Publications & conferences
Research results are regularly published in peer-reviewed journals and presented at international conferences in robotics, computer vision, AI and embedded systems.
Perspectives on industrial transformation.
Insights on automation, connected manufacturing, industrial operations and operational performance across manufacturing ecosystems.
.jpg)
Frequently asked questions
