AMPLIF-AI - PROJECT DESCRIPTION
The territory of South Tyrol hosts numerous infrastructures that require continuous monitoring to ensure their operability, efficiency and safety for the community. These installations are of fundamental importance for the sectors of transport, tourism, energy generation and transport, and telecommunications. The AMPLIF-AI (Adaptive Mission Planning for LIve inFrAstructures Inspection) project aims to create a system based on artificial intelligence that will enable inspection teams to assess the condition of infrastructures in real time in various application cases in the Alpine context. The partners will develop a system consisting of a ground platform and unmanned aerial vehicles (UAVs) supervised by the inspection team. The UAV will host on-board intelligence that will manage navigation and localisation sensors and case study-specific sensors. The ground platform will be a mobile base and will act as a communication relay with the UAVs. It will host high-performance computing processors to create the three-dimensional model of the infrastructure under inspection and enrich it in real time with the images transmitted by the UAVs. The model may be merged with data available from other sources. The project includes a final validation of a case study.
The AMPLIF-AI project is carried out together with the partners:
AMPLIF-AI - Research activities
The main objective of the AMPLIF-AI project is the automation and streamlining of infrastructure inspection operations (e.g. bridges
power lines, ski lifts, ...). In particular, the project will focus on the realisation of a hardware/software platform for the
real time monitoring through the use of (semi-)autonomous unmanned robots, constrained by the boundary conditions typical of the
Alpine regions.
The platform will integrate new solutions for adaptive trajectory planning during the data acquisition phase and innovative Artificial Intelligence-based methods for processing them (e.g. to find faults in the infrastructure).
More specifically, the objectives of AMPLIF-AI can be summarised in seven points:
- Develop new methodologies to model inspection activities, ensure that the monitoring team acquires the necessary information to support decisions;
- Develop new algorithms for the shared inspection and coordinated planning of air and ground platforms in order to
- mitigate risks during site inspection and to make surveys accurate and complete.
- Make inspection more efficient in terms of emissions and timing.
- Develop innovative algorithms for in-situ analysis and storage of inspection data through Machine Learning methodologies;
- Validate and evaluate the developed approach through in-situ experiments in order to fulfil the above objectives;
- Integrate research results with academic courses in robotics and inspection and in doctoral programmes;
- Disseminate results by contributing to international conferences and/or publications in Scopus-indexed journals, organising workshops and seminars;
- Facilitate collaboration between companies and research institutes, technology transfer and cross-fertilisation by sharing collected data on OPEN digital platforms such as the OpenDataHub of NOI Techpark and by establishing an advisory board to capture the needs of the area and the most critical use cases.
Expected outcomes
- Quality of measurements: differences in information of the state of the infrastructure obtained from the system developed in AMPLIF-AI and traditional methods will be evaluated
- Algorithm reliability: semi-autonomous navigation in the monitoring scenario; acquisition of key points for infrastructure monitoring
- Experimental validation: monitoring missions will be carried out to validate the developed system
- End User Involvement: different end users will be involved for the identified case studies
- Monitoring efficiency: a reduction in the time spent by resources on monitoring activities is expected
- Effectiveness of monitoring: additional information content will be acquired compared to traditional monitoring
- Usability of the system: an evaluation questionnaire will be filled in by the end users, requesting feedback on the usability of the developed system
Project details
Project details:
Project name: AMPLIF-AI
Projecd code: EFRE-1019
CUP: B57H23003920007
Project budget: € 619.000,78 (co-financed: € 529.758,21)
Funding programme: ERDF 2021-2027
Project partner: MAVTech Srl., Fraunhofer Italia, Free University of Bolzano - Bozen
Project duration: 01.2024 – 12.2026