MountAIn – The digital mountain for testing AI-based innovations for alpine outdoor activities.
Alpine innovation lab for AI-based innovations in alpine outdoor sports with analysis of human movements in mountain scenarios in real environments.
The objective of the project “MountAIn” is the development and prototypical demonstrations of a technically and economically sustainable concept of a test environment for AI-based innovations in alpine outdoor sports. The main purpose of the test environment is to provide a mobile digital test infrastructure and suitable test tracks on which a new generation of AI-based systems for analysing human movements in mountain scenarios can be tried and tested in a real environment. This provides the necessary basis for developing new products and services for various professional or general outdoor sports activities, either to improve health and/or individual performance. The test environment will be available for innovation projects and test scenarios of multiple stakeholders in an AI innovation ecosystem such as sports equipment manufacturers, technology providers, mountain lift operators, service providers, sports research institutes, tourism regions and the like.
MountAIn aims to develop the concept for a sustainable test environment for AI-based human motion analytics systems in mountain scenarios in the province of Salzburg and neighbouring mountain regions (Tyrol and Styria). The main objective of the test environment is the concept development for a mobile test infrastructure on suitable test tracks (physical test infrastructure) for different test scenarios of alpine outdoor sports activities, either related to performance or health improvement of individuals in professional and recreational scenarios. The test environment should be available in the future for innovation projects and test scenarios of stakeholders such as sports equipment manufacturers, technology providers, mountain lift operators, service providers, research institutes, etc., with the following sub-objectives:
- Non-discriminatory access to physical and digital infrastructures for different test scenarios of AI-based human motion analytics systems of different technology readiness levels (from labs with in-field experience to real-world environments);
- Creation of physical infrastructures for testing the effectiveness, validity and reliability of AI-based human motion analytics systems under real-world environmental conditions (e.g. weather conditions, slopes, undergrounds, etc.);
- Creation of mobile digital infrastructures for the development of (future) sports, fitness & health services;
- Use of the test environment for easy implementation of demonstration and evaluation projects (e.g. tests of sports equipment manufacturers, usability tests with customers, tests of new services) with continuous user involvement.
Based on these overarching goals, the following three objectives are defined for MountAIn:
- Definition of a strategy & conception for the test environment (Goal 1).
- Definition of test scenarios (goal 2)
- Definition of the physical and digital infrastructures for the test environment (Goal 3).
With the set objectives 1 to 3, MountAIn aims to realise a proof of concept of the planned test infrastructure within a realistic outdoor environment. The proof of concept demonstrator consists of a reusable data pipeline for soft real-time data transmission, an intelligent Body Sensor Network (BSN) and interfaces to the AI system to enable testing and prototyping of an interactive AI-based human movement analysis system. Such a real-time data infrastructure ensures a Quality of Service sufficient to meet the requirements of concurrent feedback systems. The intelligent on-body sensor network provides personalised sensor data of the end user and is capable of providing concurrent biofeedback. This feedback is provided by an interactive AI that analyses the sensor data in real time. In parallel, experts observe the user and receive the AI’s enhanced coaching output. Based on their expert knowledge, the feedback model can be validated or corrected.