SmartMowAI

There are around 128.000 km of roads in Austria. The road maintainers have the legal task of keeping the vicinity of roads free of endangering vegetation, an essential and cost-intensive task. Considering both sides of the road this results in about 260.000 mowing kilometres for the Austrian road maintenance services. Furthermore, there is a legal obligation to progressively replace current fossil mowing vehicles by zero-emission vehicles till 2030.

Today, there is no way to get an overview of the current state of the meadows along the entire road network. In order to ensure legal requirements, today this results in regular, high-frequency mowing, even though this is not required on large sections of the road network. Urgent mowing decisions are made by maintenance personnel on the basis of very local observations, resulting in mowing vehicles driving back and forth across the large road network to fulfill these urgent mowing tasks. The lack of information about the meadow state along the entire road network leads to non-optimal use of resources (increased CO2 emissions, working time, number of mowing vehicles, etc.).

The main objective of SmartMowAI is to research and develop technologies which allow to monitor the meadow state along large road networks, to use meadow state information to devise an optimized mowing strategy for the entire road network, and to validate the researched technologies and methods through a proof of concept for optimized mowing on the Styrian road network.

SmartMowAI explores novel adaptive learning and situational AI algorithms for the automated detection of relevant intensive and extensive mowing areas along roads and cycle paths, and for the classification of these mowing areas with respect to their phenological state (the state of development of meadow plants over the entire vegetation period), as well as various approaches for optimising the resource efficiency of the AI algorithms. The data collection required for the AI research within the project, and later for the operational monitoring of the meadow condition, will be carried out by road maintenance vehicles as part of their regular operations, resulting in no additional emissions.

Based on the phenological state of intensive and extensive mowing areas an optimized mowing strategy for the entire road network is devised, resulting in the following benefits:

• Reducing the number of mowing cycles, using fewer vehicles, and avoiding short-term use of mowing vehicles significantly reduces the associated CO2 emissions and working time
• Enable the use of zero-emission mowing vehicles (as their range is limited).
• Increasing the biodiversity of meadows through later and less frequently mowing.
• Ensuring visibility to avoid traffic accidents at crossroads, entrances, bends, areas with high wildlife traffic.
• Leaving traffic signs, delineators, cycle paths, and wildlife warning devices unobstructed.
• Ensuring the road drainage functions.

SmartMowAI results in novel adaptive learning and situational AI methods for the segmentation and state classification of roadside meadows and through that will provide the base for future market-ready solutions needed for the mobility transformation for biodiversity optimized mowing scheduling, enabling the use of zero-emission mowing vehicles in the first place.

The SmartMowAI project builds on the experience and results of FloraMon.