High commuter traffic volumes in urban areas, as well as the high volume of leisure traffic in tourist and excursion areas, lead to congestion of transport infrastructures. This applies above all to access and transit routes and overcrowded parking lots with very high parking pressure and parking search traffic. The lack of information often results in additional traffic volume and thus also pollutant and noise emissions.

An AI-based model is used to predict the occupancy of parking areas. The approach is based on floating car data (FCD, position data that vehicles send while driving)as well as occupancy information from existing sensors in the parking lot. In order to predict the occupancy of the parking spaces, in addition to the main data source of the existing sensors at the parking lot, other different traffic data sources, including digital parking tickets, movement, weather, and on-site information, are also included.

In addition to real-time occupancy information on parking areas, municipalities can use the information on forecast occupancy to enable more efficient use of existing infrastructure, for example, alternative parking spaces can be released or traffic control measures can be used to respond and point out alternatives. Road users are enabled to react according to the situation and make a choice of means of transport before the start of the journey; In this way, they can make the most efficient route and parking space selection on the way to their destination by means of motorised individual transport. The additional information generated by the parking forecasting can be used to increase economic activity as an added value for various stakeholders, such as tourist regions('image' of the tourist destination), local transport operators (additional offer), and local population (less negative externalities). In the long run, entrenched patterns are avoided by taking action early. Overall, a parking space forecast helps to produce less parking search traffic and thus energy savings, fewer air emissions (global: CO2; local: nitrogen oxides, particulate matter), and less noise and to initiate a modal shift in favor of environmentally friendly means of transport. Example calculation: For the following example calculation, the following assumptions are assumed:

Average consumption	7l / 100 km	Occupancy rate of passenger cars	2
Average CO2 emissions	150 g CO2/km	Follow-up costs CO2	180 € / ton
Price fuel	2€ / litre

Example: Forthreeparking areas at ski lifts in Oberstaufen, in addition to occupancy recording in real-time, there is a parking space forecast. The information on occupancy information of the parking areas is imported and visualized via the local database. From the train station in the center ofOberstaufen a ski bus to the ski areas is in use(distance to the lifts 6.7km / 9.3km / 4km). On 25 days a season, the parking space forecast shows that the parking spaces at the ski lifts will be fully occupied from 9:00 am. On this basis, 15% of overnight guests (120 people) who would otherwise have driven to the lift in their own car decide to take the ski bus and not drive their own car to the lift. People use the three available lifts in equal parts. Furthermore, 80-day visitors from the Stuttgart area decide to travel directly by train and use the ski bus from Oberstaufen station instead of traveling by car.

Overnight guests		Day tourists
Fuel savings	Avoided emissions	Fuel savings	Avoided emissions
28l / day	60kg CO2 / day	1232l / day	2,64 tons CO2 / day
700l / season	1.5tons CO2 / season	30.800l / season	66tons CO2 / season
1400€ / season	270€ / season	61.600€ / season	11.880 € / year

By using a parking space forecast, a modal shift can be achieved, which not only contributes to the avoidance of parking search traffic and a reduction of emissions in the region itself but can also immensely increase the economic benefit through the use of public transport. A modal shift towards public transport ensures that this offer becomes more attractive in the long term through additional connections and/or an increase in frequency and can thus also sustainably change commuter and everyday mobility in a region. An increase in the attractiveness of a region then attracts people (center of life, visitors) and strengthens the economic performance of the region.

The merging of data from several parking spaces as well as "external" data sources, such as weather data, mobile phone parking, etc., which are needed for the AI to calculate forecasts. Data can be passed on to other systems via data platforms and visualized, e.g. cockpit or apps, in order to make it available to users and to direct them.

Sensors for occupancy detection at the parking lot (camera, floor sensors, barrier system)

LTE, LoRaWAN or NB-IoT

The following screenshot shows the Datalab "Parking Space Forecast", which shows occupancy forecasts for a P&R parking space in Eschborn. The forecasts are compared here with the actual occupancy at the parking lot. The forecasts of the occupancy status are calculated from historical data as well as real-time data from existing sensors for occupancy recording at the parking lot as well as on the basis of floating car data and other context information.