The parking distribution function module can be used for various traffic-related tasks such as the continuous monitoring of existing parking infrastructure or when planning new parking infrastructure, as well as for the management of parking traffic at major events. In addition, the analysis results for parking distribution can be used as an important input parameter for a data-based location analysis in the retail sector or for measuring the attractiveness of properties.

Decisive added values of the parking distribution function module are:

• Reducing the search for a parking space:By analysing parking traffic data, urban decision-makers and specialist users (e.g. traffic planners) can gain insights into the availability of parking spaces and develop solutions to minimise the search for parking spaces for drivers.
  
  
• Improve traffic flow: Urban decision-makers can analyse traffic data to identify bottlenecks and bottlenecks in traffic flow and take measures to improve traffic flow, for example by suggesting alternative routes or optimising signal control.
  
  
• Reduce environmental impact:By reducing search traffic and congestion, data-based parking traffic analysis can help to reduce the environmental impact of exhaust fumes and noise.
  
  
• Support planning and development:Analysing parking traffic data can help clients make informed urban planning and development decisions by providing insights into parking usage and traffic patterns.

A Kepler-based web application for the spatial representation of parking distribution analysis data and filter options for different time periods is provided as a specialised application.
The figure below shows an interactive map for visualising the parking distribution

Areas with high parking volumes are identified on the basis of floating car data (vehicle movements measured by navigation devices and / or mobile devices) and are made available for motorised private transport. A detailed description of the data fields included can be found below.

The analysis covers the period of the last 15 weeks prior to the order. Spatial cells are differentiated with an edge length depending on the size of the area under consideration (100 m-10 km), as well as a differentiation between weekday / weekend and four different traffic-relevant day phases as time slices.

The parking processes are recognised and classified as such in the floating car data (FCD) of a region with the help of a machine learning model. For this purpose, transitions in which a vehicle falls below a certain minimum speed are searched for in each sequence of telemetry data. For each of these transitions, various properties are extracted from the driving behaviour of the previous minute and the two following minutes. These properties are then used as a basis for the machine learning model to make a classification decision (parking process / stop at traffic lights or stationary traffic).

The following data fields are provided as results of the analysis:

• lat / lon: The latitude and longitude of the centre of a spatial cell in the EPSG:4326 (GPS / WGS84) coordinate system.
  
  
• relative_parking: The relative number of parking events in the spatial cell. Normalised for the duration of the time slice in order to establish comparability between time slices.
  
  
• day: The day or the day span for which a value was collected.
  
  
• hour: The hour or the time span for which a value was collected

No further data sources are required for this function module, i.e. no connection of connectors. Floating Car Data (FCD), which is used by [ui!] to carry out the analyses, is used as the data source.

The parking space monitoring and forecasting function module provides local authorities with a powerful tool that enables them to monitor the use and availability of parking spaces in real time and predict future developments.

An overview of the added value of this function module:

• More efficient parking space management: By continuously recording and analysing occupancy data, local authorities can closely monitor the use of parking spaces and react quickly if necessary. Measures such as the dynamic adjustment of parking charges or the introduction of parking guidance systems can be implemented in a targeted manner. Early detection of bottlenecks and targeted traffic management can make it easier to find a parking space at major events, for example.
  
  
• Optimised planning of parking infrastructure: With data-based insights, municipalities can identify areas with high demand and make needs-based decisions, such as the placement of new parking spaces and the improvement of existing facilities. Forecasts of future demand enable proactive planning. Historical occupancy data as well as external contextual information such as weather data are included in the calculation for the forecasts. By utilising the data that is added over the term, the forecasts are continually adapted and improved.
  
  
• Combining recording technologies: The use of different recording technologies such as camera and ground sensor systems offers local authorities a flexible and comprehensive data basis. This diversity makes it possible to collect data on car park utilisation at different levels of granularity depending on local requirements. This promotes long-term, well-founded decisions and supports effective car park management.

The specialised application of the parking space monitoring and forecasting function module is provided as a [ui!] DATALAB. The illustration shows four visualisations of different aspects of parking space monitoring and forecasting.

Depending on the characteristics of your car parks, different technical sensor or camera-based solutions are used in the parking space monitoring function module. These are not part of the [ui!] PARKING solution and are assumed to be an existing installation for the car park monitoring function module. If no sensor technology has yet been installed, this can be ordered, for example, via the Smart City marketplace [ui!] AGORA (https://agora.umi.city) be obtained.

Three different types of sensor technology for car park monitoring are presented below.

• With the help of a camera system or ground sensor system at the entrances/exits of a car park, a balance of the currently occupied parking spaces can be drawn up. Each entry of a car is added, each exit is subtracted.

Balancing camera system for car park monitoring

• Overhead sensors (mounted on light poles or integrated into LED lights, for example) and an AI-supported computer vision system recognise free parking spaces and forward them as data records. This solution is particularly suitable for large, open car parks with unclear access situations.

Overhead camera system for recording the occupancy of car parks 

• Floor sensors (embedded or mounted on parking space surfaces) are used to precisely detect whether a parking space is occupied. In further expansion stages, active car park management can also be implemented. For this purpose, special rights (e.g. residents' parking) are stored on NFC chip cards; these cards can then be placed in the glove compartment, for example. The floor sensors detect whether the vehicle is equipped with an authorisation and issue an alarm at a central location if this is not the case.

Floor sensor for recording the occupancy of car parks