Townsman are constantly moving. Whether going to work or to pick up children from school, mobility is an inherent factor in urban life. However, cities are now congested by traffic that continues to grow and bring its share of inconvenience (traffic jams, pollution …).
According to ITS America, 30% of congestion in urban areas is caused by motorists looking for a parking space. Despite emerging solutions, flow management has never been such an important topic for the future of urban planning. Mobility is proving to be a major issue in the development of the cities of tomorrow, Smart Cities.
Digital is not only revolutionizing transport: mobility in general is impacted. Open data can be the solution to save cities from congestion while providing new services.
Datategy has developed Octocity Parking, an innovative solution based on an Artificial Intelligence predictive algorithm to reduce fraud.
For a transport operator, fraud represents a direct shortfall. For example, fraud represents more than one million euros per day in “Ile de France”, or 1000 buses per year.
For the regulator (the city, the client), fraud represents a lack of information on the actual circulation of users, preventing a match between the deployment of the network and the real needs.
The goal of the DATATEGY solution’s goal is to reduce fraud and its effects. Improving the revenues of the operator and the quality of the service. The goal is also to facilitate verbalization, dematerialize the administrative system by centralizing all administrative services on the same platform.
Network structure *: list of roads, P&D machines , geolocated sensors, GTFS.
Validation *: Access to payment data and place occupation on real-time or periodic extracts.
Optional: Access to verbalization data if existing. If not existing, automatic learning by reinforcement.
The validation data can be in real time or through a periodic extract.
GPS-equipped mobile terminals: Datategy is in partnership with suppliers (Winmate, Bluebird)
The solution guides the enforcers to the right roads at the right schedules.
A mobile application: for agents in the ground, allows to enter a verbalization, visualize the history of verbalization and to have at real time, the list of zones to visit and to guide agents to places where fraud is the highest at the moment T.
A cloud application: for managers. Allows a global monitoring of the situation on the ground on the one hand, but also a statistical vision of the effectiveness of control and mediation units.
And this, according to several axes: by zones, city organisation(market/commercial zone), by commune, by day of the week, by time of day … The mobile application will be compatible Android and the Back Office will be a web application accessible on all popular browsers (Chrome, Firefox, Safari, IE, Vivaldi …).
The technical stack is the JS stack, namely Node.JS for the Back and Angluar for the front office. The data will be on Cassandra in part and on Postgres for transaction data.
A pedestal technological base, allowing data scalability and record execution time. The use of open source allows the removal of hidden costs.
The cloud and mobile solution allows a complete externalization of technical aspects and removes hidden costs (storage, servers, maintenance …)
When collecting data, all sensors are connected to P&D machines, smartphones, agents who patrol the field, the work tool used by tow drivers and user’s smartphones. To enrich the model we will also use related data including maps of the area.
These allow motorists to pay their parking bill, record data dates and times of all people who took a parking ticket. This information will make possible to estimate the number of vehicles occupying “legally” a street and to estimate the number of unoccupied places.
Thanks to the recording of cases of fraud in the field, it is possible for us to improve the driver guidance algorithm while adapting the characteristic indicators of the city (occupancy rate according to the time of day, duration average vehicle parking, fraud rate). The application offers towing agents an optimal route to recover and route poorly parked vehicles at the headland.
Users will also help to collect data via our app on their smartphone. Indeed their GPS position sharing will allow us to trace their route and to make a follow-up that can be compared to the information sent by P&D machines and by field agents. In addition, the user view will allow drivers to confirm whether they have found a parking space on the street indicated. If not, these data will help to improve the calculation of the probability of fraud on the street concerned.
The maps allows us at the initialization of the work, to have an estimation of the maximum occupancy capacity of each street. as one goes along this figure can be changed. For example if there is public work in the street or if connected P&D machines record a filling rate greater than it should at a given time.
This is the intelligence of the system, the seat of all the algorithms that will be applied. We will calculate several indicators allowing a better visualization of what is happening in the field. For example we can calculate:
Once the data processed, they can be disseminated via a dedicated application. And depending on the type of users the data will not be the same. For example :
Indicateurs de performance adaptable dans la vue « statistique ».
Enforcers will receive an optimized street control schedule to patrol the areas where they will be most effective.
The tow drivers will receive an optimal route so that tows require the least amount of travel and the least fuel costs.
Impact on the fraud
Depending on the complexity of the network and the degree of effectiveness of the system already in place, the impact of the solution will vary.
On average, the impact on our anti-fraud solutions (all operators combined: transport and parking) oscillates between an increase in verbalizations of 3 to 8% the first month up to an increase between 12 and 22% of the volume. verbalization on average after 6 months of deployment.
Along with this increase in verbalization, we see a stagnation, or even a decrease in the number of long-term verbalizations, linked to an increase in the number of tickets sold (transformation of repeat offenders into customers).