[CASE STUDY]
Taking into account traffic data for route optimization
For logisticians, road traffic is a difficult variable to predict and control, with direct impacts on their organization and quality of service (especially in terms of delivery times and failure to meet delivery slots). Road traffic is also a major source of stress for drivers, deteriorating the quality of life at work in an industry already known for being difficult and unappealing.
However, the digitalization of transport and the technological solutions that go with it can be a precious help for these professionals. Real-time traffic information, route planning, real-time route calculation, ETA (Estimated Time of Arrival) calculation and updates with customer alerts: traffic data is widely used by software publishers to meet the needs of carriers.
Kardinal provides a route optimization solution based on data from technology partner HERE Technologies and integrating traffic into its calculations to provide logistics companies with increasingly accurate and efficient planning.
Where does the traffic data used in the software solutions come from?
This is one of the main questions our customers ask about our route optimization solution, reflecting the importance of traffic on their business.
In order to provide traffic-smart answers, software publishers like Kardinal access traffic data through partnerships with map data providers. But where does the traffic information provided by these partners come from and how is it compiled?
The traffic data provided by HERE Technologies comes from fleets of vehicles using the “Floating Car Data” (FCD) method, which is based on collecting location, speed and direction data from vehicles that can compile this data (or that have a device capable of doing so: GPS tracking, mobile apps, geolocation devices).
Our partner, HERE Technologies, is the leading provider of location-based data and services. It was acquired in 2015 by the German automaker consortium Daimler, BMW and Audi and is powered by the data of millions of private vehicles.
In order to create and keep its maps up to date, hundreds of HERE LIDAR-equipped cars travel the world, each recording 1 TB of data per week. In addition to these cars, there are over 80,000 additional data feeds, including sensors from connected vehicles.
What does Kardinal do with this data? What are the benefits of our route optimization solution?
The traffic data collected can be used:
- for analysis after collection, as traffic models
- in real-time
Kardinal’s route optimization solution exploits traffic data in its different aspects:
Before vehicles go on the road, during planning
At Kardinal, we have worked hard to design a route planning system that accurately incorporates the predictive traffic data available through Here models.
Our team of Operations Research mathematicians has designed our route optimization algorithms to take into account traffic patterns. Our algorithms are thus able to estimate the duration of each trip before the optimization process begins. This has no significant impact on our calculation times.
Kardinal’s route optimization solution provides more realistic results, of course, but also more efficient ones. The algorithms are able to make smarter optimization choices, avoiding recurring traffic jams in the morning and evening, for example.
In real time, once the vehicles are on the road
Kardinal’s route optimization solution is unique in that it continues to perform optimizations even once the vehicles are on the road. Thanks to data from the drivers’ mobile app, Kardinal continues to optimize the routes underway in real time.
This feature from our software allows us to take into account a large number of unexpected events that may occur in the field (such as an unavailable customer, for example), but also, and above all, those related to the traffic itself. Indeed, a traffic jam resulting from a traffic accident cannot be predicted before calculating the routes, so it is necessary to take it into account on the go, using the real-time traffic data sent by Here.
Thanks to this system, our solution can update the ETAs for each location based on real-time traffic, but it can also suggest “better solutions” based on changing traffic conditions.
This ensures that the routes proposed by our solution remain efficient and realistic while remaining reliable in the event of unexpected events in the field.
After operations, for continuous improvement
Kardinal’s Research & Development team is currently working on Machine Learning models to combine generic traffic models provided by our partner Here with field data collected by the mobile apps our customers use.
This allows the generic models to be adapted individually for each customer according to the nature of their business. The data from the field is then used for continuous improvement to refine the route durations to become increasingly accurate day after day.
Real-time route optimization solution
Conclusion
The last mile delivery business is heavily impacted by road traffic, its uncertainties and variations. As a software publisher, it is impossible to tackle planning issues dedicated to urban logistics without integrating traffic as a fundamental element in route optimization calculations.
As a result, having highly accurate location-based data and services has become a fundamental element in the performance of many organizations.
At Kardinal, our commitment to “continuous” optimization requires us to take into account traffic prior to calculation, in real time, and then afterwards using machine learning. For all these steps, we have chosen to rely on HERE’s services, which allows us to have accurate and reliable data that best reflects the current and future road context.