Deutsche Telekom organized the #AIHack4Mobility on March 18th and 19th. After the success with #AIHack4Diversity in 2019 (link) and #AIHack4Ladies 2018, this year everything revolved around the topic of “Future Mobility” under the motto “Artificial Intelligence for Diversity Teams”. The partners are Telekom Innovation Laboratories, Telekom Mobility Solutions, Telekom Code + Design and the smart mobility provider Ioki.
The issue of sustainable mobility is a moving one – not least for Deutsche Telekom in times of a changing world of work. After all, mobility, especially in a professional context, must adapt to the working environment, which can change daily, and at the same time be as cost-efficient, comfortable and environmentally friendly as possible (see „Stuck in Traffic“, link). The goal of this year’s Hackathon #AIHack4Mobility is to use machine learning algorithms to understand the specific demands that the increasingly dynamic world of work already places on different mobility service providers.
There is prize money for the first, second and third place winners of the hackathon with 5,000 Euro for the first place, and with 3,000 and 2,000 Euro for the second and third place. In addition to the financial reward, the proximity and possibilities of hubraum, the technology incubator of Deutsche Telekom (link), are also attractive. Hubraum, with its locations in Berlin, Kraków and Tel Aviv, has been an active member of the start-up scene and digital ecosystems since 2012. E-scooter provider Tier, the European market leader, is one of the most successful hubraum exports, founded in 2018 by Lawrence Leuschner (CEO), Julian Blessin (CPO) and Matthias Laug (CTO) (see “Intermodal Simulation”, link).
This year the hackathon was again hosted on the Telekom Data Intelligence Hub (DIH) platform. The DIH is a data hub with an attached AI workspace (Artificial Intelligence), everything from the cloud and as a platform-as-a-service (PaaS) offer (link). It relies on open standards and offers hackathon participants a browser based Jupyter notebook with a curated selection of Python libraries for the fast, scalable development of algorithms.
According to Wikipedia, the term Intermodal passenger transport refers to “the use of different means of transport along a route” (Wikipedia on ” Intermodal passenger transport,” accessed 2020, July 23). Intermodal transport is interesting in that it provides more flexibility on the way from A to B. In particular, technologies for the last mile, e.g. e-scooters, have been very positively perceived, as they can “be an ideal complement to bus and train for the last few kilometers to the destination and thus make local public transport more attractive” (Achim Berg, President of Bitkom, the Digital Association of Germany, Link, 2019).
But despite these highly visible transport options for the last segment of travel to the destination, no intermodal service has yet been developed. And that raises some questions: Why not? What are we still waiting for? Why is it not possible to leverage the potential as a feeder and enabler for intermodal travel?
The data scientists at the Telekom Data Intelligence Hub have succeeded in proving that the “smart”, i.e. intelligent, linking of means of transport can shorten travel times for end users – initially based on a fictitious location with average values (#4 Intermodal Simulation, link). This model was then applied to the city of Berlin at specific times of day and with routes that are crucial for urban traffic, and the results confirm the trend quantitatively (#5 Berlin Digital Twin, link). With the help of Mobility AnalyticX Innovation, the data scientist experts succeed as pioneers to launch a real experiment instead of simulations in the area of Mobility Data Spaces.
Deutsche Telekom’s Data Intelligence Hub is building a first data space demonstrator for the 2021 ITS World Congress in October as part of the real-world laboratory Hamburg initiated by the NPM (National Platform Future of Mobility) (link). It shows how the new data standard IDS (International Data Spaces) can help to enable innovative, sustainable mobility offerings nationwide, which require the interaction of different mobility service providers and platforms.
The DIH introduced itself at 12.55 pm in the slot “Mobility Data Space with IDS using the example of Intermodal Travel” and the audience got first exciting insights into the economic potential of data for intermodal travel using the example of the Real Laboratory of the Hamburger Hochbahn (NPM press release, link in German).
The mobility analytics experts and data scientists at the Telekom Data Intelligence Hub have succeeded in proving with a simulation experiment that the “smart”, i.e. intelligent, linking of means of transport generates actual end user benefits and could shorten travel times (see “Intermodal Simulation”, link and “Berlin Digital Twin”, link).
According to Federal Minister of Transport Andreas Scheuer, leading German companies can use the possibilities offered by the RealLabHH “to use the digitalization push for our mobility and make innovative new offers tangible for people [gain] valuable insights through the parallel testing of diverse intelligent mobility solutions and thus make a contribution to climate protection. (Federal Minister Scheuer, link). According to the Handelsblatt, the automotive industry is also interested in secure data exchange: “The Federal Ministry of Transport is financing a first trial in Hamburg. There, the city is trying to use the funds to set up a “real laboratory” that will ultimately become the prototype of the nationwide data space. (Handelsblatt, 07.09.2020, link in German).
Guests that attended came, for example, from the fields of new technologies and mobility trends such as micro mobility and EV sharing. Together with T-Systems experts, they discussed the requirements for open data sharing for the mobility of the future.
Using the architecture blueprints of the International Data Spaces Association (IDSA), participants were able to share and bundle data while maintaining data sovereignty. Deutsche Telekom’s IoT business with the Telekom Data Intelligence Hub is a pioneer in the first implementation of these blueprints. At the 2019 Hanover Fair T-Systems introduced its IDS-based connector solution for sharing data with control over data governance (see IDS RAM 3.0, link)
While the hackers are developing, a virtual conference with mobility experts for B2B decision-makers and players and took place on the second day (3/19/2021). Individual lectures from practice and data science are planned, as well as interviews and panel discussions. Experts such as Dr. Daniel Dettling (Zukunftsinstitut), Florian Pronold (State Secretary for the Environment), as well as other leading figures from research and practice with a focus on mobility, shift in transport and sustainability were invited to participate in the discussion.
1st place: Team “GenAI”
The winning team proposed an all-in-one mobility app that connects all modes of transport from various mobility providers, including company internal modes of transportation. The app itself used AI and ML to recommend optimized routes with the idea of “eco-points”. The whole idea behind these eco points is to raise awareness regarding the environmental impact of each travel suggestion, as an example a (private) combustion engine vehicle will generate 0 points. To add to the gamification idea behind the app users can exchange these “eco-points” for rewards, ensuring that users are incentivized and motivated to follow the route of least CO2 emission.
Congrats to Bianca Schloo, Christian Esther Enrique Morales, Vipul Malik, Wenzhang Xu, Sakib Hasa, Aleksei Karetnikov
2nd place: Team “Future Mobility”
This team applied a DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm on the given dataset. The whole idea a DBSCAN algorithm is the clustering of points in a dataset that are closely packed. For bus stop location a K nearest neighbour was applied on the bus stop as well as on the DBSCAN data, if there was no bus stop as a neighbour in a given distance (e.g. 300m of user) a new bus stop can be constructed. Regarding less populated areas the team proposed that car sharing can be utilized. The amount of CO2 reduction was also calculated in order to give the user an idea of the environmental footprint
Congrats to: Peter Duronelly, Aron Asztalos, Marton Forgach, Zsolt Neszmelyi, Yasaman Abdollahian
3rd place: Team “MobAI”
This team created an app that offers an End2End solution for reducing congestion and CO2 emission during commuting. As shown by GPS data, coworkers often lived in the same areas in the vicinity of the office and share large parts of their commute. MobAI clusters these people into groups and offers a collective commute based on individual patterns and overall traffic load. Offering sustainability and individual preferences.
Congrats to: Cini Razzaghi, Andreea Rogojan, Jiaming Song, Luiza Sivestru, Alketa Bardhsohi, Niklas Schwertner, Daniel Garcia