Artificial Intelligence in mobility

The future of mobility is one of the most discussed topics of today. Against the background of increasing traffic, limited availability of space in urban areas and the need to reduce traffic-induced noise and pollution, there is no doubt that the way that people and goods move will have to change tremendously. Artificial intelligence is a key enabling technology to master the transition to highly-individualized, environmental-friendly and autonomous mobility systems.

appliedAI is part of the mobility HUB in Germany, funded by the government. Naturally, we look at mobility in detail and see it as a field that will be transformed by AI quite heavily. The reason is that many of the applications on modern mobility rely on very basic assumptions that AI questions, namely that computer systems are not capable of handling certain problems and tasks well or at all. This has nothing to do with "intelligence" or "automation", but rather with what machine learning and its bigger brother (ro goal) artificial intelligence enable machines to do: See, listen, interpret and understand, discovery, predict and much more. They are all at the core of the market hypothesis of modern mobility companies or rather the idea that machines do not have these capabilities.

Below you find an overview of the potential applications of AI technologies in mobility that can show you how AI enables the future of mobility. This is not a complete list but shows the more obvious examples and areas of applications. Reach out to us for more on "AI in mobility".

Passenger transportation

Driver assistance systems: Autopilots run with artificial intelligence models can support the driver by taking over tasks such as adjusting the car’s speed to that of surrounding cars, detecting obstacles, steering or braking (Level 2 Autonomy).

Demand management – Car sharing: With the help of Floating Car Data (FCD) procedures, car-sharing providers can use and analyze this data with ML algorithms to detect movement patterns and predict demands. Deployment of the cars can thereby be optimized and utilization of cars can be increased.

On-demand public transportation: On-demand public transportation is operated by artificial intelligence: real-time data collected on traffic conditions and customer requests is used to calculate the vehicle's route optimally.

Autonomous trams: Autonomous trams navigate throughout the city without a driver. Multiple sensors capture data from surrounding traffic and the environment, which is analyzed and used to safely operate the tram in an unknown urban situation

Automated valet parking: Through communication of the parking garage with the vehicle. Cameras and sensors in the garage help depict free parking spots, while the car autonomously navigates and parks itself in the designated parking spot.

eVTOL vehicle (Lilium): Lillium’s all-electric, vertical take-off air taxi transports 5 passengers and flys autonomously using AI. Thereby, it aims to offer autonomous on-demand air travel in urban areas in the future.

Connected services

Predictive maintenance (airplanes): Airlines can decrease delays and flight cancellations significantly by using predictive maintenance for their airplane fleet.

Fuel planning and optimization: AI helps airlines to save money and decrease CO2 emissions due to optimized fuelling of airplanes. Based on route distance, weather conditions, aircraft types, altitudes, and further data, ML models predict the optimal amount of fuel needed for a flight.

Prediction of flight movements: Air traffic controllers can be supported by artificial intelligence in the future, which will reduce their workload while enabling them to handle more air traffic. The AI is then used to predict flight movements in a four-dimensional space.

Logistics on the ground

Unmanned last-mile delivery vehicles: Unmanned ground vehicles can use artificial intelligence to autonomously deliver packages from distribution centers to the end customer. The last mile delivery robot can help take over the last step in the delivery process.

Autonomous trucks: Costs and time can be saved and road security increased by autonomous trucks. Using machine learning techniques, trucks are learned to navigate autonomously on highways and thus safely deliver goods across countries.

Route optimization of trucks: Neural networks help logistics companies in planning trips optimally, so that the amount of empty runs is minimized. This helps to save time, money and CO2 emissions and increases the productivity of the trucks.

Platooning of trucks: A number of trucks create a convoy on the highway using a vehicle to vehicle communication, sensors to detect the surroundings and autonomous driving functions. Thereby, costs can be saved due to lower gas consumption and road safety can be increased.

Predictive planning of demands: Machine Learning can be used to predict the demands of goods. Thereby, supply chains and truck allocation can be optimized and costs can be saved by avoiding an over or undersupply of goods and adjusting inventories to the demand.

Autonomous Trains: Autonomous trains are used in freight train networks to increase safety, productivity and flexibility as well as to reduce bottlenecks.

Real-time railway system optimization: Railway systems can be optimized by using sensors that are placed on the switches of the rails and collect real-time data. The data is analyzed by an AI, which helps to predict possible failures of the rails and thus decrease costs and system downtimes.

Logistics on the sea and in the air

Drone delivery: Unmanned flying vehicles use AI to navigate within an unknown surrounding in the third space. Thereby, they can be used to deliver goods in urban areas through the air without boundaries on the ground, such as delays through traffic or human errors.

Remote-controlled ships: Space on freight ships can significantly be increased by remotely controlling them. The ship navigates autonomously while being monitored from a shore control center, which makes crews onboard unnecessary. On-board drones can be used for inspection flights.

Stowage plan optimization: Simulation of optimal loading plan for freight ships bearing in mind ship-related constraints (capacity, layout) and limiting factors at the harbor (crane intensity, cargo handling) for optimizing loading efficiency and overall turnover