In the past two years AI4Cities challenged the market to come up with AI-based solutions that can help cities reduce CO2 emissions in their energy and mobility domains. Amsterdam, Copenhagen, Greater Paris, Helsinki, Stavanger and Tallinn went through a PCP process and in the end selected seven suppliers to pilot the prototypes developed during the project.
Out of the seven suppliers, four are working on energy, three on mobility. We have created a page for each of these solutions, providing some information about where they will be piloted, which challenges they are addressing, and how they have experienced AI4Cities so far. As the pilots are progressing we will also provide updates of the first insights and results.
Holoni leverages AI beyond the prediction of solar production and predicts how much solar surplus can be generated from positive energy buildings. It also powers up a digital exchange by using IOTA, the next generation of green, fast, feeless and scalable blockchain, currently a candidate to equip European public smart service infrastructure. Energinet’s ORIGIN, now called Energy Track & Trace, is integrated to verify the origin of energy, hour by hour, anticipating the shift towards more granular certificates of origin
Enerbrain’s SPIKE is an all-in-one, cloud-based software/hardware platform supporting data exchange with proprietary IoT-enabled devices and communication with other IoT devices/platforms, knowledge extraction for situation-aware user interaction and engagement and building performance assessment. In addition it provides dedicated energy management services capable of innovative and effective optimisation of energy efficiency and flexibility in commercial, service and residential buildings.
The BEE(Building Energy Efficiency) solution combines several of the latest technologies to connect buildings with the energy grid and their environment and to optimise their overall emissions impact. Leveraging Platform-of-Trust – a common interface into different building management systems – ensures easy integration into as many different building types as possible. The AI engine uses the latest Deep Learning algorithms to predict the utilisation of the building for the next day.
C-in.Cityprovides cities with near real-time carbon emission monitoring. It empowers citizens by providing them with near real-time transparent information on how they can reduce their emissions, which also enables policy makers to design the most effective and actionable policy options, ranging from behavioural changes to technical investments, with more citizen buy-in. Finally, the solution also allows local SMEs to identify and rank cost-effective greenhouse gas (GHG) mitigation opportunities, and to offer localised insights to boost climate investments.
AVENUE is an innovative solution aiming to predict the impact of shared mobility services on urban transport GHG emissions. Avenue puts the data it receives into an AI-based analytical engine that enables demand monitoring and, based on it, it creates demand prediction models and GHG emission models, which could then be used for policy optimisation. Based on these policy adjustments, new demand prediction models are created, giving city planners a good overview of the impact of their policy actions.
The MPAT Tool (Mobility Policy Auto Tuner) is an engine to optimise the CO2 emission-reduction potential of city mobility policies, with a focus on shared micro-mobility. By understanding the geographies where a trip on an electric-powered shared bicycle, scooter or moped is most likely to create an emissions saving, the tool is able to make recommendations for areas to implement new policies (e.g. subsidies on rides, or removals of fleet caps). The impact of these policies can then be monitored with a specific view on CO2 emissions savings.
The Ix3 solution works by taking information from traffic cameras and from other traffic sources and then integrating it into the city’s traffic light system. Using deep learning-based artificial intelligence and visual sensing, the system drastically reduces the amount of sensors needed, and increases the amount of relevant data, leading to increased traffic fluency and a decrease of unnecessary stops and waiting. MarshallAi and Dynniq ensure that their system only identifies that an object is approaching a certain intersection, without giving an identity to that object.