Minimizing the Impact of Regional Flash Floods
IAV (Ingenieurgesellschaft Auto und Verkehr GmbH) is an automotive engineering company that specializes in the development of innovative solutions for the automotive industry. IAV recently partnered with FiW, the Research Institute for Water and Waste Management of the technical university RWTH Aachen, and EG/LV, the regional water authorities for the Emscher and Lippe rivers, in a research and development project to create an early warning system that can predict urban floods due to severe rainstorms. The project uses the rain and environmental sensors that are found in many cars to record the amount of rainfall in specific locations for a specific period of time.
The potential for our automotive industry partners to develop new services with this data approach is nearly unlimited. We find MQTT is a reliable and cost-effective protocol for moving the data between the car and the cloud.
Flash floods due to heavy rainfall are an increasingly frequent occurrence. For example, on June 29, 2017, nearly 200 liters of rain per square meter fell in the Tegel district of Berlin, Germany in just one day. The average yearly rainfall of Germany as a whole is only 800 liters per square meter. Severe rainfall events such as this can cause extensive damage to city infrastructure as well as fatalities and injuries to residents. Forecasting exceptional rain events is difficult since they tend to occur in a micro-location. Typical rain forecasts focus on large geographic areas. A system that can warn authorities about localized heavy rain events could minimize the impact of the events.
“Our goal for this project is to use the rain sensors found in cars to help warn local authorities of a heavy rain event,” explains Sven Holzendorf, IAV Development Engineer. “The rain sensor information from a car, in a specific location, can then be combined with data collected from local weather stations and radar maps to visualize the location of a heavy rain event. Traditional weather radar maps provide a resolution of one square kilometer. Data from cars can provide information down to the square meter. This increased resolution gives local authorities a much more accurate map.”
IAV required a protocol to send the telemetry data to the cloud and over-the-air updates back to the devices in each car. IAV considered the use of either MQTT or HTTP for the protocol but quickly decided upon MQTT. IAV appreciated the lightweight nature of MQTT when compared to the overhead of using HTTP. Additionally, HTTP lacked the quality of service features that MQTT includes and HTTP client authentication was less straightforward.
The IAV development team first deployed the open-source Mosquitto MQTT broker on a local development machine. However, this was a short term solution since the Mosquitto MQTT broker could not be operated as a reliable service. Close collaboration between the IAV development team and the IT department led to the decision to use HiveMQ. HiveMQ allowed the IAV IT department to operate a reliable, high-availability MQTT broker that could be used by many different project teams. HiveMQ’s clustering features and professional support made it an easy choice for IAV.
"We knew that support for MQTT was going to be an important requirement for many new IAV projects”, explains Karsten Meichsner, IAV Internal IT Consultant. “For this reason, we wanted an enterprise-ready MQTT broker that can support numerous projects simultaneously. HiveMQ has proven to be an ideal solution that allows us to efficiently service the MQTT requirements of our development teams."
The flood management project has been successfully deployed on 100 vehicles equipped with a device that can access the windshield wiper data. The device transmits approximately 50 pieces of data every second to the HiveMQ broker. The MQTT data is then transferred to Apache Kafka for integration with radar maps and weather services information. Over-the-air updates to the devices are also published through HiveMQ to make maintenance of the devices transparent to the end-user. The end result is that the project was able to provide real-time awareness to local authorities about heavy rain events.
"Modern automobiles often generate tremendous amounts of location-specific real-time data”, explains Dr. Etienne Costa-Patry, IAV Technical Consultant. “This pilot project to combine automobile data with radar weather information is just one example of the type of new solutions that can be developed. The potential for our automotive industry partners to develop new services with this data approach is nearly unlimited. We find MQTT is a reliable and cost-effective protocol for moving the data between the car and the cloud."
