Traffic jams during rush hours have characterised the picture in and around all the world's populous and economically powerful major cities for years. There are no signs of a trend reversal. According to figures from the German Federal Environment Agency, the mileage of passenger traffic in Germany increased by 28.5 % and that of freight traffic by 67 % in the period from 1991 - 2018. The constantly growing number of commuters clearly shows that an expansion of the available infrastructure will be necessary in the future. One promising technology that could help in this regard is the airborne transport of people and goods, especially on highly frequented commuting routes. An essential function for the long-term scalable and economical operation of such electric aircraft is the automation of the aircraft. However, a core component to ensure safe and automatic operation is a system for safe environment perception during take-off, landing and in-flight. A combination of different sensors is necessary for reliable perception of the environment in different weather and environmental conditions. Especially for the semantic mapping of the surrounding, i.e. the recognition of objects or persons based on the sensor data, classical rule-based methods are limited because deterministic methods are not real-time capable due to the large amount of data.
Artificial intelligence approaches, especially machine learning, have the potential to remedy this situation, as they can be highly parallelised and, despite their high performance and statistical properties, also enable a large data throughput and thus guarantee real-time capability.
SPLEENLAB develops and tests safe and certifiable AI algorithms for environment perception for automated flight in various projects. True to the motto Safety by Design, the young start-up is able to use the advantages of neural networks for safety-critical applications. Determinism and neural networks in harmony enable the incorporation of artificial intelligence in functions such as collision avoidance or ground risk estimation in real time and embedded, which SPLEENLAB is constantly developing.