KI4HE

In the two projects KI4HE and MaiSHU the work from the AHEAD project is continued and extended by modern technologies of environment perception, localization, semantic scene analysis, human-machine interface and teleoperation. In the KI4HE project, methods and technologies are being developed and implemented to enable semi-autonomous and remotely controlled SHERP vehicles to operate in a flooded and therefore difficult-to-access area by means of telepresence from a safe location. For this purpose, real-time monitoring of the vehicle's environment is implemented, and a representation of the situation is created to provide the operator in the safer control station with a clear overview of the overall situation. To protect humanitarian workers, teleoperation of vehicles will expand safe movement areas and enable organizations to deliver aid from a safe distance without risk to health. An important goal of the project is the presentation of the developed technologies and methods in a demonstrator. The innovative approaches developed in this project, such as environment perception, localization, semantic scene analysis, human-machine interface, and teleoperation, offer a broad intersection to autonomous driving and the digitization of traffic, respectively, and offer a high potential especially with respect to challenging situations in unstructured application areas.

The following work packages will be processed:

• Environment modeling based on multimodal sensor data (stereo cameras, IMU, GNSS)
• Localization of vehicles in a-priori unknown environments
• Semantic annotation of the environment model
• Innovative telerobotic operation using haptic rendering
• External view generation by means of a drone