InnoProm

The aim of the project is to use AI methods to safely control the suction excavator during operation in order to avoid dangerous situations and accidents. Robust detection of people in the vehicle environment is a fundamental prerequisite for this - even under difficult conditions such as fog, drizzle, swirling sand and dust, as is typically the case on construction sites. At the same time, productivity and precision are to be increased by automating complex and error-prone processes. At the end of the project, a comprehensive autonomy and safety concept will be implemented, which will be demonstrated with an MTS suction excavator both in the real environment and in simulations.

The project addresses the safe (autonomous) operation of suction excavators, with the concept being implemented using machine learning. Several goals are being pursued to facilitate the automation of both driving operations and suction arm operation while increasing safety and availability:

1. Definition of generic and universal AI interfaces for monitoring the driving functions of the suction excavator as well as the robotic arm through autonomous functions.
2. Environment and object recognition using AI, which is of central importance for the safe execution of all tasks.
3. Planning and securing the suction excavator based on the object detection results using suitable safety concepts.