Skip to main content Skip to main navigation

Publikation

Bauabfälle: Einsatz von Künstlicher Intelligenz und Robotik für eine nachhaltige Kreislaufwirtschaft

Martin Wittmaier; Sebastian Wolff; Marco Wöltje; Ole van Laaten; Thomas Vögele; Babu Ajish; Yuhan Jin; Yi-Ling Liu; Tim Tiedemann; Matthis Trost; Philipp Meyer; Timo Lange; Peter Schaeidt; Joschua Marquart
Mineralische Nebenprodukt, TK Verlag, 6/2025.

Zusammenfassung

In the field of construction and demolition waste disposal, as well as commercial waste, largesized waste are generated that need to be sorted for high-quality recycling. While the sorting of small-sized waste is already well advanced in conveyor belt-based facilities, large-sized waste is still handled and sorted on construction sites, transfer stations, and sorting yards using manually operated excavators and cranes, much like 50 years ago. In the R&D project SmartRecycling-UP [14], the control of hydraulic cranes and excavators was automated using AI. The equipment was enhanced to not only handle waste but also sort large-sized waste according to material types. An AI controller was intended to learn the movement patterns of the machines, thereby enabling control with significantly less manual effort. The concept is based on three system modules: a Smart Motion Controller (SMC), a Smart State Estimator (SSE), and a Smart Process Controller (SPC). - The SMC controls the position and speed of the joints, moving the end effector of the excavator or crane along the desired trajectory. Instead of modelling each joint and controlling it via data from joint sensors – the AI controller acts as a black box. - The SSE module provides the feedback required for effective SMC operation. Instead of retrofitting excavators and cranes with joint sensors, the SSE uses AI to estimate the joint angles based on RGB camera images. - The SPC serves as the interface between object localization and material classification and the SMC. It plans a suitable trajectory for waste manipulation, based on data from object detection. For large-scale industrial use, the technology still needs further development – for example, by increasing the speed of machine movements, improving object and material recognition, and developing safety concepts for industrial-scale deployment. However, the results and operation of the AI-automated crane and excavator demonstrate that both existing and new machinery can be automated with manageable effort using AI, offering potential for cost reduction and environmental impact mitigation in the processing of large-sized waste. It is expected that AI will increasingly play a significant role in the circular economy in the future, particularly in the automation of hydraulically operated heavy machinery.

Projekte

Weitere Links