Autocrane: Level 3 Autonomy on a Log Yard Crane

A 350 ton timber loading crane in a paper mill unloads trucks and then loads a conveyor belt for further processing or a stack of timber up to 25 m high for intermediate storage of the timber. For each lifting operation there are tree trunks with a total weight of 35t in the crane's gripping device. The crane operator aims at an even loading of the conveyor belt and a fast unloading of the trucks. This process is error-intensive and dangerous for all parties involved, therefore the process should be automated to a large extent and equipped with a level 3 autonomy. The difficulty here lies in the large physical forces and the inertia of the crane inherent in the system. An artificial intelligence must be designed to meet these challenges and developed with the necessary expertise. The scale of this undertaking is unique worldwide.

Cameras, laser scanners and internal sensors allow the crane to "perceive" its surroundings, its own position and its own condition. As a result, the Artificial Intelligence created by PSIORI detects nearby trucks, the conveyor belt and the shape and condition of the stack of wood. On the basis of these inputs, the Artificial Intelligence determines the crane's plans for efficient and safe achievement of its objectives - unloading the trucks, loading the conveyor belt, temporary storage on the stack -, determines from them the necessary actions and carries them out. The crane is monitored via the data provided in real time on the current and planned behaviour of the crane by a human being who can intervene in what is happening at any time.

The AI of the crane can be classified as Level 3 autonomous due to its extensive autonomy with the possibility of human intervention. By optimising the variables speed, even load and safety, consistently high performance can be achieved in these areas. This minimizes labor costs and downtime and improves the quality of the wood chips in production.