Industrial autonomous system uses swarm intelligence

March 05, 2021 // By Christoph Hammerschmidt
Intralogistics vehicles organise themselves in swarms
The Fraunhofer Institute for Material Flow and Logistics (IML) is developing a swarm-based control of autonomous industrial equipment called LoadRunner

Using artificial intelligence and 5G wireless communication, LoadRunner can organise itself in a swarm and accept orders independently.

In order to cope with the continuously increasing parcel masses in online logistics centres, the LoadRunner should be of help to retailers in the future. The driverless transport vehicle can sort objects and transport them from A to B at a speed of up to ten metres per second.

"The vehicles can accelerate like a sports car and are entering a whole new performance class," emphasises Prof. Michael ten Hompel, Executive Director of the Fraunhofer IML in Dortmund, Germany.

The basis of LoadRunner is a swarm of drones that the researchers at Fraunhofer IML had previously developed. Here, 20 drones imitate the behaviour of a flock of birds. The individual drones orientate themselves on the behaviour of their neighbours and continuously adjust their flight direction and speed. Collisions are avoided, groups are formed and the interactions of the individual decisions result in a swarm intelligence that does not require central coordination. The swarm system uses simulation-based artificial intelligence methods.

The LoadRunner is equipped with an omnidirectional chassis, and the direction of travel and rotation are completely independent of each other. This means that the new transport vehicle can also turn in any direction during travel without having to manoeuvre. The robots orient themselves with the help of an on-board ground camera. This camera takes 400 frames per second of the previously scanned ground. This enables the vehicle to locate itself precisely even at high speed in tight formation drives.

Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.