With AI into the Silicon EconomyWith new services, technologies and applications, Silicon Economy research is creating the basis for the commercial use of artificial intelligence in the platform economy.
THE NEW “ECOSYSTEM OF ARTIFICIAL INTELLIGENCE”
Open and federated platforms are the basis for the economic use of AI solutions and for new services, technologies and applications in logistics and supply chain management. As part of the Silicon Economy large-scale research project – funded by the Federal Ministry for Digital and Transport Affairs – scientists at the Fraunhofer Institute for Material Flow and Logistics IML develop highly scalable solutions for widespread logistics applications and processes. They make these basic principles available to the logistics industry as open source. Open source – both the active development and the pure use of software – is now hugely important in logistics.
The three Silicon Economy communities – “Transport Information”, “Computer Vision” and “Automation & IoT” – offer companies and scientific organizations a wide range of opportunities to collaborate on solutions for the future of logistics.
All research results from the Silicon Economy are published open source in the repository of the Open Logistics Foundation.
Communities & projects
Electronic Transport Information
Electronic transport documents offer a great opportunity to automate and link previously manual processes and activities. A large-scale introduction has so far failed due to a lack of compatibility between the different solutions of many players. The open source-based developments are intended to help harmonize the various approaches and establish a network of federated platforms.
The paper-based consignment note for international road freight transport (CMR) requires a great deal of administrative effort with numerous manual activities due to many different formats, media disruptions and transmission errors. As part of the Silicon Economy, the researchers developed a web-based open source service for generating, storing and forwarding digital consignment notes (eCMR), which is on its way to becoming a European standard.
In future, it should be possible to combine several accompanying transport documents in one digital folder, including the digital consignment note eCMR. So far, there is no solution for bundling digital freight documents, but there is great interest from the industry. The project builds on research results from the Silicon Economy project „Digital consignment note – eCMR“.
The reference implementation of an eFTI platform, which enables authorities to retrieve electronic freight documents, is based on Silicon Economy’s own eCMR development using the example of the digital consignment note.
Computer vision (CV) and video analytics (VA) as methods of artificial intelligence and machine learning now promise innovative approaches to increasing process transparency, optimizing processes and thus reducing costs and emissions. The open source developments of the Silicon Economy should now create the basis for the reliable industrial use of AI algorithms for logistics.
The training and validation of algorithms are crucial for the development of machine learning processes. Therefore, the focus is on the development of a continuous ML-Ops pipeline – a software process from training on a (cloud) platform to (simulation-based) validation and optimized deployment on hardware, e.g. on mobile robots, on cameras in incoming goods or at the truck gate.
The loading of parcels and goods is currently a highly manual process in logistics. As a result, there is no direct feedback between the current status, e.g. the fill level of a truck, and the planning system. As a result, trucks are often not loaded optimally or load carriers are not loaded at all, for example due to previously unrecognized pallet overhangs or a lack of stackability.
Yard Lense on Edge
Process optimization through AI is a decisive lever for competitiveness in logistics. Reliable camera-based localization, classification and identification of logistical objects – such as pallets and means of transport like pallet trucks, forklift trucks or robots – using AI algorithms provides the decisive data basis for autonomous process optimization.
Automation & IOT
The internal material flow is the lifeline in warehouses and production. Autonomous and embedded systems make it possible to generate and control data. Whether as mobile robots or in the form of IoT devices, they are indispensable for the digitalization of supply chains.
The VDA5050 standard developed by the VDA and VDMA is an interface for standardizing communication between automated guided vehicle systems or vehicles (AGVs/ATVs) and a master controller. Today’s systems are primarily stand-alone solutions, i.e. the control of AGVs/ATVs is proprietary and therefore manufacturer-specific. The libVDA5050++ – a development from the Silicon Economy – now offers a generic open source implementation of the standard that enables manufacturers of AGVs to convert their vehicles to the standard quickly and without major effort.
There are currently hardly any heterogeneous fleets of mobile robots on the market. Initial efforts towards standardization such as VDA5050 have created great added value, but are limited to the interface between robot and fleet controller and neglect robot-specific topics such as fleet coordination or process mapping and connection to the customer’s system landscape.
In many industrial sectors, defined transportation, handling and storage conditions must be observed for goods. Despite this, transportation is often handled by standard trucks and the regular logistics locations of logistics service providers. The compact Sensing Puck track & trace device now enables shippers and recipients to document and track the transportation conditions and status of the products.
O3dyn Simulation Model
In contrast to most automated guided vehicles, which are either highly efficient, dynamic or flexible and designed for indoor or outdoor use, the O3dyn mobile transport robot combines all three characteristics. O3dyn is not only highly dynamic and autonomous, but also designed for hybrid operation. It can transport large loads the size of a pallet omnidirectionally at a speed of up to 36 km/h. The O3dyn simulation model is a virtual representation of the real outdoor robot.
The Silicon Economy is an initiative of Fraunhofer IML and a research project of the Federal Ministry for Digital and Transport. The project partners – Fraunhofer IML, Fraunhofer ISST and the TU Dortmund University – are creating the foundations for a federated and decentralized platform ecosystem in logistics on the basis of open source.
For further information please contact: