3D & AI for safer, smarter, and more efficient cargo handling in terminals

What was the project aiming to achieve?

Optimization of Vehicles and Infrastructure: Utilizing sensors, data, and AI to enhance the efficiency of vehicles and terminals.

Environmental and Traffic Safety Improvements: Reducing energy consumption in internal transportation and enhancing traffic safety at the system level.

Competitiveness and Innovation: Promoting interest from major logistics players in research and collaboration with innovative SMEs.

Throughout the project, it became evident that integration into existing IT systems would be complex. Therefore, partners focused on independent solutions through web interfaces or apps to effectively present data and insights.

How was it done?

Needs-Defining Activities: Through workshops, site visits, and meetings, we identified the requirements for data and insights. Dialogues with partners and observations of operations yielded insights into how data on goods placement and movement could enhance operations.

Data Collection with Viscando 3D&AI Sensors: Towards the project's conclusion, data collection took place using Viscando OTUS3D sensors at DB Schenker's terminal in Gothenburg. The use of anonymized data allowed for the analysis and visualization of space utilization. Algorithm adjustments were made to enhance results.

Visualization of Insights from Measurement: 3D point cloud data from Viscando sensors was employed to create space utilization maps. These were updated twice per second and included dynamic maps of occupied areas, the age of placed goods, and statistics on fill levels.

What were the outcomes?

Real-time Solutions for Terminal Optimization: The project introduces innovative solutions through real-time data, visualization, and data analysis. These facilitate strategic decision-making, real-time monitoring for swift adaptation, and automated control based on current measurements, including indications of available truck bays.

Proposals for Effective AI Methods: To enhance the efficiency of the proposed solutions, two AI methods were introduced. Method 1 focuses on short- and long-term predictions of space utilization using established AI techniques such as Imitation Learning and LSTM networks. Method 2 combines pattern recognition, time-series prediction, and optimization to effectively identify and manage situations.

Prioritized Features for Optimized Terminal Management: DB Schenker identifies key features supported by the project's solutions. This includes smart control of incoming trucks based on available space, optimal resource planning with real-time information and prediction, and optimization of terminal space utilization through area adjustments based on needs and optimal cargo management.

Available Dataset for Future Development: The project provides an anonymized dataset with space utilization data from a receiving area over a week. This dataset serves as a foundation for future development of data- and AI-based solutions proposed within the project.