AWaCER
Automation of warehouse management through the use of an autonomous drone with multi-channel radar sensors and AI.
The AWaCER research project aims to revolutionize warehouse management by introducing a sophisticated drone system equipped with a terahertz radar sensor system with real-time image processing capabilities. This system is designed to automate the complicated process of item-level inventory counting by seeing through opaque packaging, thereby improving the accuracy and efficiency of inventory management. By integrating cutting-edge technologies such as artificial intelligence, LiDAR and time-domain terahertz spectroscopy, the project aims to develop a comprehensive solution that can be seamlessly integrated into existing warehouse management software via a generic application programming interface.
Our role in the project
In the project, the institute takes on the role of scientific partner for the development of new, AI-based methods for the automated counting and identification of objects and boxes in the warehouse. In addition, processing algorithms are being developed that enable accurate localization of the objects. One of the main competencies at the August Wilhelm Scheer Institute is the transfer of proven technologies from one domain to new use cases in new industries. The institute's tasks fall into work package 1 and work packages 3 to 7.
work package 1
Goal: Defining requirements and framework conditions for sensor, software and drone development.
Tasks of the institute:
- Definition of the required software output and the necessary data input. Contributing expertise in the field of machine learning and estimating the necessary data volume to meet the specified data analysis requirements.
- Participation in the interface definition and data transfer Coordination of the requirements for the resource expenditure of the analyses on the drone and in the downstream processing
Total result: solution architecture, requirements analysis, interface specification
work package 3
Goal: Adaptation of the existing drone to integrate the developed sensor including the required software
Tasks of the institute:
- Adaptation of software requirements based on real drone development in the areas of computing resources and data streams used
overall goal: Fully functional drone for use in the warehouse area with integrated radar sensor system.
work package 4
Goal: Development of algorithms for processing the raw radar sensor data via data fusion with the data generated in the drone
Tasks of the institute:
- Annotation of the generated data in preparation for AI development in Work Package 5
- Use of existing annotation tools for manual counting and identification of pallets, boxes and objects
- Development and evaluation of automated annotation software using metrics
- Creation and annotation of test data to detect anomalies
Total result: Software component for drone data preparation, data aggregation and annotation
work package 5
Goal: Development of AI-based analysis and control components based on the data stream generated in the drone
Tasks of the institute:
- Development of AI-based methods for:
- detection of pallets, boxes and objects
- Detection of anomalies on or in pallets, boxes and objects such as broken objects or leaking liquids
- Identification of free spaces for comparison with the storage system and for optimizing storage capacities
- Detection of objects in a warehouse with a wide variety of stored goods
- Processing of the individual results to create a complete record of the warehouse and the contents contained therein
- Creation of automated recommendations based on the results of the individual analyses
- Support in developing the interface to existing warehouse management systems
Total result: Fully automated warehouse analysis software based on the sensor data generated by the drone in the GCS (Ground Control Station)
work package 6
Goal: Early tests, evaluations and iteration loops of the system components to ensure the applicability of the developed overall system
Tasks of the institute:
- Software testing of AI-based analyses and data transfers
- Validation of results based on varying lighting conditions, obstacles, boxes, objects and storage types
- Iterative development and optimization of AI models to maximize performance
- Real-world testing of the entire system in warehouses in Germany
Total result: evaluation report, optimization of the developed system
work package 7
Goal: Ensuring project success and subsequent utilization of results
Tasks of the institute:
- Project management
- Scientific exploitation through the presentation of scientific publications at international conferences
- Communication in Germany
- preparation of project reports
Total result: project reports and ensuring the success of the project
The initial situation
Overview of the need for action
Warehouses are a central part of the economy, serving not only to buffer supply chain disruptions but also to dispose of inventory. Efficient inventory management is crucial, but companies are facing new challenges as demand for warehousing and logistics solutions soars. Current warehouse management is mostly manual, leading to errors and inefficient processes. Item-level inventories are hardly possible, and growing pressure from e-commerce is exacerbating the problem. To meet these challenges, global warehouse management is increasingly moving towards automation, with drones, robotics and AI-based systems playing a central role. Advanced technologies such as terahertz radar sensors are offering new opportunities, especially for drone-based inventory management solutions, which can reduce labor and compensate for labor shortages.
Inefficient and error-prone manual warehouse processes
Lack of precise inventory overview at item level
Lack of real-time data for optimization and decision-making
Increasing pressure from e-commerce and lack of automation
Your contact person
AWaCER
AI-based counting at carton and item level in the warehouse by an autonomous drone equipped with a multi-channel radar sensor
Laura Bies
laura.bies@aws-institut.de
+49 162 2934 499
Your contact person
AWaCER
AI-based counting at carton and item level in the warehouse by an autonomous drone equipped with a multi-channel radar sensor
Laura Bies
laura.bies@aws-institut.de
+49 162 2934 499
Our solution approach in focus
AWaCER's solution approach is based on the integration of a specially developed multi-channel radar sensor into an existing drone system to enable precise detection and counting of inventory at item level.
- Development of a multi-channel radar sensor
The radar sensor developed is specifically designed to meet the requirements of warehouse management and integration into drone systems. It uses short-range radar technology to enable detailed detection of objects through pallets and boxes. - Integration of the radar sensor into the drone
The multi-channel radar sensor is integrated into a drone system, enabling data fusion and thus the creation of a complete 3D image of the warehouse and inventory. This visualization provides a precise overview at item level. - Development of AI-based software for counting and identification
Based on the data captured by the radar sensor, software is being developed that uses artificial intelligence to uniquely identify and count pallets, boxes and the items they contain. This automation reduces manual errors and significantly speeds up the counting process. - Evaluation and integration into existing warehouse management systems
Finally, the entire system will be tested in real warehouse environments to verify its efficiency and ensure that it can be seamlessly integrated into existing warehouse management systems. The solution is scalable and customizable so that it can meet the different needs of warehouse operations and can be quickly put into operation.
This approach ensures accurate inventory tracking and helps reduce errors and increase efficiency while leveraging cutting-edge sensor technology and AI-powered processes.
Funding notice
The AWaCER project is funded by the Federal Ministry for Economic Affairs and Energy.
Funding code: KK5612101MS3
Running time:01.07.2024-30.06.2026
