"In order to significantly reduce the manual effort in the rental process, the automated assessment of completeness and condition when returning products plays a central role for us. The Quick Check enabled us to analyze various AI-based approaches and build a prototype with the most promising solution. We can now see that there is a lot of potential here and would like to thank the Fraunhofer IPA team!"
Edgar Müller
Akkurent GmbH
Contact at the AI Innovation Center
Maximilien Kintz
AI-based valuation of returns when renting out valuables
Initial situation
Sharing economy models are becoming increasingly relevant due to the new possibilities offered by the digitalization of our everyday lives and the growing desire for sustainability and resource efficiency. In the context of the sharing economy and especially when renting out valuable items, the efficient checking of the return for completeness, condition (cleanliness, damage, etc.) and function is of great importance. One particular challenge here is to minimize manual checks, as these are time-consuming and costly. In addition, the return process should be as convenient as possible for users of the sharing economy offer in order to achieve a high level of acceptance of such models. This Quick Check looks at the automation of the return process for electrical construction and gardening equipment. A particular focus here is on completeness, e.g. chargers, batteries, manuals, etc.
Solution idea
Different concepts for the automated return of borrowed items based on image data were initially developed and compared with each other. The aim was to maximize the benefits of AI-based processes while at the same time ensuring a high level of user acceptance. An approach was developed in which the end customers or operators (users) take a picture of the borrowed items when they return them. In this image, the return is automatically evaluated based on AI and the operator receives feedback on whether the return was successful. The AI should decide for the operator whether they need to check the device again and repair/clean it if necessary. To train the AI models, users can correct the AI-based completeness check at the beginning. This correction is incorporated into the training so that a reliable model is gradually created.
Benefit
In addition to the product benefits (professional products), renting and sharing also contributes to environmental protection and waste avoidance through the multiple use of products. Low fees and a high level of user-friendliness are of great importance for good acceptance of such systems. A high degree of automation for returns in the sharing economy can save costs, as the manual time required for a return can be significantly reduced. Furthermore, the process can be made convenient for users thanks to AI-based image analysis. Automated image analysis can therefore support an efficient rental process.
Implementation of the AI application
The return is checked for completeness, condition and function using a neural network based on the images taken by the end customer or operator on return. For this purpose, a network for object recognition was adapted to the task of checking for completeness. Compared to object recognition, the exact size of the detected objects is irrelevant; instead, the focus is on deciding whether a searched object is contained in the image. The objects detected in the image are compared with the database of lent objects and missing objects are identified.