In response to the escalating challenges of last-mile delivery (LMD) due to various factors such as urbanization and the rise of e-commerce, this study explores the innovative integration of autonomous delivery robots within a Van-and-Robot (VnR) system, exemplified by the BeIntelli project.

In recent years, the integration of fully or semi-autonomous vehicles has attracted considerable attention. Technological innovations are propelling the logistics industry toward the adoption of autonomous delivery robots for LMD. This approach aims to enhance operational efficiency through optimized routing and extended operational hours, reduce costs primarily by lowering labor expenses, and promote sustainability through the use of electric vehicles. However, the integration of autonomous delivery robots into urban logistics faces multiple challenges, including limitations in their range or capacity. The conceptual framework that combines vans with robots offers a promising solution to extend the operational capabilities of robots and utilize vans as mobile depots, directly addressing the challenges of LMD. Despite the growing interest from both industry and academia there is currently a noticeable gap in the conceptualization and practical implementation of this concept, even though the necessary technologies are increasingly accessible. This study aims to broaden its research focus to include the technical design and development of such a VnR system, specifically within the context of the BeIntelli research project.

Employing a qualitative morphological analysis, including a morphological box and cross-consistency matrix, the research delineates feasible VnR configurations. This methodology, enriched by patent review, identifies essential categories and variants for the VnR system. For this concept, the integration into the logistics process takes precedence over specific technical construction details. Collaboration with experts, including developers of the delivery robot, the autonomous driving stack for the transporter, and integrators of the vehicle's material flow system, is crucial in constructing the cross-consistency matrix to determine viable combinations. A specific, project-constrained variant is developed and executed as a proof-of-concept within the BeIntelli project, showcasing the practical application of the theoretical framework.

The study aims to contribute to urban logistics innovation, emphasizing the importance of structured analytical methodologies and collaborative development for the advancement of autonomous delivery solutions. The implementation of constructive projects could be enhanced by a conceptual framework, facilitating detailed and practical investigations into the challenges described. Additionally, the utility is derived from the definition and identification of subcomponents and process steps within the concept, enhancing understanding of its constituents and, consequently, informing the design of the concept. Through the BeIntelli project, this research not only presents a viable implementation of the VnR system but also sets a precedent for future exploration in enhancing urban delivery systems.

There is plenty of research on the vehicle routing problem (VRP) and its application in last mile logistics. For home deliveries with an in-person handover offering time windows (TW) increases customer satisfaction and first attempt delivery rate. In general, it can be stated that the smaller the time windows the higher the successful first attempt delivery rate. To improve the predicted delivery time and keep the time windows offered traffic and congestions risks can be integrated in the VRP. In the logistics context most analyses focus on the B2C and B2B market with small package deliveries. At the customer location a fixed or stochastically chosen stop time is used.

In the field of service management the technician routing and scheduling problem (TRSP) considers varying services times at the destination. Also here, time windows have to be considered.

As the e-commerce market is expanding in terms of product portfolio heavier and bulky items like furniture and household appliances are added to the online stores. At the same time the vendors add services like final assembly or installation at the customers site to increase customer satisfaction. Due to the size and weight of the items to be delivered and the additional services offered the stop (and service) time at the customer site is longer and therefore relevant for an adapted VRP-TW.

The scope of this research is to analyze the processes at the customer location in more detail with a special focus on the process steps beginning with the search for a parking spot to unloading and the delivery of multiple items into the customers apartment or house. The objective is to identify patterns and indicators that allow for a more accurate prediction of the total stop time. One of the hypotheses is that deliveries in the city center take longer due to limited parking spaces which result in longer walking distances combined with multi-story buildings with narrow staircases compared to deliveries in suburban areas where the delivery van can park in the driveway.

To identify these patterns and indicators delivery service teams of a logistics service provider have been equipped with motion tracking devices. Based on the Motion Mining technology by MotionMiners which have been successfully applied in the context of manual warehouse picking process analysis the motions are classified into basic activities (e.g. walking, carrying, unloading, documentation) from which the duration of the individual process steps can be derived. Linking the empirical data with the logistics aspects of the delivery task (e.g. number of pieces, volume, weight, address, ZIP code) allows for a stop time prediction of similar delivery tasks in the future.

Overall, the results enable a more precise transport planning which could on the one hand optimize the utilization and workload of the delivery crews and on the other hand increase customer satisfaction as more accurate delivery time windows can be offered.