We present a motion controller that generates collision free trajectories for autonomous Tugger vehicles operating in dynamic factory environments, where human operators may coexist. The controller is formalized as a dynamic system of path velocity and heading direction, whose vector fields change as sensory information varies. By design the parameters are tuned so that the control variables are close to an attractor of the resultant dynamics most of the time. This contributes to the overall asymptotically stability of the system and makes it robust against perturbations. We present several experiments, in a real factory environment, that highlight different innovative features of the navigation system-flexible and safe solutions for human-aware autonomous navigation in dynamic and cluttered environments. This means, besides generating online collision free trajectories between via points, the system detects the presence of humans, interact with them showing awareness of their presence, and generate adequate motor behavior. Index Terms-Tugger vehicles, flexible and safe autonomous navigation, obstacle avoidance, dynamic environments shared with human operators

The identification of stop locations in GPS trajectories is an essential preliminary step in obtaining trip information. We propose a neural network approach, based on the theoretical framework of dynamic neural fields (DNF), to identify automatically stop locations from GPS trajectories using their spatial and temporal characteristics. Experiments with real-world GPS trajectories were performed to show the feasibility of the proposed approach. The outcomes are compared with results obtained from more conventional clustering algorithms (K-means, hierarchical clustering, and HDBSCAN) which usually limit the use of the available temporal information to the definition of a threshold for the duration of stay. The experimental results show that the DNF approach not only robustly identifies places visited for a longer time but also stop locations that are visited for shorter periods but with higher frequency. Moreover, the self-stabilized activation patterns that the network dynamics develops and continuously updates in response to GPS input encode simultaneously the spatial information and the time spent in each location. The impact of the obtained results on systems that automatically detect drivers' daily routines from GPS trajectories is discussed.

This paper describes an implementation of the Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology for a demonstrative case of human queue waiting time prediction. We collaborated with a multiple domain (e.g., bank, pharmacies) ticket management service software development company, aiming to study a Machine Learning (ML) approach to estimate queue waiting time. A large multiple domain database was analyzed, which included millions of records related with two time periods (one year, for the modeling experiments; and two year, for a deployment simulation). The data was first preprocessed (including data cleaning and feature engineering tasks) and then modeled by exploring five state-of-the-art ML regression algorithms and four input attribute selections (including newly engineered features). Furthermore, the ML approaches were compared with the estimation method currently adopted by the analyzed company. The computational experiments assumed two main validation procedures, a standard cross-validation and a Rolling Window scheme. Overall, competitive and quality results were obtained by an Automated ML (AutoML) algorithm fed with newly engineered features. Indeed, the proposed AutoML model produces a small error (from 5 to 7 minutes), while requiring a reasonable computational effort. Finally, an eXplainable Artificial Intelligence (XAI) approach was applied to a trained AutoML model, demonstrating the extraction of useful explanatory knowledge for this domain.

No abstract

One of today's automotive research focus is the development of vehicles for the future, with their own intelligence, aware of their occupants, able to give support to its users, striving for natural and efficient interaction, and giving rise to the concept of the cognitive vehicle. Furthermore, truly adaptive intelligence can be achieved with assistance systems capable of adapting to different drivers. Vehicles with the potential to learn users' routines and preferences, and make decisions to prepare the next trip (e.g., manage comfort; check if the usual objects are being transported), is a concrete example that has started gaining attention. To accomplish such a challenge, data-driven approaches are required. Hence, datasets that include information on the habits of different vehicle occupants and their preferences are essential for building cognitive computational models. To the best of our knowledge, there is no tool capable of obtaining these data in a real-world situation. Thus, this work proposes a mobile application capable of collecting real data and creating datasets about: (1) where and when the driver and passengers get in and out of the vehicle;(2) objects brought/taken by the occupants; and (3) vehicle settings preferences. Collected data are internally structured in files that can be uploaded at any time. The developed mobile application can be described as an easy-to-use, flexible, and free of charge solution for collecting data on the travel routines of vehicle occupants, to support the development of personalized assistance systems.