This paper presents an on-board multi-sensor system which is able to detect obstacles and estimate their distances in railway scenes in different illumination conditions. The system was developed within the H2020 Shift2Rail project SMART (Smart Automation of Rail Transport) and aims at increasing the safety of rail transport by detecting obstacles on the rail tracks ahead of a moving train in order to reduce the number of collisions. The system hardware consists of cameras of different types integrated into a specially designed housing, mounted on the front of the train. Multiple vision sensors complement each other in order to handle different illumination and environmental conditions. The system software uses a novel machine learning-based method that is suited to a particular challenge of railway operations, the need for long-range obstacle detection and distance estimation. The development of this method used a long-range railway dataset, which was specifically generated for this project. Evaluation results of reliable obstacle detection in various environmental conditions using the SMART RGB camera in day light illumination conditions and using the SMART Night Vision sensor in poor (night) illumination conditions are presented. The results demonstrate both the potential of the on-board SMART obstacle detection system in the operational railway environment and the benefit of the use of different cameras to be more independent of light and environmental conditions.

One of the most challenging technical implementations of today is self-driving vehicles. An important segment of self-driving is the ability of the computer to “see/detect” objects of interest at a distance which enables safe vehicle operation. An algorithm for the detection of railway infrastructure objects, namely, track and signals, is proposed in this paper to enable detection of signals which are relevant for the track the train is moving along. The algorithm integrates traditional computer vision (CV) algorithms, including Canny edge detection, Hough transform, and You Only Look Once (YOLO) algorithm, based on convolutional neural networks (CNNs). Each of the concepts (CV and CNNs) deals with a different object of detection which together form a unique system that aims to detect both the rails and the relevant signals. This approach ensures that the artificial intelligence (AI) system is “aware” of which route the signal belongs to. The reliability of the proposed algorithm in detection of a relevant signal, verified by the performed tests, is up to 99.7%. The metric method used for validation was intersection over union (IoU). The obtained value of IoU applied on the entire validation dataset exceeds 0.7. Calculated values of average precision and recall were 0.89 and 0.76, respectively. The algorithm created in this way solves the problem of detection of relevant signals along the train route, especially in multitrack scenarios such as stations and yards.

In this paper, an advanced thermal camera-based system for detection of objects on rail tracks is presented. Developed system is powered by advanced image processing algorithm, in order to achieve greater reliability and robustness, and tested on set of infrared images captured at night conditions. The goal of this system is to detect objects on rail tracks and next to them and estimate distances between camera stand and detected objects. For that purpose, different edge detection methods are tested, and finally Canny edge detector is selected for rail track detection and for determination of region of interest, further used for analysis in object detection process. In determined region of interest, region-based segmentation is used for object detection. For estimation of distances between camera stand and detected objects, homography based method is used. Validation of estimated distances is done, in respect to real measured distances from camera stand to objects (humans) involved in experiment. Distances are estimated with a maximum error of 2%. System can provide reliable object detection, as well as distance estimation, and for improved robustness and adaptability, artificial intelligence tools can be used.

Accurate models for heat load prediction are essential to the operation and planning of a utility company. Load prediction helps a heat utility to make important and advanced decisions in district heating systems. As a popular data driven method, artificial neural networks are often used for prediction. The main idea is to achieve quality prediction for a short period in order to reduce the consumption of heat energy production and increased coefficient of exploitation of equipment. To improve the short term prediction accuracy, this paper presents a kind of improved artificial neural network model for 1 to 7 days ahead prediction of heat consumption of energy produced in small district heating system. Historical data set of one small district heating system from city of Nis, Serbia, was used. Particle swarm optimization is applied to adjust artificial neural network weights and threshold values. In this paper, application of feed forward artificial neural network for short-term prediction for period of 1, 3, and 7 days, of small district heating system, is presented. Two test data sets were considered with different interruption non-stationary performances. Comparison of prediction accuracy between regular and improved artificial neural network model was done. The comparison results reveal that improved artificial neural network model have better accuracy than that of artificial neural network ones.

In this paper, the concept of Obstacle Detection and Track Intrusion Detection (OD&TID) systems related to the operation of trains is introduced, along with a potential concept for such a system. The main focus of the work presented here is the identification and description of system requirements and Use Cases (UC), their detailed classification, including general UCs for mainline railway and UCs specific to freight, as well as an analysis of the UCs and of the method used. The identified UCs have been organised with respect to the mode of operation, Grade of Automation (GoA), and operating conditions. The UCs were further analysed in different UC scenarios, including the pre-conditions, system response, actions made by OD&TID and associated systems and the post use conditions of the scenarios. The priority for implementation and complexity of each UC are discussed with respect to the probability of scenario occurrence and required interfaces. This work has been carried out as part of the process to evaluate implementation constraints, risks and requirements, and the operational scenarios of the OD&TID developed within the EU-funded Shift2Rail project SMART2, which aims to design and develop a prototype OD&TID system.